• Word-A-Day: I built and launched a vocabulary game called Word-A-Day. Every day a new word drops. Players come to the app to learn the word. If they choose to accept the challenge, they have until the end of the day to use the word in a post on social media. After each day ends, their submission gets scored by an AI for correctness (did you demonstrate word mastery?), authenticity (do you sound human?) and originality (independent thought). Players build streaks and can climb the daily, weekly and global leaderboards. It’s currently only available to play on Farcaster, but will be coming soon to other platforms. If you want to learn more, you can read my launch announcement, or check out the website where you can also get on the waitlist.

• Warpee.eth AI Bot: The first real project I worked on when I got back into coding was an AI bot with access to all the video content in the GM Farcaster library. We had been livestreaming for years and built up a huge knowledge base of Farcaster history but since video isn’t inherently searchable like text, there was no way for people to interact with all that content. Agents had not yet taken off when I first designed the bot using a basic RAG system, so as much as the bot could answer some questions, it wasn’t reliable or consistent. I finally got around to upgrading Warpee’s basic framework from workflow-router/RAG to agentic workflow with skills. Warpee now has a handful of skills, so now when someone tags him and asks him a question, he decides which ones to use. This framework is much easier to build on top of because when I want to give him more capabilities, I can just add more skills to his library.

• x402 Experiment: If you believe agents will be doing more and more activity on behalf of humans, and if you believe that APIs will be how agents communicate and transact with each other, then it follows that agents need a native way to exchange value (aka pay for things). Enter the world of x402 and more recently MPP. These are protocols that wrap APIs behind crypto payment methods and only process the API request after payment is sent. I really wanted to test this concept out so I built a simple prototype and tested it to see it work.

• Paid Down Tech Debt: Thanks to a scary breach at Vercel a few weeks ago and two recent high profile supply chain attacks, I decided to spend time hardening up some of my code repositories and coding practices. I wasn’t impacted by any of the recent attacks but it was a good wake up call.

  • Rotated all secrets and keys across all projects (took 3 days - probably because it’s one thing Caude can’t do it on its own!). Also cleaned up unused env vars and documented the process so it will be easier in the future.

  • Migrated transcript embeddings to a new index with larger and overlapping chunks and newer embedding model. Pointed all consumers to new index. The plan was to sunset the old one, but search results are not actually better, likely because larger chunk size means less precision, especially for keyword searches.

  • Migrated the lore database from a JSON file in S3 to Postgres SQL on Neon. Pointed all consumers to new database and sunset the S3 files.

  • Migrated the lore scraping job to my admin dashboard in the cloud so I don’t have to run it locally on my PC anymore.

  • Consolidated the gmfarcaster superfan snap into the gmfc monorepo so all gm-farcaster projects are together.

  • Refactored the gmfarcaster website to make the episode detail pages fully crawlable by search engines and bots. Previously the transcripts were generated client side so it wasn’t being indexed.

  
  

Reflections

• I started using Claude in the terminal and have given it more and more control over decisions. When Claude code first came out and everyone said they were using -dangerously-skip-permissions , I was still reviewing everything Claude was outputting. As I had time to get comfortable I started reviewing less and less code and I’m pretty much at the point where I’m not looking at the code at all. It’s still extremely important to understand your risk profile, blast impact, technical architecture, etc. while building, and if I had a database with PII or was collecting payments that would be different than protecting a database with a list of vocabulary words and podcast transcripts.

• Micro payments were cool to experiment with. No one actually used the prototype I built (the wallet has exactly $0.001 in it which was the payment I sent for my test) and I’ve since taken it down. Does that mean it’s not worth experimenting with? Of course not; it’s cool to get hands-on experience with emerging patterns. When the time comes for me to have an idea that a bot would want to buy, I’ll be ready to go.

Other Projects

• Clanker Ecosystem Fund was announced. The fund’s mission is to distribute protocol fees to creators and communities who positively contribute to the Clanker and Farcaster ecosystem. It was also announced that GM Farcaster will be acting as the stewards of the fund. There has been a lot of discussion about grant distribution and how fees are best distributed to a community - from pure KPI and algorithm based to opinionated, human directed decisions. There are also lots of questions about who is eligible to receive grants. As we’re just getting started, a lot of these questions will be made clear over time, but the one thing I keep telling people is we will always stay true to the mission of rewarding people who are positively impacting Farcaster/Clanker. Prof and I have been daily users of Farcaster for almost 4 years and have been documenting and highlighting all the incredible builders and we’re both very excited to be acting as stewards of the fund.

• Bat, A Friend to All is coming soon! The children’s book I started with my sister over three years ago is closer than ever to being ready for publishing. She finished all the art (check out the website for previews) and now we’re working with a graphic designer to prepare the final manuscript file for uploading to self-publishing platforms. I’ve continued my research into the world of self-publishing and for our book we’ll likely do a combination of:

  • IngramSpark for distribution through bookstores and libraries

  • Amazon KDP for distribution through Amazon

  • Direct Sales for distributing autographed copies through our website

  • If anyone is interested in the economics, estimated earnings per book for a 24 page children’s book with a list price of $24.99 is: $1.50 thru book stores, $6.00 on amazon, and $15 direct sale

Some of the (other) things

Roundup of some interesting things I’ve come across:

• On the internet

  • Free web hosting for agents, here.now

  • Enclose the horse game, enclose.horse

  • Duo Lingo for everything, wondering.app

  • Ne scal ic nefre hit forȝeten, naht uuhiles ic libbe! (If you can’t read that, check this out)

  • Correlation does not equal causation. Fun site for data people. Get Spurious.

• Watched movies: Sinners, Dog Day Afternoon, Whiplash. Read book: The Vegetarian (had no idea what i was getting into but found it so unexpected and good payoff, was described as Kafkaesque). Saw show: Dog Day Afternoon (on broadway).

• Quotes/mantras that have been staying on my mind:

  • Insults are mirrors more than they are weapons

  • When something is painful, do it more often.

  • There is an infinite supply of love.

Some of the recent pics

Winter into Spring here in the north east

A Note to My Readers

I started this substack in December 2022 as an experiment to see if developing a writing habit would help clarify my thinking and/or provide other benefits. It has evolved into different formats over the years, but it currently serves as my public journal.

I use it to share musings, meditations, and links to things I’m finding interesting as I build out a farcaster-native media company, a modern technology consulting company, raise my kids, and have fun creating and learning in the worlds of crypto, tech, finance, science and wellness.

Thank you for supporting my writing and journey. If you’d like to get in touch you can reply to me here or find me on X and farcaster.

Until next time, keep putting good into the world. —adrienne🌏❤️

Introducing Word-A-Day — a Farcaster miniapp designed to help you grow your vocabulary, sound smarter, and make learning social.

How It Works

1. Learn the word. A new word drops at 00:00 UTC. Open the miniapp to see the word, its definition, an example sentence, and hear its pronunciation.

2. Use it in a cast. If you accept the daily challenge, you have until the next word drops to use the word naturally in a Farcaster cast or reply. This proves you actually learned the word.

3. Get scored. At the end of the day, an AI evaluates your cast on correctness (accurate, contextual usage), authenticity (natural, human voice), and originality (independent thought). If your cast reads like it was generated by an LLM or copied from someone else, you’ll lose points.

4. Climb the leaderboard and build your streak.

The miniapp is powered by $wordaday, an ERC-20 token on Base. To participate in daily challenges, the app requires holding a minimum balance of the $wordaday token. The token is used solely for access and gameplay and has no use outside of the app.

What Makes Word-A-Day Unique

There are a million learning apps that try to create a daily habit where you open the app, take some action, get a quick dopamine hit, and repeat. What makes Word-A-Day different is that it introduces a learning mechanism that really only works on Farcaster:

• SOCIAL. It’s played out on a public social feed. The app itself is very light. It just shows you an index card with the word and its definition. There’s not much to do inside the app. The action happens on the Farcaster feed. When you cast, you don’t need to announce you’re playing. You don’t have to tag anyone or link to the miniapp. You just use the word. It’s subtle. And won't it be interesting when people who aren’t even playing start noticing the same word appearing across the feed?

• ACTIVE. Learning is most effective when it's active, not passive. Most learning apps show you a concept and ask you to read or watch something. This app forces you to apply what you’ve learned during the course of your day. You’re far more likely to remember a word you’ve actually used than one you’ve only seen on a screen.

• CRYPTO-NATIVE. There’s a revenue model built in. Because it’s built on Farcaster, where everyone automatically has a crypto wallet, I can add a simple token gate requiring users to hold a token before they play. It’s like asking for a quarter at the arcade. I earn small fees when someone acquires the token, which helps cover infrastructure costs and fund rewards. Adding payment rails to a traditional game (payment providers, subscriptions, credit cards) introduces significant overhead that usually doesn’t make sense for a small app.

• ROOM TO GROW. We often say Farcaster is powerful because you can get your first 100 or 1,000 users. A small group of beta players helped me test this app and surface edge cases I wouldn’t have found on my own. Farcaster works as an incubator. If this apps takes off on Farcaster, I can expand it to other platforms. I’d love to see a Word-of-the-Day mechanic take off at the scale of something like X. And there are many other “___ of the Day” ideas that could follow the same pattern of app → social feed → AI scoring.

Ready To Play?

If this sounds interesting, I invite you to play.

Open the miniapp inside Farcaster:
https://farcaster.xyz/miniapps/eDqZkiLWJ0IZ/word-a-day

If you don’t hold 10M $wordaday, the app will prompt you and guide you through acquiring it.

If you’re not yet on Farcaster, you can create an account here using my referral code:
https://farcaster.xyz/~/code/EM02EP

Have fun playing. I'm building this in public, so tag me on Farcaster with feedback or ideas.

Thank You to the Early Users

I want to give a quick shout out to the people who helped test the game and shape the mechanics over the past few weeks while I was building it:

Completed 1-5 challenges:

greg, christopher, logonaut.eth, trh, bertwurst.eth, pjc, yerbearserker, catfacts.eth

Completed 6-10 challenges:

tracyit, bradq, links, jacy, nounishprof, sgniwder

Completed over 10 challenges:

yes2crypto.eth, jachian, music2work2

Part of the reason is obvious. Humans won’t be coding anymore because AI will write all the code, so why would someone choose to study Computer Science when it prepares you for jobs that won’t exist anymore?

I’m seeing the shift with my own eyes. I’ve been using AI for coding for over a year, and it’s allowed me to finish several end-to-end projects using frameworks I had no prior experience with. Projects that would have taken a professional team of engineers several months to build, I finished in weeks. And it keeps getting better and faster. Just a few weeks ago I switched to Claude Code, and I’m blown away by how much better it is than what I was using before. Every spare minute I have, I’m at my laptop telling it to write more code. If Claude can spin up real, working software in minutes, what does “learning to code” even mean anymore?

And it’s not just me. Every day I see more examples of people with zero coding experience building fully functional, professional, polished apps. At the same time, we hear stories about how dire the job market is for new CS grads.

So, the question for the day is: is computer science still worth it? Why should anyone choose to study it in this day and age?

I can’t answer “is computer science a good college major today.” I really don’t know. But I had an experience helping my son with his homework a few weeks ago that made me sure of one thing: there is still real value in learning computer science.

My son is taking AP Computer Science in high school. He’s learning object-oriented programming in Java. A lot of the coursework is what you’d expect: syntax. Variables. Loops. Arrays. Classes. All the rules you have to memorize just to make the code compile.

And he hates the moment when he hits the little triangle “run code” button, waits a few seconds, and his screen fills with a wall of error text that tells him he failed, but doesn’t tell him why.

A few weeks ago, he started getting really frustrated working on his homework. I gave my usual advice, reassuring him: “it’s okay to struggle, work it out, you’ll get it.”

He rolled his eyes because I can be a broken record, but he pushed back and told me, “I don’t mind hard work. I like to be challenged. I just don’t like this. This is different.”

He’s wicked smart. We’ve been playing chess recently and I can’t even come close to beating him. He sees the entire board and is always ten moves ahead of me. School has come easy to him. So, when he tells me this is different, it’s not because he can’t understand the concepts. It’s because it’s a different kind of hard.

Here’s what his homework assignment looked like: they were given a CSV file (Pokémon data—name, type, image URL). They had to load it, split strings into arrays, store the data in a class, and then write methods to query it. The assignment starts with tiny tasks and then, at the end, asks you to put everything together into a working program.

When he struggles, it’s almost always with that last “put it all together” step. The earlier parts of the assignment are small and self-contained: load a file, split a string, print values inside a loop. They’re approachable and not overwhelming. Sometimes they’ll even nudge you towards the right answer with embedded code comments and hints. It’s easy to get through those first tasks with a little trial and error without fully understanding what’s happening. But the final “put it all together” step builds on all of it. If you didn’t really grasp the earlier pieces, you can’t fake your way through the last question. You get stuck. The homework is no longer about knowing the right syntax to get your code to compile, but about knowing how to make a program work. And ultimately, this “put it all together” step is where real learning happens.

He comes to me for help with his CS homework because “mom is the tech person.” And he expects me to answer instantly, as if he asked me the square root of 64. But I can’t. I have to sit down, read, think, and troubleshoot.

After helping him a few times this year, I’ve realized the problem-solving steps are always the same, regardless of the assignment:

1. Understand the question. Don’t touch the keyboard until you’re clear on what you’re being asked to do.

2. Simplify. Start with a working program by simplifying it until it compiles successfully. Comment out everything you can. Get to a baseline.

3. Add back one change at a time. Write one line of code and then test to see if it still compiles successfully. Do this until you find your first error.

4. Instrument. Use print statements to see what the computer is actually doing.

5. Repeat until it works. Once you find the first error, fix it, then repeat step 3.

When his code finally compiles and he submits the homework, he feels satisfied. But what did he learn? Did he learn that “arrays are 0-indexed” and “you can’t store a decimal in an integer variable”? Well, yes, but the real learning is knowing how to approach a messy problem without panicking.

He is learning how to break down problems. How to reduce complexity. How to test assumptions. How to get comfortable not knowing something. And he’s learning that he can figure things out.

This is why computer science still has value. I have no doubt we no longer need to know code syntax because AIs will be writing all the code. But these problem-solving skills, what I like to call the “engineering mindset” is priceless.

Some of the (other) things

As always, a list of some things I’ve been doing and finding interesting:

• Finished reading The Correspondent, by Virginia Evans. This book made me feel something, which is honestly all I look for with fiction. Definitely recommend it, if you like stories about the human experience and care more about character arcs than plot.

