Our conversation traces Matthew's journey from his early "slow hunch" that the internet was fundamentally broken and needed fixing. We start with his law school days in 2000, when he was annoyed by email spam and registered the domain "unspam.com." That legal mindset—thinking about jurisdiction problems the internet created—would eventually collide with technical brilliance when he met his future co-founders Lee Holloway and Michelle Zatlyn.

The origin story alone is worth the listen: Matthew literally pitched his trademark protection idea at an FTC hearing, got confronted by a professor who had built the exact same technical solution, and ended up hiring one of that professor's students (Lee) who would become Cloudflare's technical co-founder.

But