• Continuing to read Eastern Body, Western Mind; Chess for Dummies; and just started The Jews by Howard Fast. I used to read just one book at a time, but I’m experimenting with reading multiple at a time to see if it leads to more reading hours.

• Personal opsec: I use two-factor authentication for all email and finance apps and avoid SMS whenever possible, preferring authenticator apps instead. I finally added SIM protection to my phone, since SIM swapping is one of the few attacks that can nullify 2FA. If you haven’t done this, search “[your carrier] SIM protection” or just tell ChatGPT your carrier and device and ask how to enable it.

• A suicide prevention project that reaches people at difficult moments through anonymous letters written by volunteers: https://www.reasonstostay.co.uk.

• Someone I know through Farcaster fed Claude the Epstein files and turned it into a 64-episode podcast: www.epsteinfiles.fm. Three days after launching, it had ~12.5k plays and a 2.8k audience size. Interesting times we live in. It’s making me want to experiment with an AI-generated Farcaster podcast, not to replace GM Farcaster (I still think people want human hosts… or at least I need to believe that for now), but as a way to extend the brand and add to the content library.

Some of the pics

I usually include a little photo section here, but apparently I haven’t taken many pictures since my last update. It’s either because it’s been too damn cold, or because I’ve been spending all my free time with Claude Code. Probably both. Here’s one lonely photo of the frozen Hudson River.

A Note to My Readers

I started this substack in December 2022 as an experiment to see if developing a writing habit would help clarify my thinking and/or provide other benefits. It has evolved into different formats over the years, but it currently serves as my public journal.

I use it to share musings, meditations, and links to things I’m finding interesting as I build out a farcaster-native media company, a modern technology consulting company, raise my kids, and have fun creating and learning in the worlds of crypto, tech, finance, science and wellness.

Thank you for supporting my writing and journey. If you’d like to get in touch you can reply to me here or find me on X and farcaster.

Until next time, keep putting good into the world. —adrienne🌏❤️

Today I’m sharing a few things that are on my mind, from yoga as a long-term practice to reflections on power and agency. My annual recap and goals post will come later this month.

Subscribe

All yoga is good yoga.

I first got into yoga in the early 2000s, living in NYC after college. I took classes at New York Sports Club or New York Health & Racquet Club, often with my now husband. Invariably, after each class, we would evaluate it. What did you think? Did you like the teacher?

I could always find something to critique. It was too hot. The music was annoying. There was too much hamstring work. Or too much shoulder. It was boring. I didn’t get a good workout.

Matt had a saying back then that has stuck with me ever since. After I complained, he would say, “All yoga is good yoga”. What he meant was that the class didn’t have to be perfect. I didn’t need to find enlightenment in a single class. Taking an hour to move your body, breathe, and connect mind and body is never a bad use of time.

That mindset has helped me over the years. We should do things we know are good for us without expecting a payoff in the moment.

I’ve had an on-again, off-again yoga practice for 25 years, but the benefits have compounded. It’s probably the single most valuable practice of my adult life.

I was thinking about this old saying, “all yoga is good yoga”, recently and was wondering what other things could fill in for yoga.

All ___ is good ___.

What else applies? Time spent in nature? Time spent learning new skills? Quality time with loved ones?

Share

Third Chakra and Power from Within

Speaking of yoga, I wrote last time about the book Eastern Body, Western Mind and predicted it would be one of those books that changes me. I’m still less than halfway through and can’t recommend it enough.

I’ve finished the chapters on the first two chakras. The first chakra is about creating a solid foundation with grounding and making sure our physical needs and safety are taken care of. The second chakra is about pleasure and sexuality. And I’ve just started the chapter on the third chakra, located at the solar plexus, which all about power.

The author, Anodea Judith, says we are a culture obsessed with power. “One need only pick up the daily newspaper to see that we are a culture obsessed with power. Headlines of violence, warfare, victimization, and dominance reveal a world continually beset by conflict.” It’s interesting to note that this book was written 30 years ago, but I can imagine reading this line at any point in history, including today.

When we hear the word power, we often think of domination and control, in terms of power over someone or something. But just a few pages into this chapter, it’s very clear to me that power has much more to do with agency, rejecting victimhood and reclaiming power of self.

“It is not power we seek so much as the overcoming of victimization—the ability to determine our own lives. For what greater responsibility than to allow the divine within the freedom of its unique unfoldment? … To have true power emanating from within renews the joy of being alive.” - Anodea Judith

Subscribe

Some of the (other) things

Collection of things I’ve read or seen that I’ve enjoyed.

• Artist BrettDrawsStuff wrote a short story about a guy who goes to urgent care for a sore throat. It’s a nice break from the doom and gloom which tends to permeate throughout the internet. A Frog in Your Throat.

• Farcaster CTO Varun Srinivasan wrote about the problem with passkeys for securing crypto wallets that want a consumer grade UX.

• “Your personality is where your premium is” - I liked this quote from current Cornerstone CEO about what matters in the workforce when anyone can use AI to write code.

• Watched Pluribus, Death by Lightning, and Aziz’s new movie Good Fortune; finished reading 1929; started reading Chess for Dummies in an attempt to be more competitive when I play with my son.

Some of the pics

Florida Gulf coast, the Everglades, Naples Botanical Garden and what my teenage daughter makes for dinner when she’s home alone - a kale salad with roasted chickpeas, sweet potatoes, avocado and lemon tahini dressing

Note: if you’re reading this newsletter on Paragraph or another RSS syndicate, you might only be seeing 1 image instead of a gallery of 6. Substack publication will have all the photos if you are interested in seeing them

A Note to My Readers

I started this substack in December 2022 as an experiment to see if developing a writing habit would help clarify my thinking and/or provide other benefits. You can read about my original intentions in my first post or my more recent reflections after sticking with it for a year.

I write about twice a month and share musings, meditations, and links to things I’m finding interesting as I build out a farcaster-native media company, a modern technology consulting company, raise my kids, and have fun creating and learning in the worlds of crypto, tech, finance, science and wellness.

Thank you for supporting my writing and journey. If you’d like to get in touch you can reply to me here or find me on X and farcaster.

Until next time, keep putting good into the world. —adrienne🌏❤️

Leave a comment

I started reading Eastern Body, Western Mind, a book my yoga teacher recommended that blends the chakra system from ancient India with modern Western psychology. I’m still only ten percent in, but in true Adrienne fashion, I’ve already started recommending it to others. I can tell it’s going to be one of those books that changes how I see things, especially as someone interested in the relationship between mind, body, and soul and wanting to develop a healthy nervous system.

One line from the book has stuck with me all week:

So many of life’s problems stem from too much or too little of something. We spend our lives searching for balance. — Anodea Judith

Finding balance is an impossible goal. Every action you take to restore balance knocks something else off, and you have to try again. Think of balance as a north star; it can guide you, but you can never reach it.

I felt this recently. Regular readers know I used to write twice a month, but this is my first post in two months. I stepped away from writing to make room for other activities. I’ve been coding more. Both are creative acts, and for these past few months, coding has fulfilled me in the same way writing had. But eventually, I felt out of balance and needed to come back to my words.

The idea of balance came up recently during a Jeweled Horse Foundation board meeting.

Jeweled Horse is a nonprofit newly formed in 2024 to support wellbeing and cultural preservation in Mongolia. I serve on the board, and the foundation is the vision of Janis Chilton, a fellow Kalmyk Mongolian. From the start, the plan was to balance large grants and major donors with smaller, community contributions.

Unfortunately, that balance hasn’t materialized the way we expected. Recent changes in the political landscape eliminated several nonprofit funding sources, with ripple effects across the broader grant ecosystem that have made funding harder to secure. So far, the foundation has been funded almost entirely by small donations, mostly from family and friends.

And yet, even to my pleasant surprise, we’ve been able to make tangible impact with small donations alone.

Because we partner with existing organizations like the Diluv Hutugthu Foundation which are already doing meaningful work on the ground in Mongolia, even our modest funding has translated into real impact. Small donations have supported initiatives like a Zero TB for Kids program, launched in partnership with Johns Hopkins and local health authorities, alongside compassion training for educators and public service workers, social and emotional learning programs, and winter emergency preparedness.

Preserving Mongolian culture matters to me personally, but it matters even more broadly to the world. Mongolia is unlike anywhere else. As technology accelerates, and as AI threatens to flatten the world into a single monoculture, we risk losing ancient wisdom that doesn’t easily digitize.

I’m deeply pro-technology and pro-humanity, and I don’t believe in choosing one over the other.

I believe we need balance between seemingly opposing forces: compassion alongside innovation, communities that cross borders alongside preservation and pride in where we come from. Balance is not a fixed state; it’s a practice we return to so we don’t stray too far in any one direction.

So, I have two simple asks.

1. First, if you’re able, consider making a donation. We currently have a generous donor matching up to $3,000, and contributions as small as $20 genuinely help. Donations are tax deductible and all donations will be matched and go twice as far through Jan 6. Donate here. (If you do, please let me know so I can shout you out in my next post!)

2. Second, we’re looking for help securing a stable funding source of roughly $50k per year to help the foundation get off the ground. If you’re able to help, that could look like:
   – introductions to aligned donors or high-net-worth individuals interested in supporting compassion-focused work, Buddhist-inspired initiatives, or cultural preservation in Mongolia
   – information about relevant grants or foundations we should be paying attention to
   – or guidance navigating the nonprofit grant-writing world, including help with writing or strategy (simply reply to this email if you can offer any help)

Thanks for reading, and for supporting the Jeweled Horse Foundation in whatever way feels right to you. I’m grateful to everyone who’s supported Jeweled Horse in these earliest days, and to those who take the time to learn about the work we’re doing. Please consider following JHF on Facebook, Instagram and LinkedIn.

Some of the (other) things

As always, a quick roundup of some of the other things I’ve been getting into:

• Attended DevConnect, the first Ethereum World’s Fair in Buenos Aires, where I volunteered at the Farcaster booth, recorded a podcast with Will from Splits, Debbie from Privy, and Limone from Builders Garden and FarCon Rome, attended the base BUILDher brunch, ate lots of steak, and made time for an afternoon of sightseeing.

• Read the book Culpability by Bruce Holsinger which explores AI ethics and morality. As we give more control to technology, are we losing our sense of personal accountability? Who is responsible when a self-driving car crashes and kills someone? When a defense drone kills a civilian? When pre-teens get negatively influenced by chat bots?

• Hosted Thanksgiving this year and we added a few new items to our traditional menu: baked mac and cheese as requested (and made) by my 14 year old son, chimichurri (because our summer vegetable garden happened to have a ton of green parsley so why not? turns out it goes great with turkey), and a cranberry pico de gallo (a huge crowd favorite - simple recipe with raw cranberry, onion, celery, sugar, lime, jalapeno, cilantro, salt). We also got our turkey from a Catskills, NY farm, and it was delicious after a 30-hour brine.

• “It’s cheaper to cover a hole in the wall with a flat screen TV than fill it. Stuff is cheap, services are expensive. AI is about to fix that.” I got a kick out of the subtitle in this recent Fintech Brainfood newsletter.

• After thanksgiving I decided I would no longer work on any new projects for the remainder of the year and only allow myself to do refactoring and maintenance. I surprised myself how much I’ve gotten done already - so many things on my list that just weren’t a priority but I’m glad to have knocked them all out! I moved from local disk storage to S3 /redis/shared cache for the lore project; added natural language search to the gmfarcaster website (was easy to do using existing components already created); integrated full transcript files and featured casts into our website, ran some security updates, and changed S3 buckets to be versioned.

Some of the pics

Top row: Saw Foreigner at the Capital Theater; Cranberry Salsa, a street in Buenos Aires where trees, cars and buildings are in balance (there’s that word again)

Bottom row: devconnect main entrance; dog walk in the autum; and playing in the snow

Note: if you’re reading this newsletter on Paragraph or another RSS syndicate, you might only be seeing 1 image instead of a gallery of 6. Substack publication will have all the photos if you are interested in seeing them

A Note to My Readers

I started this substack in December 2022 as an experiment to see if developing a writing habit would help clarify my thinking and/or provide other benefits. You can read about my original intentions in my first post or my more recent reflections after sticking with it for a year.

I write about twice a month and share musings, meditations, and links to things I’m finding interesting as I build out a farcaster-native media company, a modern technology consulting company, raise my kids, and have fun creating and learning in the worlds of crypto, tech, finance, science and wellness.

Thank you for supporting my writing and journey. If you’d like to get in touch you can reply to me here or find me on X and farcaster.

Until next time, keep putting good into the world. —adrienne🌏❤️

When I was in my early twenties, I came across a short passage with some life advice. I have zero recollection of where it came from or who wrote it. All I know is I saved it by emailing it to myself, and over the years I’ve reread it and shared it with friends more times than I can count.

Here it is:

Don’t concern yourself with being better than everyone else. That will mainly get you resentment, disappointment, or both. Focus instead on being better today than you were yesterday, on being even better tomorrow than you are today.

When you compare yourself to others, you’ll either end up being disappointed or you’ll end up being arrogant. Instead, compare yourself to yourself. The important thing is to make progress in your own life. Upstaging someone else simply won’t bring you any true gain. Improving yourself will.

We all share certain things in common, and yet no two people are the same. There is no one who is better than you. And at the same time, you are better than no one else. You are unique. You are beyond compare. Make the most of it. Make it better and better each day. Stop worrying about everyone else and you’ll free up enormous energy to be the best you can be.

— Author Unknown

Growing up, I was a pretty happy, confident kid. I didn’t really get jealous of other people who had more than me, whether it was money, popularity, achievement, whatever. But I did measure myself against other people, and I took some comfort in knowing I was “doing better” than others.

But when I read the line in the passage about arrogance, it kind of stopped me in my tracks.

The reason you don’t compare yourself to others is not only to protect yourself from feeling unworthy or badly, but it’s also about not letting yourself become someone you don’t want to be.

In other words, don’t compare yourself into misery, and don’t compare yourself into being an asshole either.

I internalized this advice more than any other life advice I can remember and I’ve tried to live by two simple rules:

• Focus inward

• Focus on improvement

It’s great to be competitive. Just compete with yourself and try to be better today than yesterday.

This passage has lived in my email drafts for two decades. When I catch myself comparing myself to others, feeling like I haven’t accomplished enough, or even getting a little too proud of what I have, I come back to it and reread it. More than any book or mentor or piece of advice, this short passage has helped me the most, not just in terms of personal and professional growth, but in staying grounded and mentally balanced.

I’ve been meaning to share this publicly, in case it helps someone else the way it has helped me. An occupational hazard of building a business on top of a social network is that I end up spending a lot of time in social feeds. Social media (especially crypto social media) is ground zero for comparison.

Who has the most followers?
Who got the airdrop?
Who’s on the leaderboard?
Who has the biggest wallet balance?
Who made the most money on a trade?

It’s interesting and often fun. But it’s also a great reminder that the only comparison that actually matters is you vs. you.

Subscribe

Related Reading:

• Status as a Service by Eugene Wei, “humans are status seeking monkeys”

• Mindset by Carol Dweck, which introduces the concept of fixed vs. growth mindset

A Note to My Readers

I started this substack in December 2022 as an experiment to see if developing a writing habit would help clarify my thinking and/or provide other benefits. You can read about my original intentions in my first post or my more recent reflections after sticking with it for a year.

I write about twice a month and share musings, meditations, and links to things I’m finding interesting as I build out a farcaster-native media company, a modern technology consulting company, raise my kids, and have fun creating and learning in the worlds of crypto, tech, finance, science and wellness.

Thank you for supporting my writing and journey. If you’d like to get in touch you can reply to me here or find me on X and farcaster.

Until next time, keep putting good into the world. —adrienne🌏❤️

Leave a comment

Subscribe

Engineer-y, Serious, and Warm: Notes from Farcaster Dev Day

I attended Farcaster Dev Day in NYC this past Friday. About 100 developers building on Farcaster were invited to come together for face-to-face networking and collaboration, listen to an update from founders Dan Romero and Varun Srinivasan, attend a Q&A, and meet the Merkle team.

It was an awesome day. The vibes: engineer-y, serious and warm.

I shared my detailed feedback on Farcaster, but I’ll also summarize my notes here.

Temperature Check

I walked out of Dev Day feeling just as confident in Dan & Varun’s judgment and leadership as I have every time I’ve heard them speak since first learning about Farcaster 3 years ago.

Their presentation centered on the current product strategy of Discover. Trade. Create. which leans heavily into crypto, a bit of contrast to past messaging which was far more social oriented (e.g. Farcaster is a Social Network not Social Media, URL to IRL, make friends and make money). What stood out to me wasn’t the strategy itself but the consistency in messaging.

There’s an underlying message Dan has repeated in every single presentation I’ve heard him deliver and it’s all about growth:

• If you don’t solve for growth, nothing else matters.

• Growth must happen in orders of magnitude.

• Keep experimenting, keep iterating, keep learning.

People invariably get upset whenever priorities shift (devs should build alt clients! no, frames and miniapps! we’re onboarding journalists and creators! channels is the key! investing in vertical video! crypto everywhere and everything!), but if you pay attention to the underlying message that growth is the most important factor, it’s hard to stay upset.

As for the current direction toward crypto and the trader experience: it’s great! I’m not a trader. I’m not a gambler. As an independent developer, I’d love to see better dev experience, mature tooling, better docs, and a DevRel team. That might help me the most. But I can’t make the argument it would drive growth.

And their logic for focusing on what will drive growth is sound. Early signs indicate it’s the right move. So, let’s go for it!

The Best Part of Dev Day: People

Honestly, I just really like being around the Farcaster dev community. Being together was a good reminder of why I build here. The community is filled with people who are smart, technical, and kind.

I loved meeting members of the Merkle team face to face—Horsefacts, Deodad, Shane, Linda, Daniel—all first-class humans. And shoutout to Grin and Maxp, who helped me unblock an issue with my AI Bot that I’d been ignoring because I didn’t have a good solution to. Nothing beats in-person time: a few minutes of hallway talk saves me days of research and debugging.

And finally, huge kudos to Linda and Emma for organizing the day. Logistics were smooth as butter.

I left Dev Day feeling confident in Farcaster’s future, and excited to keep building miniapps, bots and livestream news shows on top of Farcaster.

If you’d like to try Farcaster, you can create an account using my referral code here. And feel free to reach out if you want help onboarding.

Note: If you’re reading this via my syndicated Paragraph publication, you may only see one photo above instead of a photo gallery. To view the full set of photos, refer to my Substack newsletter.

Notes from the SoGal Annual Meeting

I attended the SoGal Ventures Annual General Meeting and Summit last week. Wow do these women know how to put on an event! The location was a cool soho loft, decorated in style, with yummy specialty snacks and drinks, 20-minute founder talks, a fun two-minute pitch competition, dinner and drinks with DJ, and an epic swag bag.

I’m really impressed with GPs Pocket and Elizabeth, not only through a financial ROI lens, but for building such a passionate and large community. SoGal is led by all-women GPs, backed by 80% women LPs, and invests in women founders across industries shaping a better world.

Many conversations started with, “So, why are you here?” For me, it’s because I’ve spent my entire career in male-dominated spaces, and I believe investing is one of the most powerful ways to help shape the future you want to see. I want a future with more women in tech, more women in leadership, more women in finance, and more women owning and running their own businesses. When I wanted to learn more about venture investing in early 2022, SoGal was the first investment I made as an LP, and I’m really glad I did.

Same Note As Above: If you’re reading this via my syndicated Paragraph publication, you may only see one photo above instead of a photo gallery. Refer to my Substack to see the full gallery.

Some of the (other) things

Here’s a quick rundown of some of the other things I’ve been reading, watching, and doing.

• READ Careless People by Sarah Wynn Williams, who worked on public policy at Facebook and went from being optimistic about its potential for good to deeply disappointed by internal politics, content moderation challenges, and how the company engaged with governments. It’s an easy/quick read and I liked the firsthand account of what it was like working there in those earlier years, especially the tension between Facebook’s “growth at all costs” and “engineering mindset” and the nuances of social media policy. Another theme I found interesting was the tension of her being a mom in a leadership position, working closely with Sheryl Sandberg during the Lean In era. Her belief was she could influence Facebook “from the inside” but you can imagine how that ends up. Overall, an interesting book but her conclusion felt a bit lazy; she sums her entire experience up as “it didn’t have to be this way” without giving any serious vision for an alternative way.

• READING The Hobbit. I wanted fiction I could get lost in and somehow had never read Tolkien before. Enjoying it so far!

• READING 1929, Andrew Ross Sorkin’s new book about the people and events leading up to the Great Depression.

• WATCHED Task, on HBO. The plot isn’t entirely believable, but as long as I kept reminding myself it is a series about characters and not plot, I came to really love it. Definitely recommend. It was by the same creator as Mare of Easttown.

• WATCHED The Network State Conference. 9 hours, watched over a few days. Really good. A lot of optimism for the future.

• LISTENED to Brian Eno on The Ezra Klein Show. A beautiful conversation about art, humanity, modern life. And yes, we got a scenius shoutout.

• TRAVELED to Ann Arbor for a quick tour of UM with my middle daughter who is in the middle of college applications.

• ATTENDED the Aptos Experience in Brooklyn.

• PLAYING with Sora, Open AI’s new video creation app. It’s fun. While the output isn’t so great yet, I like experimenting with new tools. My favorite output so far is me as an Eagle Huntress.

A Note to My Readers

I started this substack in December 2022 as an experiment to see if developing a writing habit would help clarify my thinking and/or provide other benefits. You can read about my original intentions in my first post or my more recent reflections after sticking with it for a year.

I write about twice a month and share musings, meditations, and links to things I’m finding interesting as I build out a farcaster-native media company, a modern technology consulting company, raise my kids, and have fun creating and learning in the worlds of crypto, tech, finance, science and wellness.

Thank you for supporting my writing and journey. If you’d like to get in touch you can reply to me here or find me on X and farcaster.

Until next time, keep putting good into the world. —adrienne🌏❤️

I want to take a quick minute to welcome new readers to my substack (or paragraph newsletter if you are reading there).

If you’re new, welcome! I’m happy you’re here.

I don’t have big plans for this substack besides it being a fun experiment in writing in public.

The inspiration for the title Some of the Things came from my desire to start a writing habit and recognizing that if I tried to write about All the Things I was getting into or thinking about, I’d never succeed.

Some of the Things has evolved over the years into my public journal. Some posts are more reflective and prose heavy, and some tend to be simple bullet point lists of things I’m getting into.

It’s been a long 3 months since I wrote last so I won’t try to document everything I got into this summer, but I'll hit some of the highlights.

Coding Updates

From Standalone Web Apps to a Single Monorepo

One of the first projects I worked on this summer was converting the GM Farcaster web projects from 3 separate standalone repos into a single, consolidated monorepo. This type of behind-the-scenes work is some of my favorite — invisible to the end user but improves the developer experience and makes it easier to manage and ship code in the future.

Apparently you can take the girl out of a devops role, but you can’t take the devops out of the girl.

For example, I had 3 web projects for GM Farcaster: our main website, a media kit, and a miniapp. There was duplicated code across the 3 (sponsor lists, testimonials, our latest episodes, etc.) and having to manage it across 3 separate GitHub repos made development cumbersome. After merging them into a single monorepo, it has made making changes much easier to manage.

Podcast Admin Dashboard, S3 Refactoring, and OAuth

The second big project I worked on this summer, also invisible to end users, was an admin dashboard for GM Farcaster. I had a set of internal tools I’d built for managing the podcast, but they ran from my local PC, which meant only I could access them. For example, I have a tool that sends a push notification to all our miniapp users when we go live. But only I could trigger it, and I wanted to give my cohost NounishProf the ability to use it too, especially ahead of my travels this fall.

I loved building this because it gave me hands-on experience with libraries I hadn’t used before, like Google OAuth for authentication. I first tried using SIWF, since I prefer to use Farcaster-native tools whenever possible, but I struggled to get it working. The flipside of using emerging tech (aka leading/bleeding edge) is there is less documentation, fewer examples, and harder for LLMs to help. After struggling with SIWF for a few days, I got Google sign-in working in a day.

I also had to retrofit the “database” our miniapp was using so the new admin dashboard could manage backend data. Previously, I just had a few local JSON files in the codebase acting as the database, which meant updating data required a git commit and a code deploy. To make the data easier to manage, I refactored the storage to use S3, decoupling data from the code. Updates to data can now be managed outside of code deployment. I’m no longer the bottleneck either since we can update everything using the UI in the tools dashboard now.

One Click Mint in a MiniApp

One smaller thing I worked on recently was adding the ability for users to mint podcast episodes directly from our miniapp. Since the start of Season 3, we’ve been posting all our episodes to Pods.media, which means anyone can mint or collect an episode as an NFT. One of the Pods founders reached out to ask if I could add a link to Pods from our miniapp, but instead of kicking users out of our app, I decided to call the Pods smart contract directly from within our app.

Calling a smart contract to mint an NFT is one of the most basic web3 actions. And yet I can’t get over how wildly excited I was to do it.

I come from the enterprise world, where up to 80% of my time was spent on coordination. Coordination is expensive! I’m used to scheduling meetings to decide on how apps should integrate. I’m used to having to register and get permission to use APIs. I’ve been resigned to accepting the risk that eventually someone else’s API will change or stop working.

But when you call a smart contract - even one someone else wrote - you have a guarantee it will never change. You don’t need permission. You just use it.

MAGICAL! And you can’t convince me it’s not.✨✨

Lore Finder App

Our live stream Farcaster news show is centered around casts. We curate the most important casts of the past few days and share them on screen and talk about them. We link to the casts we talked about in our show notes. And there they sit. Technically available but virtually unusable to answer questions like:

• “Was I ever mentioned on GM Farcaster?”

• “How many times was I mentioned on GM Farcaster?”

• “Who is most often mentioned on GM Farcaster?”

• “What casts were popular back in Feb of 2024?”

So I decided to build an app to make it easy to scroll back through casts we’ve featured on the show.

I first built a utility that scrapes all our show notes and parses out links to casts. I then used Neynar API to enrich the URL and fill in the cast author and cast date. And stored it all in JSON files.

Now that I have casts in a queryable format, I built a miniapp front end to let people scroll through all the bangers we’ve featured, complete with search and a leaderboard.

I also created a page so any user can see how many of their casts we’ve pulled, and I implemented a dynamic share image to take advantage of social virality to encourage users to share their accomplishment and draw more people to the app.

You can see the app at lore.gmfarcaster.com but you only get full features when opening as a miniapp in Farcaster or The Base App.

Oh, I also open sourced the code: https://github.com/atenger/farcaster-lore-miniapp

Refactoring Email Storage for the Bat Book

The last coding project I worked on this summer was a bit of refactoring for the website for the children’s book I’m working on with my sister.

The bat book was the first web site I vibe coded and when I had originally created it, I used Supabase (managed PostgreSQL) to hold email addresses for people who want to be notified. I found it really easy to configure and set up.

The downside ended up being limitations of its free tier.

My website gets virtually no traffic. Without usage, Supabase kept shutting down my database to save resources. For a while I kept manually spinning it back up, but it was starting to get annoying. I could have paid $25/month to keep it running, but that’s massive overkill for a database with a single emails table which is expected to grow only into the hundreds, thousands if I’m lucky. So, I refactored the website to use S3 as its backend to store emails and now I’ll pay pennies to AWS each month.

Much better.

Coding Roadmap

• Build out an episode library on our website and create an episode details page that has each episode, the full transcript, GPT summaries and more.

• Combine various miniapps into a single one for better UX. Add the lore finder to the main gmfarcaster miniapp. Create an API endpoint for our AI search through transcripts so people can search through our content using web interface and not just the bot.

• Automate the lore scraper and transcript creation scripts and deploy somewhere so that I can trigger them from our tools dashboard and not need to run them locally.

Some of the things I did this summer

• Saw comedy! I’ve been on a standup comedy kick. So far this year, I’ve seen Whitney Cummings, Nikki Glaser, and Mark Normand doing shows. Then we happened to be in Halifax when the Great Outdoors Comedy Festival was there so we snagged tickets and saw a phenomenal show with Fred Armisen, Mike Birbiglia, Nick Kroll and John Mulaney.

• Traveled a bit. Went on a road trip through PA, DC and VA and visited colleges - Penn, Georgetown and UVA. Also traveled to Cape Breton Island and Halifax in Nova Scotia.

• Tried to read. I started a lot of books but kept putting them down after a few chapters. The two I managed to finish are Genghis Khan and the Making of the Modern World by Jack Weatherford and The Music of Bees by Eileen Garvin. Have you read a book recently that you just absolutely loved or changed how you see the world? Send recommendations!

Some links to interesting things

• A website to translate from English to Dolphin and the source code for the translation. e EEEE EE EEE EE EEE EEE eee eEeE eee eee EeEE eEeeEE eEee eee EEe EEEE Ee EEEe E e eee e EeE eEee EE e eee EEe e

• Loved this Ted talk from the author of Free Range Kids, where she encourages people to spend less time with our kids.

• I learned more about the history of the Kalmyks (1/2 my ethnicity) and how they were exiled to Siberia in the 1940s and then returned home to Kalmykia in the 50s, after Stalin’s rule. Many sad stories during this time, and the current poverty rate is still high today. It was during this time when my grandfather escaped and fled west to Europe, eventually making his way to the US. Makes me think of sliding doors and how much harder life was just 2 generations ago.

• Anyone need a reminder to separate self worth from your business?

• “We help small companies act big and big companies act small” is one of my favorite ways to describe what Tenger Ways is all about.

• New book from IT Revolution about Technological Jerk. They are referring to the physics term jerk which happens when the rate of acceleration changes and is a jarring, discomforting experience.

Some of the Pics

• Row 1: Chimney Corner beach on Cape Breton Island, Table Doucet

• Row 2: Pleasantville Farmers Market x 2, Halifax Harbour Kayaking, Great Outdoors outdoor comedy festival

• Row 3: Old rowboat near Peggys Lighthouse, goats eating a tree in Rockefeller State Park, our lodging in Cape Breton featured wood burning hot tubs and epic sunsets

A Note to My Readers

I started this substack in December 2022 as an experiment to see if developing a writing habit would help clarify my thinking and/or provide other benefits. You can read about my original intentions in my first post or my more recent reflections after sticking with it for a year.

I write about twice a month and share musings, meditations, and links to things I’m finding interesting as I build out a farcaster-native media company, a modern technology consulting company, raise my kids, and have fun creating and learning in the worlds of crypto, tech, finance, science and wellness.

Thank you for supporting my writing and journey. If you’d like to get in touch you can reply to me here or find me on X and farcaster.

Until next time, keep putting good into the world. —adrienne🌏❤️

As cohosts of Farcaster's longest running and number one podcast, GM Farcaster, Nounish Prof and I have been at the forefront of this shift. We’ve built a Farcaster-native media brand from the ground up, with live shows, podcasts, newsletters, community events, and an AI assistant—all integrated with the Farcaster protocol.

This playbook is open-source and based on everything we’ve learned since launching our media business on Farcaster in 2023. Farcaster has been an incredible place to grow both audience and community, and we’d love to see more media brands lean into this growing ecosystem. We’ve found success with tools like NFT mints, AI bots, and MiniApps, and we hope this guide helps you try them too.

→ Feel free to use this playbook, remix it, or reach out if we can help.

📝 The Playbook

The four primary building blocks we’ve found effective for establishing a media presence on Farcaster are channels, group chats, mini apps, and AI bots. Together, they form a lightweight but powerful foundation: a place to gather and connect with your community, a way to showcase your content inside the feed, and a smart assistant to help people explore everything you’ve made.

1. A Farcaster Channel

Think of a channel like a subreddit or hashtag you control. Channels create a dedicated space in Farcaster where you can post your updates, where your community can find you and get some basic information, and where anyone can indicate they're talking about your brand.

Having a channel in Farcaster is like putting up a shingle in the real world. A channel is dedicated real estate for your brand.

For GM Farcaster, we used our channel as the central hub for our show, the way traditional companies use websites. In fact, we used our channel as our "home page" long before we even had a website.

In this screenshot of our /gmfarcaster channel, you can see a few features of channels:

• Name & Description - Establish your brand presence

• Follower counts - People can follow channels the way they follow accounts. We have 17k followers for the GM Farcaster channel.

• External links - You can easily link to your podcast or newsletter or website for people to learn more

• Membership - You can set up membership to limit who can post to your channel

2. A Private Group Chat

Today’s media companies understand that a smaller but active community can be more valuable than a larger but passive audience. Farcaster supports private group chats - dedicated spaces where your most engaged followers can connect directly with you and with each other.

We use our GM Farcaster group chat to share previews of what we’re working on and to gather early feedback from our most loyal listeners.

3. A Custom MiniApp

MiniApps (formerly known as Farcaster Frames) are lightweight web apps that run directly inside the Farcaster feed, on both mobile and web. They feel like native apps but are just regular websites, so you can launch quickly without worrying about app store approvals.

Because MiniApps live inside the social feed, usage is almost always higher than standalone apps. Users can interact with your content without leaving Farcaster, and when they engage (watching, collecting, commenting) it happens in public, where others can see and join in. That’s where network effects kick in.

Every media company should consider launching a MiniApp. Why? Because the hardest part of running a show or newsletter is getting people to come back. Subscriptions help, but we’ve all seen what happens when Gmail tosses your email into the “Promotions” tab. A MiniApp gives you a direct line to your audience.

When someone saves your MiniApp to their favorites, you can send them push notifications. That means you can tell fans when you’ve released a new episode or published a newsletter, without relying on third-party algorithms or email filters. (But of course, be careful and don't abuse it, or people will disable alerts or remove your MiniApp.)

We built a MiniApp for GM Farcaster so users could watch our live stream without leaving Farcaster. When someone is scrolling the feed, they don’t want to jump to YouTube or Spotify. Now, every time we go live, we send a push notification, and fans can watch the show right inside of Farcaster.

We also added a “Cast that you're watching” share button to the MiniApp that encourages users to cast about our show to their followers, further amplifying the episode to others on Farcaster and drives more views, more discussion, and more community engagement.

4. An AI Assistant

Farcaster is a social network, which means everything happens in the feed. Users post casts, reply to others, and all of it is public and searchable.

So what if your brand could be part of that conversation—literally?

That’s where an AI Assistant comes in. Instead of thinking of it like a chatbot hidden on a website, think of it as a Farcaster user in its own right: replying to posts, answering questions, and surfacing content from your back catalog at the right moment.

For GM Farcaster, we built an AI bot called GMFC101. It looks like any other Farcaster user, but behind the scenes, it has access to transcripts from 300+ episodes of our show. It can help new users learn how to use Farcaster or point longtime fans to past discussions on specific topics. It lives in the feed, participates in real conversations, and is always on-brand.

It’s also open source. You can find the code here: github.com/atenger/Farcaster-101-API (Apache-2.0 licensed), and I wrote a detailed walkthrough of how I built it: How I Built GMFC101.

Why do you need an AI Assistant?

If you produce podcasts or video content, you’ve probably run into the same issue we did. Audio and video aren’t inherently searchable like text, so all that valuable content becomes hard to access. YouTube search only gets you so far as it relies on your titles, descriptions, or chapters. But how does someone search inside one of your conversations?

That’s what an AI Assistant solves. It unlocks your full catalog and makes it interactive, discoverable, and shareable—right where your community already is.

🧰 Tools and Tech Stack

Beyond the core playbook items above, Farcaster is made of legos, and that’s where the fun begins. These additional tools can be used to increase your brand presence and build deeper connections with your audience. Using crypto-native tools signals that your brand is aligned with the values of the Farcaster ecosystem: openness, experimentation, and community-driven growth.

For Distribution & Engagement

Paragraph.xyz - Syndicate newsletters directly to Farcaster
Users can read your posts in the Paragraph miniapp without leaving the feed, and subscribers get push notifications when you publish. It’s like Substack, but crypto-native.

Ponder - Create native polls inside the Farcaster feed
Ask your audience things like “Which topic should we cover next?” or “What’s your favorite episode so far?” Polls are a great way to boost interaction and crowdsource ideas in public.

Buoy - Real time keyword alerts
Set up real-time alerts whenever someone mentions your brand on Farcaster. It's a powerful social listening tool to track keyword mentions and stay in tune with the conversation about your content.

For Onchain Presence and Credibility

Pods.media - Host Your Podcast Onchain
With Pods, you can publish your podcast episodes onchain, letting users listen, mint, and collect them as NFTs. It creates a permanent record of your work and lets your audience own a piece of it. We’ve been using Pods since Season 3 of GM Farcaster to make our episodes more discoverable and collectible in the crypto-native world.

Zora, Highlight, Rodeo - Mint Media That Matters
These platforms let you mint full episodes, clips, artwork, newsletters, or one-off moments and share them on Farcaster. At GM Farcaster, we’ve experimented with all of them - from minting sponsor receipts to posting clips. Each platform offers different mechanics, and trying them helps you understand how your audience wants to engage.

Clanker - Launch a Meme Coin
Clanker lets anyone create a meme coin in seconds using nothing more than a cast. It’s a Farcaster-native tool that turns community memes into shareable, tradable assets. We launched $CastOut using Clanker during our CastOut competition. It’s playful, sticky, and a great way to gamify participation.

Hyersub - Build Memberships Onchain
Hypersub is a protocol for onchain memberships, similar to Patreon or NPR’s supporter model. You can offer perks like bonus content, early access, or shoutouts to fans who subscribe. We've had success with our GM Farcaster hypersub which allows people to directly support us and take advantage of membership benefits and perks.

💜 If It Worked For Us, It Can Work For You

Farcaster is still early. Farcaster users are curious, engaged, playful, and global. They don’t just consume media passively, they build with you, remix your memes, show up live, and give feedback that actually helps.

We’ve seen this firsthand at GM Farcaster:

• Our community consistently joins live episodes and participates in the chat.

• Livestream viewership jumped ~25% after we launched our MiniApp.

• Our AI assistant has become a go-to for onboarding and content discovery.

• Our Hypersub members actively help us shape the show.

If you’re looking to build not just an audience, but a community, Farcaster offers the most direct, flexible, and creator-friendly rails on the internet today.

💬 Want Help Getting Started?

We pioneered this model with GM Farcaster, and would love to see more media brands take the leap and join us on this sufficiently decentralized social network.

Now’s the time to get in early, grow with the network, and build a trusted presence while attention is still easy to earn. As Farcaster expands, early media brands will benefit from credibility, community, and momentum that’s hard to replicate later.

If you want to:

• Launch a channel

• Build a miniapp

• Set up a meme coin or AI bot

• Mint media onchain

• Or just understand what’s possible...

Start experimenting with the four building blocks in this playbook, try the tools we recommended, and let us know how it goes!

Or if you'd like some help getting started, reach out and let’s talk

But here we are in the first week of June and summer weather has finally arrived.

I know this not just because it’s been in the high 80s all week, but because I had trouble falling asleep last night. I was too busy scratching my legs. I counted at least 10 mosquito bites. Ugh. If only we could have summer without mosquitos.

But the flip side of summer in New York is seeing new life everywhere:

• Two baby raccoons crossed the street in front of my car last night

• A baby bunny was sitting on the side of my running path this morning

• I’ve been waking up to birdsong coming in through the open window

• Flowers are abloom everywhere I look

September is a good month for getting back to hard work. January is good for setting intentions and goals.

But there’s no better time for birthing creativity into the world than summer.

I always notice I have more manic energy this time of year, an increase in desire to start new things. Maybe it’s the weather, the longer days, or all the new life all around. Does anyone else feel this rhythm during summer too?

Instead of writing about things I’ve done lately, I’m making a list of things I want to work on:

• Add features to the GM Farcaster miniapp: haptics, more onchain capabilities, and a “chat with the AI bot” experience

• Build a utility that scrapes our show notes for every cast we’ve ever mentioned and adds them to a searchable database

• Auto-generate video clips when the bot recommends an episode, so it shares a clip instead of just linking to YouTube

• Create admin tools for processing shows—transcripts, cast metadata, and selecting which episode to spotlight in the miniapp

• Publish the children’s book I’ve been working on with my sister (her hand-painted illustrations should be done in the next month or so!)

• Start writing The Joy of Coding. I’ve already collected one story from an engineer, and I’m hoping for more

Let me know if you’re feeling the summer energy too or if you’ve got a project you’re hatching this season. I’d love to hear what’s on your list.

Some of the pics

Top to bottom, left to right: S’mores, beets and herbs, a sign for a postcard show in NYC, peony bloom, chamomile bloom, hammock, cows at pasture

A Note to My Readers

I started this substack in December 2022 as an experiment to see if developing a writing habit would help clarify my thinking and/or provide other benefits. You can read about my original intentions in my first post or my more recent reflections after sticking with it for a year.

I write about twice a month and share musings, meditations, and links to things I’m finding interesting as I build out a farcaster-native media company, a modern technology consulting company, raise my kids, and have fun creating and learning in the worlds of crypto, tech, finance, science and wellness.

Thank you for supporting my writing and journey. If you’d like to get in touch you can reply to me here or find me on X and farcaster.

Until next time, keep putting good into the world. —adrienne🌏❤️

Leave a comment

Apologies in advance for making this one such a long one. It’s been over 6 weeks since I last wrote so I had a backlog of topics I had to write about - including an idea for a book I want to write, notes from our family trip to London, and the usual roundup of short updates.

Oh, and new this week, all of my substack posts (SomeOfTheThings.substack.com) will be auto-published to my paragraph publication (https://paragraph.com/@adrienne), so you might be reading this on substack or paragraph.

Enjoy.

If you were to write a book, what would it be about?

The book I want to write is The Joy of Coding.

I don’t think most people think of joy when they picture computer code. Ask a room full of engineers and you’re just as likely to hear about frustration, toil, tedium, bugs, and middle of the night deployments as you are about delight.

But when I look back on the 25 years since I learned to code, the one word that keeps coming to mind is joy.

On my podcast, GM Farcaster, we end interviews with a lightning round of rapid-fire questions. One of my favorites is: If you were to write a book, what would it be about?

Someone recently flipped the question back on me and I didn’t have an answer. But it must have planted a seed because a few weeks ago, in the middle of a yoga class, a thought popped into my head out of nowhere and it was loud and incessant:

I should write a book about the joy of coding.

First Spark

My first memory of coding was a course I took in college. It was a required computer science class for my Environmental Systems Engineering major. According to my transcript, it was called CMPSC 203: Bus Pgm Aplcn. I believe we spent the entire semester learning Turbo Pascal. Using a DOS-based terminal, we set variables, used conditionals, and wrote loops to print sequential numbers 1 to 10 on a screen. Very, very basic stuff. I am certain I didn’t feel “joy” during this class, yet why is it that almost 30 years later I still remember the feeling of excitement of watching my code compile and seeing the expected results print out on a screen.

Continuing to Code

My first job out of college was at PwC as a technical consultant. I went through a training program where we learned to build client-server applications using Java and Oracle, and web programs using PHP. It didn’t feel like work. It felt like playing a video game or solving a puzzle.

Then came Cityfeet.com, where I picked up web development—HTML, CSS, VB6, ASP, SQL Server. We used tables for page layouts and CSS red borders for debugging. We deployed changes over FTP. Everything was manual - writing code, testing changes, deploying to live, rolling back when something didn’t work as expected. I remember pressing F5 to refresh the live website, holding my breath each time. Did I break it? Did it work? A sigh of relief when it worked out and I didn’t have to spend the evening trying to fix a mess.

Next was Aon, where I joined a small, talented team to build one of the first digital benefits enrollment systems. I worked on the data exchange platform, my first time working with large batch processing and systems and data integrations. I don’t remember those years as necessarily joyful either, but I know I was never bored at work. Year after year brought new and challenging problems to solve. I can still close my eyes and picture the 300 line SQL statement we used for a bulk export and the hours upon hours I spent profiling and tuning it to shave off seconds of runtime.

Looking back at my early professional years spent coding, I’ve come to realize there’s something deeply satisfying about coding that maybe my friends in other careers didn’t experience.

I was developing hard, practical skills, and using them to build something tangible. At the end of the day (or week, or quarter, or year), I could point to something tangible in the world and say, “I contributed to that”. I discovered the kind of personal fulfillment that comes from creating something that didn’t exist before.

Joy Fades

As I moved into leadership and management roles, I stepped away from hands-on coding. Work became… work. Meetings. Spreadsheets. Power points. More meetings.

As my salary increased, I experienced less joy at work.

And Joy Returns

Then came DevOps.

I’ve written before about how adopting DevOps practices at work brought joy back into my career. Even though I wasn’t writing the code myself, shifting our planning from annual to quarterly cycles and increasing our deployment frequency from quarterly to weekly had an incredible effect. Work became fun again.

Once I saw firsthand how DevOps not only led to better quality software, but made the team happier too, there was no turning back.

For the first time, I started to pursue happiness at work intentionally.

Work doesn’t have to suck.

A few years later, that pursuit led me somewhere unexpected: crypto. And then Farcaster. Crypto is an open invitation for builders. It’s open source. It’s permissionless. It’s global.

And for the first time in 15 years, I started coding again.

I worked my way through a 30-hour, self-paced full stack crypto dev course. I built an AI bot. I created a Farcaster Mini App so people could watch GM Farcaster livestreams without leaving the app. I vibe-coded a few websites. I joined a hackathon.

Am I loving it? Am I having fun?

Not exactly. It’s been brutal. Frustrating. A total slog.

I’ve spent entire days stuck, making zero progress. Picking up coding after a 15-year break is hard, especially when the landscape has changed so much. The last time I was hands-on, there was no GitHub, no open source at scale, no mobile apps, no modern JavaScript frameworks, no APIs as we know them, no cloud. Sure, I had plenty of familiarity with building modern software because I led teams building with all these tools, but I wasn’t the one typing the code so I’ve found it quite difficult to pick back up.

And yet… through all the struggle and frustration of learning how to code again, something surprising has emerged.

A feeling of joy.

Where does that joy come from?

There’s No Joy Without Slog

Maybe the Zen master was right: the obstacles are the way.

Maybe the slog isn’t a barrier to joy, it’s the path to it.

Let me explain with a story. I recently participated in a Farcaster hackathon with two people I know from the community: Tiago and Graham. On the day of the event, we had to decide what we wanted to build together.

A few hours before Builder Day kicked off, we dropped a bunch of ideas into a shared Google Doc. No judgments, just brainstorming. When it came time to choose, none of us felt strongly about any one idea, so we set a timer: ten minutes to decide.

To make the call, we quickly aligned on what mattered most to us:

• We wanted to ship something end-to-end—a real working MVP.

• We each wanted to learn something new.

• We weren’t trying to win; that wasn’t the goal.

• We wanted to build something that would capture attention.

• And we wanted to have fun.

Once we had those criteria, the answer was obvious: Coinaroid. It was the only idea that checked every box.

One idea we passed on was a Farcaster birthday horoscope mini app. Users would get a daily fortune based on their FID and could mint and share the result. Graham and I liked it—we both wanted to learn how to incorporate onchain transactions, and we’d seen how similar mini apps quickly captured attention. But Tiago, the most experienced developer on our team, shot it down. Not because it was a bad idea, but because it wasn’t novel. It was something he could build in an hour or two by forking one of JC4P’s open source repos (JC4P is one of Farcaster’s most prolific mini app developers). For Tiago, there’d be no challenge. No growth. No joy.

We chose the harder path. A project that would push all three of us.

We coded nonstop until they kicked us out of the building at 9 PM, then moved to a bar and kept going until midnight. We shipped at 12:15 AM.

It was one of the best days I’ve had in a long time.

Joy is in the Process, Not the Outcome

Growing up, my mom had a copy of The Joy of Cooking in our kitchen. Threadbare cover. Four inches thick. I never read it, but I remember the title. And I think it left an impression.

It’s not called The Joy of Food. It’s The Joy of Cooking. The joy is in the process, not the result.

That’s how I feel about code. It’s not The Joy of Software. It’s The Joy of Coding.

The Book I’m Going to Write

I’m going to write a book about Joy and Coding.

I’ll interview friends, teammates, and fellow builders who share this feeling. I know why coding brings me joy, but I want to document and understand how others describe their joy. Where it comes from. What it takes to access it. What makes it worth the pain.

Maybe it’ll inspire someone else to try coding. Or rediscover it. Or chase joy in their own way.

Or maybe not. But it’ll be written.

Check in with me in five years.

(And if you are reading this and have a personal story to share about joy and coding, please get in touch!)

Share

Spring Break in London

I spent a week last month visiting and sightseeing with my family in London. I enjoy traveling to new places and while I’ve been to London before, I enjoyed seeing it through a different lens on this trip. Two of my favorite quotes about travel are:

“All journeys have secret destinations of which the traveler is unaware.” - Martin Buber

“The most dangerous world views are the worldview of those who have never viewed the world.” - Alexander Humboldt

We did a ton! London Eye, eBikes in Hyde Park, Science Museum, Churchill War Rooms, Scone Making, Chelsea match, Tower of London, Buckingham Palace, High Tea at The Savoy, Theater (Book of Mormon), Bath, Stonehenge, Windsor Castle, Macbeth at The Globe Theater, Borough Market, Camden, Cambridge.

Some of my notes:

• At the London Science Museum I learned the first pregnancy tests were TOADS! “…urine from the woman being tested was injected into animals such as this Xenopus toad. The toad reacted to hormones in the urine of pregnant women by laying eggs. Specialist hospital and university laboratories held thousands of toads for these tests, which were only available to women who needed to know if they were pregnant for medical reasons.”

• One of my favorite things to do when traveling with teenagers is find a cooking class. We found a scone baking class. Great way to learn about local culture while doing a hands-on activity. British scones are very different than the triangle shaped pastry we call scones in America. We had ours served with clotted cream and jam. Delicious.

• Traveling with teens means eating all the TikTok famous things like chocolate covered strawberries at the Borough Market, honey truffle and parmesan pork ciabatta from the Black Pig, cucumber salad, garlic green beans and soup dumplings at Din Tai Fung, creme brulee donuts, dubai chocolate, k-dogs, and lavender haze lemonade. Note to self: if I ever open a restaurant, figure out how to go viral on TikTok.

• Visting Cambridge, home to a university that’s over 800 years old, was a really nice day. Punting down the river was peaceful and meditative. We ate at a pub that dates back to the 14 century. For all my critique of modern day universities, I do think their ideals are noble and worth defending: education, knowledge, philosophy and the pursuit of truth.

• Overall, London felt similar to NYC, but calmer and cleaner. I guess that’s why people call it called London Town and not London City?

Some of the (other) things

• FarCon Recap: The 3rd Farcaster user conference was held in Brooklyn earlier this month. I attended and published a recap.

• FarHack Winner: As mentioned above, I participated in FarCon Builders day this year. Joined a team and we built Coinaroid in just over 12 hours of hacking. It’s a mini app that lets you coin your content (image, title, caption) using the Zora SDK, and share to Farcaster, all without leaving Farcaster. We were happy to be selected as winners of the Base $2500 bounty, and maybe even a prouder moment was getting a shoutout from Fred Wilson of USV Ventures during his talk at FarCon.

• Finally added a much anticipated, third workflow route for GMFC101, where it can look up the full transcript of a specific show and use that as additional context to answer questions.

• 29 Seconds of Farcaster: Launched a new newsletter called 29 Seconds of Farcaster, which features a quick rundown of what happened on Farcaster over the weekend, curated by the one and only bradq, and delivered every Monday.

• What I’ve watched: Four Seasons (with Tina Fey and Steve Carell), Conclave (about choosing the pope), and Draft Day (an older sports movie that was perfect for watching with my 14 year old son).

• Enjoyed listening to this timely podcast, as I’ve had a few conversations with people lately about the intersection of DevOps/Observability and Agents. Honeycomb cofounder and CTO Charity Majors sat down with

  Conor Bronsdon

  to talk about building software in the age of AI. Great discussion about hiring junior engineers as an investment, shifting the center of gravity away from code generation and towards production, and seeing software as socio-technical systems.

  
  

• The other night at dinner, with my oldest recently home from her freshman year at college, we had a conversation that I know is going to stick with me. Out of nowhere, she said, “You know what you guys did well?”, and my ears perked up. How often do teenagers offer unsolicited praise? “You let us do things on our own,” she said. “My friends’ parents do so much for them. You make us do everything on our own.” She was referring to daily tasks in the life of a college freshman - booking flights, choosing classes, decorating her dorm, packing up at the end of the year, etc. I once wrote about wanting my kids to grow up with independence and self-awareness, not high achievement or perfect grades. I almost cried when I heard her say this because it’s exactly what I want for her, to grow into a young adult, independent, capable and confident.

Share

A Note to My Readers

I started this substack in December 2022 as an experiment to see if developing a writing habit would help clarify my thinking and/or provide other benefits. You can read about my original intentions in my first post or my more recent reflections after sticking with it for a year.

I write about twice a month and share musings, meditations, and links to things I’m finding interesting as I build out a farcaster-native media company, a modern technology consulting company, raise my kids, and have fun creating and learning in the worlds of crypto, tech, finance, science and wellness.

Thank you for supporting my writing and journey. If you’d like to get in touch you can reply to me here or find me on X and farcaster.

Until next time, keep putting good into the world. —adrienne🌏❤️

Leave a comment

Here's the list of events I attended, my personal reflections and what I'm excited about, and where I'm going next.

Events Attended

A day-by-day rundown of everything I did at FarCon from Wednesday through Saturday

Wednesday April 30

GM Farcaster Karaoke Happy Hour
Welcome event for hypersub members and friends of the show.

Thursday May 1

Builder Day – FarHack
Talks from 10AM–12PM, followed by 12+ hours of hacking with Tiago and Graham. After pushing through hurdle after hurdle, we finally shipped Coinaroid, a mini app that lets you coin content and share to Farcaster, all without leaving Farcaster, a little after midnight.

Friday May 2

GM Farcaster Early Breakfast
Hosted a members-only breakfast at the Summit location for our hypersub crew. Passed out VIP gift bags and enjoyed the calm before the storm.

Farcaster Summit
The main event. Dan and Varun opened with remarks and a Q&A. Dan also had a fireside chat with Fred Wilson. Jesse shared his vision for bringing social into Coinbase Wallet with the Farcaster feed front and center, and then gave us all beta access.
Community talks from Limone, Phil, Samuel, YB, Jihad, and Tayyab. Followed by Hot Ones with Rish & Cassie. Hackathon winners announced.
Meanwhile, we hosted the GM Farcaster podcast studio from a 2nd floor balcony overlooking the main stage and recorded short interviews with all the speakers throughout the day.

Base Golden Hour
Evening happy hour hosted by Base. Caught up with The Park, Humpty, Boop, Oxbid and others.

Bright Moments After Party
Great venue with distinct spaces: quiet lounges, a smoke-filled dance floor, and a few vibes in between. Reconnected with Metamonk, Charles, Les, JTGI, Matallo, Grin, Links, Pichi, Statuette. Met Coinbase's US policy manager (and fellow Nittany Lion). Finally met folks like Irina face to face. I know I'm forgetting names.

Saturday

Farcaster Brunch
Chill rooftop brunch. Demoed Coinaroid to a bunch of people, chatted with my hacking team about next steps as we all want to continue to build it out.

Padel Tournament
Just like the pickle ball tournament at FarCon Venice, I love seeing events that get people active and competitive and having fun. That being said, I was happy to be a spectator this year because by the time the tournament started I was already feeling so tired.

Lunch with Chicago + Prof
Snuck away for a lunch at 12 Chairs Cafe with Nounish Prof and Chicago. I casted a picture of my lunch and it's fun to see the feed still go wild for a good food pic.

Forever Library
A standout event - scholastic book fair meets a cocktail bar. Brought my IRL friend, who ended up buying a book and joining farcaster.

Reflections

What stood out to me, what I'm still thinking about, and what made this FarCon unique

1. URL to IRL Pipeline and Internet Friends

The best part of FarCon is just hanging out with internet friends. Most conferences and networking events can be draining - meeting so many new people and having to repeat the same surface level conversations over and over again. While there’s some of that at FarCon, a lot of my time was spent with people I consider actual friends, which makes everything feel easier and more natural.

The URL to IRL pipeline is a better way to make friends. You meet people on a text-based social network like Farcaster, where connections are built around the words you write and the topics you care about. You get a feel for someone’s personality and perspective without knowing things like where they live or work, their gender, how old they are, or what their politics might be. It feels more pure than meeting someone IRL and making snap judgments based on appearances.

2. Consumer Crypto is Here and It's Good

Ordering bagels at the Summit through the Blackbird app wasn’t just easy and better than most crypto apps, it was better than any web2 alternative. I used my phone to pay, one tap and done (Blackbird even covered it with $10 of $FLY when you sign up). When my food was ready, they called my name, even though I never told them who I was.

Warpcast has obviously been getting closer and closer to closing the UX gap with existing web2 social networks, removing all mentions of crypto during the onboarding, but with the launch of the Coinbase Wallet beta app with Farcaster as the home page, we're having more consumer apps enter the space. Makes me super bullish on decentralized social built on Farcaster.

Alexandria Books hosted a book fair and I brought my IRL friend who isn’t into crypto. He was able to use Apple Pay through their app to buy a book and got both a physical and digital copy. Easy peasy lemon squeezy.

3. Fred gives venture capital a good name

The highlight of the summit, based on chatter in-the-room and on-feed, was the fireside chat Dan hosted with Fred Wilson of Union Square Ventures. Luc casted some highlights and we'll be posting the recording on GM Farcaster YouTube channel as soon as we get it back from the AV production team.

Right before Fred went on stage I had a chance to show him Coinaroid, the mini app we built the day before for the Hackathon, and he immediately tried it out, and then gave it a shout out from the main stage.

4. FarCon needs some soul searching: Is it still a memetic icon or has it growing into a Tech Conference?

Has FarCon outgrown its original community meetup vibe?

The first FarCon happened in 2023 as a community experiment. Grin, one of the original organizers, described the FarCon formula as a decentralized conference for a decentralized protocol—an event for us, by us. He wrote:

The core purpose of a FarCon is to deepen connections and promote collaboration.

That original event had about 70 people who came together on a Saturday to give some talks and share their work. FarCon has grown a lot in two years. This year saw 420 ticketed attendees, and maybe twice that many in town for all the side events. FarCon is still entirely community-led and volunteer-run, but this year it got some criticism for being exclusive, for shutting ticket sales down months before the event, for not letting everyone who wanted to attend in.

It's understandable and expected. The organizers aren’t getting paid. They have to secure venues before knowing how many people will actually show up. Sponsors don’t always commit early. Planning something like this is hard.

But it starts to raise a question: what is FarCon now, what should it be, and has it outgrown the "by us, for us" community meet up vibe?

Looking ahead, is it still about deepening community and promoting collaboration among current active users? Or should it be more of a public, external facing, growth conference to reach new people interested in joining the Farcaster ecosystem? I don’t know the answer yet, but I can see a future where Merkle takes it over and runs it more like an open, commercial conference.

5. Hacking is my happy place

I decided to participate in Builder Day this year. I’ve been having so much fun building the GM Farcaster AI Bot, and I wanted to find more ways to build on Farcaster. I’ve tried messing around with mini apps before but kept getting stuck. Joining FarHack felt like the best way to learn. I was right.

I teamed up with Graham and Tiago. A few hours before Builder Day started, we shared a Google Doc to brainstorm ideas. Everything was fair game—no judgment. We had some open-ended chats to figure out what each of us wanted to get out of the hackathon. When it started, we realized no one was attached to any single idea. So, we set a timer for 10 minutes and said we had to pick something before the timer went off.

To help decide, we listed out what mattered to us:

1. We wanted to ship something—an end-to-end MVP.

2. We wanted to learn. Each of us needed some new surface area.

3. We didn’t need to optimize for prizes. Winning wasn’t the goal.

4. We wanted something attention-grabbing.

5. And we wanted to have fun.

One idea I threw out was a daily horoscope mini app. Tiago pointed out we could probably clone a JC4P repo and be done in an hour. That wasn’t true for me—it would’ve taken hours—but it was a fair point. We wanted a challenge for everyone on the team.

With that list, the decision became easy. Coinaroid was the only idea that checked all the boxes.

We worked nonstop until 9PM, when we got kicked out of the venue, then kept going at a bar until after midnight. We hit roadblock after roadblock—mostly around how to host images in a way that would work inside Zora. We tried Pinata with both IPFS and HTTPS gateways, Imgur, Arweave, AWS. Nothing worked. Eventually we got Cloudinary to work. I was ready to give up more than once, but my team kept pushing. It was such a good reminder of something engineers know deeply: the obstacles are the way. You just keep going.

Finishing the project and submitting it as a working mini app was already a huge win. Seeing people use it the next day was even better. Winning the Base award? Icing on the cake.

It also made me realize how much I miss working on an engineering team.

6. Final Thoughts

FarCon was really well executed. Everything ran flawlessly, and Emma did an excellent job pulling it all together.

I’m proud of GM Farcaster and everything Nounish Prof and I have built together. It’s incredibly encouraging to meet people who watch our show and appreciate what we’re doing.

Farcaster is going to make it. (I could be wrong, but hard to think anything but this after the energy at FarCon)

Next Steps

Where I'm heading after FarCon

• Keep building on Farcaster. Keep building GM Farcaster Network.

• Reach more people. Move one of our GM Farcaster episodes to a West Coast–friendly time (probably Monday or Wednesday).

• Open source the GM Farcaster AI bot to bring in more dev collaborators (shoutout to JTGI for the push).

• More dev work - build the GM Farcaster mini app, build out coinaroid, build out the AI bot etc.

Photo Album

Moments from the week

This coming September, I’ll be joining a group of around 20 people, traveling to the Altai Mountains in western Mongolia. We’ll be staying with Kazakh nomads, joining them for an eagle-hunting festival and a multi-day horse trek. I’ll sleep in a ger (a traditional Mongolian yurt), at a campsite with a single portable toilet and one makeshift shower for the entire group.

For 13 nights, in this remote corner of the world, we’ll have no electricity, no running water, no cell phone service.

I couldn’t be more excited.

There are two reasons I keep getting pulled back to Mongolia. The first is personal: it’s a connection to the culture of some of my ancestors. The second is harder to articulate, but it’s something I feel more intensely with every passing year: the faster technological advances push Western society forward, the more I want to balance it by stepping back. Back into places where people live in harmony with the land, in ways that have remained largely unchanged for thousands of years.

I recently attended the annual board meeting of the Jeweled Horse Foundation, where I serve as Board Secretary. One of our goals is to help preserve the nomadic way of life in Mongolia. Plenty of nonprofits serve Mongolians, but many are missionary organizations that seek to also spread Christianity. Our mission is different. We believe in supporting nomadic life without trying to convert or change the culture.

During that meeting, I found myself thinking about the rapid rise of artificial intelligence. We’re living through an incredible moment where AI is making collective human knowledge widely accessible to anyone with an internet connection. Anyone, anywhere, can learn almost anything. If you believe knowledge is power and that education is the greatest equalizer, this is an extraordinary time to be alive. It matters less and less where you’re born or how privileged you are. If you have the desire to learn, AI is making it easier and more accessible than ever before.

But I keep coming back to a nagging question: what about all the world’s knowledge we can’t digitize?

Mongolian nomads live in close relationship with their animals, drawing sustenance from milk, meat, and hide. They make cheeses, yogurts, and fermented drinks. They produce wool, felt, and textiles. They use humane slaughter techniques and know how to utilize the whole animal, wasting nothing. Their entire way of life — from basic nutrition and survival to transportation and leisure — is built around the close relationship they’ve built with five traditional herding animals: goats, sheep, camels, horses, and yaks.

Their deep, specialized knowledge isn’t taught in schools or written down in college textbooks. It’s passed directly from father to son, mother to daughter, generation to generation.

I’m reminded of a passage from the introduction to a book of Mongolian folktales that talks about the tradition of oral storytelling and touches on this same idea:

“…assumed by people who call themselves ‘civilized’ […], that because most of the Mongols were illiterate, they were also ignorant, and because they were ignorant, they were also stupid. But being unable to read is not the same thing as being ignorant, and a man who is ignorant is not necessarily a stupid man. […] [the Mongols] had a wide range of knowledge, especially knowledge of nature, knowledge of human nature, and knowledge, most important of all, of the nature of society.” — Owen Lattimore

This kind of wisdom risks being erased, not by malice, but by omission. It’s invisible to our modern technology systems that are cataloging and organizing digital information. Nomadic wisdom can’t be scraped, digitized, or fed into a large language model (the kinds of AI systems behind tools like ChatGPT, Claude, or Perplexity).

When it comes to this wisdom the nomads have, you have to live it to learn it.

I’m generally optimistic about the future of AI. I believe it has the potential to bring abundance and expand access to knowledge in powerful ways. But I also worry about what gets left behind. Doesn’t it already feel like the world is becoming more homogenous? Have you noticed how similar certain major cities around the world start to feel, with the same global brands, stores, and restaurant chains?

As AI accelerates this flattening effect, it is even more urgent to preserve distinct cultures. With loss of cultural and tribal knowledge, we lose more than just a lifestyle — we lose insight into different ways of being human.

The people building our AI systems probably aren’t thinking about nomads in the Altai Mountains. I wouldn’t have either, if I hadn’t spent time with Mongolian nomads six years ago. That experience gave me a deep appreciation for a people whose culture is so different from my day-to-day life. I saw their incredible talents firsthand — especially their horsemanship, but also the way they milk animals and make dairy products from scratch, every day. I saw how hard they work, how beautifully untouched the land is, and how much of their life is built around care, skill, and connection to nature. It made me fall in love with a place and a people whose way of life deserves to be preserved.

My fear is that if we lose these ancient ways of life, we’ll also lose essential wisdom that even the most sophisticated AI could never replicate.

I look forward to sharing my experience visiting Mongolia this September and continuing to support the Jeweled Horse Foundation.

Before you go, here are a few questions to consider:

• Can you think of other kinds of knowledge that can’t be digitized?

• What happens if we lose them?

• What ancient wisdom did your ancestors have, and how was it passed along?

• Who decides which knowledge is worth preserving?

• Can AI help protect culture, or does it risk erasing it?

• What role can each of us play in keeping different ways of life alive?

Support The Jeweled Horse Foundation: https://jeweledhorse.org/donate

Get in touch with me: X or Farcaster

The responses were full of stories of resilience, innovation, caregiving, and generosity - from mothers providing unwavering support, women overcoming personal and financial hardships, to builders inspiring the community with their projects.

To my surprise, the women of Shane’s team chose me, thanks to a nomination from the one and only Ivy highlighting my work building a technology consulting company, launching the GM Farcaster media network, and advocating for women in tech.

I'm grateful for the honor and recognition. And I would like to use these funds to amplify and empower even more women.

Empowering women has always been important to me. From studying engineering in the late 90s to working as a software engineer in the early 2000s, I was often the only woman in the room. As I progressed through my career into senior leadership roles, this pattern continued.

One of my early mentors shared with me the idea that as women rise, we should not just send the elevator back down but hold the door open for others and bring them up with us. I’ve taken that message to heart. I've seen firsthand how women can be excluded from STEM fields simply because they don't fit a certain mold. It's sad to me because STEM careers are incredible. They are high paying, intellectually satisfying, and fulfilling. They also offer flexibility, which is especially valuable for women, who tend to carry greater caregiving responsibilities. That’s why I've volunteered with Girls Who Code clubs, mentored women engineers and new women managers, and invested in women-led businesses.

My ultimate feminist belief is simple: women deserve equal opportunity in education, equal access to capital, and the right to choose if and when to have children.

I'm grateful for the recognition and I'd like to find a way to maximize the impact of the funds I received.

Here are a few ideas:

1. FarcastHER FarCon Travel Grants: Conferences are where some of the best learning and networking happens yet travel costs can hold people back. Providing grants for women on Farcaster to attend FarCon could remove barriers and empower them to lean further into the Farcaster scenius.

2. Community-Directed Donations: Ivy’s nomination made this possible, so why not let her choose one organization to support and have the FarcastHER community vote on additional donations?

3. Support Brad’s Mom: I was touched by Bradq's story about his mom, about her dedication to her sons, knowing how to support them especially in hard times, and how she herself started a business at age 70 which got wiped out after covid hit. Brad is supporting her and paying her mortgage. Brad has also been doing a lot of valuable curation and summarization of Farcaster happenings, notably his "29 Seconds of Facaster" Monday recaps and the post-rewards weekly recap, which have been helpful to our GM Farcaster podcast. What if I used the $2500 to fund Brad continuing doing his 29 Seconds of Farcaster, but as a GM Farcaster Network newsletter?

4. Support Women Builders: Funding could directly support builders in the FarcastHER community working on frames, bots, or companies on Farcaster. I love this idea, but there's a risk of spreading the funds too thinly; strategic, focused support might have a greater impact.

5. Reinvest in GM Farcaster: Keeping the funds to invest directly back into growing GM Farcaster. We're entirely boot strapped and the extra funds can give us a little more cushion and allow us to continue building our platform, where we naturally empower and amplify women's voices and perspectives.

I'd love your feedback. What resonates most with you? Are there any other ideas I should consider?

Introduction

A few months ago, I decided to take a stab at building a Farcaster AI agent, whose purpose would be to help users navigate Farcaster by answering their questions and pointing them to relevant content. The idea was to train it on hundreds of hours of video content from GM Farcaster, the live show NounishProf and I host to cover everything happening in the Farcaster ecosystem. People have called our video archive a time capsule of Farcaster history. The problem is video isn’t inherently searchable like text, so there was no way for people to interact with all that content.

I had been wanting to train an LLM on our content, but I hadn't done anything about it. Partly because I didn't know where to start or what was involved, partly because I'm not a developer and thought it would be hard, partly because we were focused on other priorities. But seeing the explosion of AI agents like Aether, mferGPT, AskGina.eth, Atlas, and Clanker pop up on Farcaster made me want my own special purpose bot even more. Seeing Six’s comment about Aether was the final catalyst that got me to take action. He casted:

It's true that anybody can make a bot/ai agent for any community - the tech is all out there in the open.

But anybody can do a lot of things, so much so that the statement becomes meaningless. Out of the set of people who can do something, only a few people /actually/ do it, and within that set, even fewer people actually do it in a way that is valuable.”

And I decided then I had to stop waiting and just do it.

For context, I’ve worked in tech for over 20 years, starting out as a software engineer, but it's been a long time since I wrote any real code myself. With the help of a developer friend I met through Farcaster and AI as a coding copilot, I started working my way through the steps toward an MVP.

In this post, I’ll share the journey of building GMFC101, a Farcaster AI agent that has access to all the video content from the GM Farcaster Network. I’ll cover the concept and vision for the bot, the architecture and tools I used, and the features I’ve built so far. I’ll also share the challenges I faced along the way and what’s next on the roadmap.

This post is for people who have ever wondered what it takes to build an AI bot, or who just want to hear about the highs and lows of experimenting with AI.

The original vision for GMFC101 was simple: build an AI bot to help new users onboard to Farcaster. I wanted it to act as a friendly assistant, trained on our Farcaster 101 video series, guiding users through the basics of the platform and introducing them to Farcaster’s unique culture. At first, that was the MVP—a small, focused bot that could answer questions about onboarding and point users to specific Farcaster 101 videos.

But as the project evolved, so did the bot’s purpose. With access to hundreds of hours of video content from the GM Farcaster Network (Farcaster 101, GM Farcaster episodes, Vibe Check, The Hub, and Here for the Art) the bot is now a resource for all Farcaster users, whether they’re brand new or have been around since the early days.

When users tag the bot and ask a question, it searches our entire archive of transcripts and responds using relevant information pulled from our content. If it finds an answer in one of our videos, it recommends the video and provides a direct link, timestamped to the exact moment the topic is discussed.

GMFC101 is for anyone on Farcaster, but it can be particularly useful for new users trying to understand the platform. Farcaster has a lot of unique language, culture, and history that can be overwhelming at first. The bot makes it easier to dive in by offering answers, context, and recommendations for content they might not have otherwise found.

1. Makes video content searchable – Video is not an inherently searchable medium. With hundreds of hours of content across our network, finding episodes where we talked about certain topics used to be impossible. GMFC101 makes it easy to search for specific topics and surface key moments from Farcaster’s history and lore.

2. Makes content interactive – Instead of passively watching videos, users can engage with GMFC101, turning content into a conversation, all without leaving their social feed.

3. Provides a Farcaster Knowledge Base – Farcaster is constantly changing as it evolves. The core team behind Farcaster is focused on finding product-market fit, not on documenting the journey to get there. That’s where GM Farcaster comes in. The bot helps make our content more accessible, especially for those who miss our live shows.

How does one go about building a bot with specific expertise? In my case, I wanted to give my bot deep knowledge of GM Farcaster video content. There are different ways to approach this, but two of the most popular options are LLM fine-tuning and RAG (Retrieval-Augmented Generation).

I chose the RAG approach for GMFC101.

Here’s how it works: Instead of fine-tuning the LLM on all your data (which can be time-consuming and expensive), you keep your knowledgebase in an external database. When a user asks the bot a question, you retrieve relevant information from your database and inject it into the prompt for the LLM to use in its response.

For example, if a user asks, “What does wowow mean?” which is a very Farcaster specific question, the base LLM model won't have that information or be able to answer it. But since we've covered that topic in our Farcaster videos, we can use RAG to help the LLM give a good answer. Using RAG, the prompt to the LLM would look something like this:

“Answer the question: ‘what does wowow mean?’ using the following information: ‘wowow is a meme that emerged from low effort replies and became popular, if you type wowow in your cast, the like button turns into a wowow reaction... ,’” (where that information was pulled directly from our transcripts).

This approach allows the bot to respond with accurate, context-specific answers without needing to be fully retrained. The key is having a reliable method for extracting, storing, and retrieving relevant knowledge—which I'll cover in the next section.

A note on Eliza:
Eliza is a popular open-source framework for building conversational AI agents. While powerful, it wasn’t the right fit for my project. It felt like overkill for what I wanted to do, and I was more interested in learning and being hands-on rather than relying on a library.

I built two main systems, each their own code repository:

1. The Transcript Pipeline: A process for turning YouTube videos into text transcripts in a format that the bot can interact with (aka embeddings, more on this below).

2. The Bot: An API endpoint that gets triggered when someone mentions the bot and handles the workflow of generating and posting a response.

These two systems work together: the pipeline prepares the transcript data, while the bot uses that data to answer user questions. The entire system is built in a modular, component-based way, making it easy to modify or improve individual parts as needed. The three main components are:

1. Transcript Management – This is everything involved in making high quality transcripts for the bot to interact with. High-quality data is critical because the bot’s unique knowledge comes directly from these transcripts. The old database mantra "garbage in, garbage out" is still true in the age of LLMs.

2. RAG & Semantic Search – This component turns transcript text into vector representations (embeddings) for semantic search. It’s how the bot retrieves relevant knowledge to include in its responses. By keeping this process modular, I can easily swap out components in the future, like changing the vector database or embedding model, without affecting the rest of the system.

3. Prompt & Response Generation – This is the part that decides what information to give the AI to get the best possible answer. It builds a customized prompt for each user question by including things like relevant transcript snippets, previous parts of the conversation, and controls to make sure the bot doesn’t respond with too much or too little detail. This component is what turns raw data into a helpful, human-like reply.

Of course, there are also supporting components for tasks like logging, archiving, and monitoring. These are important, but the three core components above have the biggest impact on the bot’s ability to get the job done.

The pipeline is an automated process that takes a YouTube video and turns it into a format the bot can interact with. In my case, the end result is storing the transcript in two different formats - one as embeddings and one as full text - that the bot has access to. The pipeline is built to run in either batch mode (processing multiple videos) or single-video mode (for ongoing updates or redos).

Pipeline Steps for each video

1. Generate Transcript – I used Deepgram to create a high-quality transcript from the YouTube video. Starting with a YouTube URL, the API turns it into a JSON file with the full transcript as one field, with sections that also break the transcript up into sentences, paragraphs, words, and it even has information about speakers. It doesn't have names for the speakers yet (just speaker1, speaker2, etc.) but that's something that can be added later.

2. Clean the Transcript – Corrects common spelling errors (like dGen → degen and NannishPrav → NounishProf). As good as Deepgram is, it's still hard to get high quality transcripts when we use so many names and words in the Farcaster ecosystem that are not common. I originally built this step to update the values directly in the VectorDB without re-calculating the embedding but later took the time to move it earlier in the pipeline so I could safely rerun files when I discover spelling errors. This is an important step because embedding values will change depending on spelling.

3. Chunk the Transcript – It's either impossible or not advisable to generate embeddings on large chunks of data. So, this step takes the transcript file and splits it into smaller sections for embedding. I went with 500 characters based on the embedding model I was using.

4. Generate Embeddings – Using OpenAI’s text-embedding-ada-002 model, each chunk is converted into a vector representation.

5. Upload to a Vector DB - Embeddings are stored in a Vector DB for fast semantic search. I went with Pinecone. I started by using upsert so I could rerun a file, and then later changed it to delete/re-insert so I could safely rerun files that I wanted to manually clean or fix the transcript for.

6. Store Full Transcript and Metadata – The full transcript and metadata (e.g., episode title, date, YouTube URL) are stored in a JSON database that the bot has access to. I have one single JSON file that contains the metadata for every episode that has been processed, and a single JSON file for each episode containing the full transcript.

Quick note on embeddings if you are not familiar with its meaning or why it's necessary:

Think of an embedding is a numerical representation of a piece of text. Imagine taking a sentence and turning it into a unique string of numbers that captures its meaning and context. These numbers live in something called vector space, which allows the bot to compare and search through them based on their similarity.

When a user asks the bot a question, the bot converts that question into an embedding and compares it to the embeddings from all the transcript chunks. The most similar ones are returned as context for the answer.

Embeddings is the technical term for how "semantic search" works.

While the pipeline prepares transcripts so the bot can interact with them, the actual bot itself is an API. When someone tags @gmfc101 on Farcaster, Neynar handles the webhook and calls the API which handles the bot's response.

Bot Flow:

1. Event Triggered – The Neynar webhook detects when the bot is tagged or mentioned.

2. API Call – The webhook calls my API hosted on Render.

3. Semantic Search – The API queries my vector DB (Pinecone) to find the most relevant transcript chunks based on the user’s question.

4. Prompt Preparation - A custom prompt is prepared that includes relevant transcript chunks and video metadata as context. For example, if the question is "What is a frame?", the prompt might look something like this:

   "You are an expert in Farcaster culture. Use the provided context from GM Farcaster Episode 123 which aired MM/DD/YYYY, to answer the user's question. The user's question is: 'What is a frame?' Context: 'A frame is...[information from our transcripts]'"

5. Generate Response – Using OpenAI’s LLM, the bot crafts a response based on the transcript content and episode metadata. The prompt itself plays a huge role in how good the response is. However, because AI responses are non-deterministic (the same inputs produce different outputs, results are not reproducible), getting consistent, high-quality results is tricky.

6. Reply on Farcaster – The bot posts a reply with the answer and a link to the relevant video, including a timestamp if applicable.

To build GMFC101, I used a combination of tools across several categories: dev tools and coding assistants, APIs, and infra & hosting. I am using the free or lowest usage/priced tier for all of them. Here’s a summary:

• Cursor Pro– A lightweight IDE built on vscode, with Claude 3.5 Sonnet integrated as its AI copilot that made writing code faster and easier. $20/month.

• ChatGPT – My overall project copilot. While Cursor was writing most of the code, I used ChatGPT as my mentor/guide at a high level for the entire project. $20/month (Probably don't need both Cursor AI and ChatGPT but it's a sunk cost as I was already paying for this.)

• GitHub – Used for version control of the project’s codebase and for supporting automated deployments. $0/month

• Postman – Used for local API testing and debugging webhook responses. $0/month

• Deepgram – Used to generate transcripts from YouTube videos, recommended by an experienced dev as being one of the services with the best accuracy. It came with $200 in free credits and I haven't needed to pay for anything additional yet. $0/month

• OpenAI APIs – I used APIs for two purposes: 1.) creating and searching embeddings using the text-embedding-ada-002 model, and 2.) for getting LLM responses to user queries by sending in prompt using gpt-4 model. Paying for usage as I go, my monthly bills during this early development stage with only light usage for the bot are in the $10 - $20 /month range.

• Neynar – The first (and biggest? only?) player in the Farcaster Infrastructure space. While Farcaster data is 100% open and accessible for anyone to integrate with, neyar makes it easy with APIs and tooling. I am using several Neynar tools including 1.) the webhook that detects when @gmfc101 is tagged on Farcaster and triggers the bot to respond, 2.) an API that summarizes a thread conversation for additional prompt context, and 3.) an API which sends the bot's reply as a new cast. I'm on the lowest tier, starter plan, at $9/month

• Pinecone – A vector database that stores transcript embeddings for semantic search across all content. I am on the free tier and with such light usage during this early development stage, I'm not even coming close to the limits of the free tier across storage, read and write units. $0/month

• Render – Hosts the production API, and is set up with CD for automated deployments anytime I check in new code. I'm on the free tier called hobby plan, $0/month

Here’s what the bot can do right now:

• Answer User Questions – Any Farcaster user can tag @gmfc101 and ask it a question about Farcaster. The bot responds with relevant information pulled from our video transcripts.

• Video Recommendations with Timestamps – If the answer is found in a specific video, the bot recommends that video and provides a direct link to the exact moment the topic is discussed.

• Contextual Memory – The bot can engage in multi-turn conversations, remembering the context of previous questions within the same conversation.

• Massive Knowledge Base/ Access to Full GM Farcaster Network Content – Initially, GMFC101 was trained only on Farcaster 101 videos. It now has access to all 200+ GM Farcaster episodes, as well as Vibe Check, The Hub, and Here for the Art.

I built an MVP in just 5 days of work, with the help of a mentor, AI copilots, and modern tooling. I continued to iterate over the course of the next several months, using AI as my guide. The process has been fun and rewarding with several key learnings along the way:

The project scope could have been overwhelming because going into this I knew nothing about how to develop on Farcaster, how to integrate LLMs into code, or how to get video transcripts. Instead of getting bogged down in too many details, I focused on building a small, functional MVP that worked end-to-end. I knew once I had something working, it would be much easier to identify what needed improvement and iterate from there. This approach helped me avoid rabbit holes and kept the project moving forward.

Even with AI copilots, code quality is still crucial. I prioritized modularity and simplicity, which made it much easier to swap out components later. For example, if I decide Pinecone isn’t the right vector database, or if I want to use Claude instead of ChatGPT for handling replies, I can swap them out without rewriting the entire system. Keeping code clean and maintainable from the start is a gift to future me.

Transcription errors have been a constant source frustration. The bot would call NounishProf “NanishPrav” and “degen” became “dGen.” These mistakes came directly from errors in the transcripts. Garbage in, garbage out. Someday I'll need to spend hours upon hours cleaning up every transcript (schlep). Fow now, I choose not to spend too much time on this because I'm choosing progress over perfection. But I know if I want exceptional results, I'll eventually have to revisit this step. As of now, I dont know any shortcuts.

The difference between a good prompt and a great one is enormous. It took several iterations to improve how the bot handled context, recommended videos, and to get it to stick to Farcaster’s 1024-character limit. Getting the bot to provide YouTube links in plain text (not markdown) was another unexpected challenge. The process of refining the prompt never really ends, but it's easy to iterate as I go.

I got distracted comparing GMFC101 to bots like Aether and AskGina.eth and all the other bot accounts that are just really good. They feel more human, have memory and natural conversations, and can take all sorts of actions on Farcaster and onchain. I started trying to add comparable functionality to GMFC101, but quickly realized it was a waste of time, as I remembered the purpose of my bot.

GMFC101’s value is in its unique access to GM Farcaster video transcripts. No other bot can surface Farcaster lore from hundreds of hours of video content. I decided to lean into that niche and stay focused on making GMFC101 the best at what it does: answering questions about Farcaster’s culture and history.

If you want to see it in action yourself, just tag @GMFC101 on Farcaster and ask it a question. Or, check out some of these examples of the bot replying to real users asking it questions.

▼Who were the cypherpunks?

▼Can you tell me about the special episode we held last year with chicbangs.eth focused on mental health?

▼What does it mean to lock degen??

▼Can you tell me anything about basecamp?

▼What's the deal with "tipping"?

Here are some of the challenges I faced while building GMFC101:

• Transcript Quality. Transcript quality is a constant challenge. Spelling errors, filler words like ums and ahs, and speakers talking over each other all degrade accuracy. Transcripts also miss key video context—like screen sharing, facial expressions, and visual references—which can make parts of the conversation confusing or incomplete. Since the bot relies entirely on transcripts, poor-quality input leads to poor-quality responses. It’s a problem I’ll need to tackle eventually to take GMFC101 to the next level.

• Infinite Loop with Another Bot. Shortly after its initial release, GMFC101 got into an infinite conversation loop with another bot because both were programmed to respond anytime they are tagged. It was comical, but pointless, and would rack up my AI bill if left unchecked.
  Solution: Hardcoded a stop once the conversation reaches a certain depth and added a prompt warning to the user when it’s one step away from that limit

• Search Accuracy & Context Limits. Early searches pulled transcripts that were relevant for the user's query, but the chunks were too small, so the bot's answers were often too vague or incomplete.
  Solution: Kept the chunk size the same for creating embeddings and for searching for content, but then expanded the context window around each chunk—adding 10 sentences before and after—to give the bot more to work with.

• Metadata Queries. The bot struggles with metadata-driven questions like “Who were the guests on episode 100?” or “How many times has DWR been on the show?” These require a different approach than semantic search.
  Solution: Working on a routing system that detects question types and chooses the right search path—metadata or transcript.

• Longitudinal Questions & Trend Summaries. The bot handles specific questions well but struggles with trend-based queries like “What were the most iconic cultural moments on Farcaster in 2024?”
  Solution: Experimenting with techniques to summarize content and trends over time.

• Prompt Design. Prompt design is a constant iteration process and warrants its own dedicated blog post. I've gone through many iterations to get the bot to answer questions appropriately, recommend videos accurately, and maintain multi-turn conversations. Some challenges were surprisingly difficult: getting the bot to respect character limits under 1,000 characters, ensuring it provided YouTube links as plain text instead of markdown, and figuring out how to give it the right personality (something I’m still working on).
  Solution: Continual iteration on the prompt, adding explicit instructions for handling text format, limiting response length, and improving conversational flow.

• Limited Testability. Testing was another challenge. Since there’s no dedicated test environment for Warpcast, I had to test everything in production, including sending replies. While simulating an environment is technically feasible, simulating a back and forth conversation between a bot and a user was more effort than expected, especially because I wasn’t as familiar with Farcaster’s data objects.

GMFC101 is still very much a work in progress, and there’s a lot more I want to build. Here’s what’s on the roadmap:

• Train the Bot on More Content – Continue expanding its knowledge base with additional shows and interviews.

• Introduce Question Routing – Right now, the bot uses semantic search for everything, but metadata-driven questions (like “Who were the guests on episode 100?”) need a different approach. I’m working on a routing system that will detect the question type and choose the right search path—metadata or transcript. (See this anthropic blog post which describes workflow routing as a pattern)

• Add More Farcaster-Specific Features – I’d love for GMFC101 to have more farcaster native capabilities and be able to do some stuff onchain. The bot account has both an ETH and SOL connected wallet, but it doesn't do anything with them yet. I can start teaching the bot how to tag people and channels, recommend accounts to follow, and highlight popular communities on Farcaster, like casts, and other capabilities to become more "human" like a regular farcaster user

• Memory & Conversation Tracking – The bot can handle short conversations now, but I'd love if it had more conversational ability to feel more human, and had memory across multiple conversations and the ability to surface relevant past discussions.

• Differentiate Speakers – Right now, the bot doesn’t distinguish between me and NounishProf in transcripts. Teaching it to recognize speakers will improve the accuracy of its responses.

• Make it Smarter - This is an evergreen effort with multiple aspects. Sometimes the bot answers perfectly; other times, it hallucinates. The goal is to improve reliability and consistency in its responses. There’s no single fix—this work will require improvements across several areas, including enhancing transcript quality, improving semantic search, refining its personality, and continuous prompt iteration.

Building GMFC101 has been fun and rewarding. It reminded me how much joy there is in problem-solving and coding, even after years of being away from hands-on development. AI copilots are a game-changer, making it possible to go from zero to a working MVP in about 5 days. But as powerful as they are, having a solid foundation in coding and architecture patterns was crucial to making the right decisions and catching mistakes.

There’s still a lot more I want to build, but this experience reinforced something I've long believed in: the best way to learn something new is to dive in, start small, and iterate.

If you’ve ever thought about building your own AI bot, my advice is just to do it. You’ll be surprised by how much you can learn, and how much fun you’ll have, along the way.

As technology professionals, we've been trained to use our brains, but if we want to be effective leaders, it's the heart muscle we need to develop.

I based my talk on my experience leading a technical team out of a state of dysfunction and deadlock — not by using hard technical skills, but by leaning into my heart and using softer skills — to get them back into a state where they could start shipping code that mattered again.

https://videos.itrevolution.com/watch/777439512

• Brain vs. Heart: Leading with the brain means putting logical and technically correct solutions above everything else, while leading with the heart means putting people first. When you reach senior leadership, you can’t rely solely on your brain anymore. You might have the "correct" technical solution, but if people aren’t willing to listen or engage, it doesn’t matter. In today’s complex, matrixed organizations, technology leaders don't operate in isolation. You need to work closely with leaders from other departments.

• DevOps Principles in Enterprise IT: Discussed applying DevOps principles to enterprise IT environments, particularly in low code and no code systems, to improve quality, increase stability, and enhance team morale.

• Cultural Transformation: Shared a personal journey of leading a cultural shift within an IT department, focusing on creating a collaborative environment and breaking down silos. Referenced the Westrum Organizational Model for generative cultures, and Ron Westrum was even in the room! (photos below)

• Impact of Leadership Style: Illustrated how leadership focused on psychological safety, collaboration, and a unified team goal led to successful project outcomes and improved trust with business partners.

• Practical Strategies for Change: Offered specific strategies such as automating everything, making work visible, and focusing on incremental delivery over big bang releases.

• The Role of Communication: Emphasized the importance of communication with all levels of stakeholders to ensure alignment and buy-in for IT initiatives.

• DevOps typically addresses technical constraints. I’d love to further explore and understand how culture contributes to the flow of value within the enterprise. The story in my talk was one such example whereby focusing on culture, I increased flow from a state of deadlock to one of small, incremental delivery.

• How can we turn soft skills into hard skills and normalize a new leadership style for the modern age?

I'm grateful for the opportunity to witness a traditional method of cheese-making firsthand, which deepened my appreciation and admiration for the Mongolian nomads' self-sustaining lifestyle—a way of life largely untouched by modern technology

Further Exploration About Nomadic Life in Mongolia

• Part 1 and Part 2 of my blog about my travels to Mongolia where I wrote about land ownership, cultural norms, food and diet, Buddhism and the Naadam festival.

• The Jeweled Horse Foundation — a non-profit organization whose board I joined this year — is dedicated to providing aid to Mongolia to promote well-being and help preserve its culture and traditions.

• Travel to Mongolia is increasing among millennials.

• Mongolia's “big five” herds are yak, sheep, goat, horse, and camel, though there are also reindeer herders in the north.

• Traveling to provide healthcare for Mongolian nomads

Before I ever heard about Farcaster, I was very active on Twitter.  The 3 main reasons I loved Twitter:

• Discovery & learning. It was by far the best place to learn about my favorite topics: tech, enterprise software, DevOps, crypto, AI, investing, wellness, mindfulness, science.

• Clarified & improved my thinking. The more I tweeted about my own personal and professional journey the better my thoughts evolved and took shape. Regurgitating things I was learning about helped me understand topics even deeper.

• Journaling and bookmarks. I loved having a written archive of things I had found interesting in the past; I used my twitter feed as an external hard drive or an old journal to refer back to.

But 2022 was the year when Twitter stopped being so valuable to me.  It was the year Elon purchased Twitter for $44B and in the 6 months before the deal closed, it was utter chaos.  I was no longer discovering interesting or useful content. Instead, my feed became entirely filled with people complaining about Elon and announcing they were leaving Twitter for any number of different platforms. Mastodon was most popular among my tech network.

I loved Twitter and I’ve always been a big Elon fan so I stuck around hoping the drama would subside and my experience would improve. But because I wasn’t getting value from the platform, I was open to finding an alternative.

Earlier in 2022 I had left my corporate job, a company I had worked at for more than 10 years and decided to give myself a few months to explore options.  During this time, I took a web3 full stack dev course to more deeply understand crypto, a topic I had been curious about and exploring on the side.  The course opened up my eyes even further to the benefits and promises that blockchain networks can bring, and during this time I started hearing about several twitter alternatives that were being built on top of decentralized technologies. I remember seeing mentions of Lens, Bluesky and Farcaster on my twitter feed.

There wasn’t a ton of information comparing the different options but someone I respect tweeted about Farcaster and it was enough to get me to try it.

Farcaster was still in beta at the time. To get an invite code you had to DM Farcaster founder Dan Romero on Twitter with a “link to something you built”. (Note: this is important! I'll come back to this later).  I felt like an imposter asking for an invite because I hadn't built a crypto app or didn't have deep crypto credentials, but I put together my plea and he sent me an invite. 

To get onboarded back then you had to download the app through TestFlight (Note: more on this later too).  TestFlight is Apple’s app that lets developers distribute beta versions of their apps to early users to test them out, before the apps are available in the app store to the general public.

I never used TestFlight before, so I had to first install Test Flight, and then I installed the Farcaster beta app through TestFlight.

Once the Farcaster beta app was installed, I used it to create a new account on Farcaster which involved spinning up a new Ethereum wallet. I got my seed phrase, wrote it on a piece of paper, and stored it away safely. The onboarding wasn't difficult for me at all because I had experience with other crypto apps but for the normal web user, it would have been really confusing. (Note: by the time you are reading this, the Farcaster onboarding experience does not require any crypto knowledge and you can register and onboard as easily as social networks you're already familiar with).

Once in Farcaster, it felt just like Twitter, so I knew what to do.  I started casting (Farcaster's equivalent to tweeting or posting). I replied to people. The names of the people I first remember interacting with were: Cameron, Samantha, Ted, Les, Kazi, DWR, Polluterofminds, tldr, Ivy. All people who are still very active today.

I was hooked. I kept coming back, day after day, scrolling the feed, casting, replying.

I've thought a lot about what exactly I found so special about Farcaster, and how and why I kept coming back and it comes down to 4 main reasons: 1.) builder vibe, 2.) people act like humans, 3.) URL to IRL pipeline, 4.) it's not dull.

More than anything I kept coming back to Farcaster because the people were interesting.  I didn't know a single person on Farcaster before I joined. And when I first came in, it seemed like everyone knew each other – like a tight knit circle of friends, and I was on the outside.  But I realized pretty quickly it wasn’t the case – most people had only met in Farcaster and didn’t know each other before.  I’m a natural pattern finder and I tried hard to find the pattern to figure out who these people were and what they had in common.  Did they all work in crypto? Were they all developers? Designers? Founders? Investors? Were they all Silicon Valley based? Were they all in their 20s, 30s, 40s? I tried to map them to existing communities I knew of elsewhere.  It broke my brain because I couldn’t find a common denominator. 

Then it hit me. Everyone identified as a builder.

A builder creates something that wouldn't exist in this world but for them and their effort. Whether it's software, an app, a website, a company, a brand, a blog, a poem, a meal, a photograph, a painting, a family.  True builders are inherently positive sum people. There was no one in Farcaster who was looking to extract value from someone else, they were all creating something new.

What bound people together was a value system.  Farcaster users didn’t vibe together because of who we were or what we do, but rather how we saw the world.  Builders are optimists.  Builders are generous.  Builders are confident and not threatened by scarcity or others.

The amazing builder vibes was not by accident. Remember Dan's requirement to get an invite into Farcaster was to tell him what you were building. I give Dan a ton of credit for seeding the early culture with this requirement.

Around the time I onboarded to Farcaster my mom was diagnosed with Pancreas cancer.  It was out of the blue.  She passed away within 4 months of her diagnosis.  It was a brutal time for me. I was driving back and forth from NY to PA several times a week.  I was in crisis mode trying to understand the disease and figuring out what to do to help her beat it.  So much was out of my control. 

Farcaster was a surprising oasis for me. A place to connect with people on a human level.  I wrote about her illness, my processing of it, and my grief after her passing. People were incredibly kind and empathetic.  Not what you’d expect from social media.  I’ll never know for sure, but I have a suspicion that Farcaster filled a need at a time when my brain was elastic and created a foundation and love that is now permanent.

The third thing that makes Farcaster special is something we started calling the URL to IRL pipeline: creating relationships first online and then bringing them into the real world.

Farcaster looks a lot like other social media platforms – there is a user profile where you can upload a PFP (profile pic), a place to write a short bio, and a twitter-like feed where you can post updates.  What was different is that most people’s PFPs were graphic avatars, not pictures of themselves.  When you meet someone in real life, the first thing you do is take in all sorts of visual cues.  Hair style, clothing, shoes, skin color, height, age, accent, – all things you can take in and make quick assumptions about the person in front of you.  When you meet someone for the first time what do you ask but “What do you do?” On Farcaster it’s different.  I was meeting people based on their words and thoughts about topics.  I became friends with people because of their ideas and thoughts without knowing anything about them.

I remember the first person who reached out and asked me to jump on a zoom.  It felt a little odd because that wasn't happening on other social media sites I was active on, but I said sure.  I had a 30-minute conversation with tldr and we just got to know each other.  No expectations. He wasn’t trying to get anything from me.  We just got to know each other.  I’ve since had at least a dozen zoom calls with people on the feed. 

I went to the NFT NYC meetup in 2023.  It was a little weird. I am in my mid 40s. I am a suburban mom.  Most people were in their 20s and 30s but the energy was awesome. I met Ted and Matthew and J4ck and Papa and many others for the first time. 

I then went to FarCon in Boston.  That also felt weird.  Traveling to a different city to meet 70 internet strangers.  But it was so worth it. I met Nounish Prof who would later become my cofounder and business partner.

Farcaster was still in beta for most of my first year. Only in late 2023 did it become permissionless where anyone could sign up. Even today with tens of thousands of daily active users, Farcaster is still at small platform that has not found product market fit.

But there's an upside to hanging out in an emergent space. It's exciting and energizing. It's constantly and continuously changing. It’s also where founders of other new apps and services are launching their products to get feedback.  As I wrote above, I hadn't even heard of TestFlight before Farcaster and now I have a dozen apps because I'm helping founders test their early versions. On Farcaster I’m getting a front row seat to crypto builders.

I've been a daily active user on Farcaster since my first day. I came in not knowing a single person. The builder vibes and strong human connections keep me coming back day after day.

At the time of this writing, about 18 months after first joining Farcaster, I have almost 100k followers, I've made dozens of new friends, and I am building a new media business with a cofounder I met on Farcaster.

If your experience on other social media sites has gotten worse over the years, like how my Twitter experience degraded, if you're looking for a more wholesome corner of the internet, or a place to meet interesting and kind people, I encourage you to check out Farcaster.

Ways to get on Farcaster:

• Use this link with an invitation to check our Farcaster: https://warpcast.com/~/invite-page/5818?id=f280ec84

• Download the Warpcast app from the AppStore for iOS or Android

• It costs ~$5 /year to use Farcaster. If we are connected on another platform, feel free to DM me for a code and I'll cover the cost of your first year.

I wrote this essay as part of the first Farcaster Writing Hackathon. If you have any questions or you’d like to get in touch you can reply to me here or find me on X and farcaster.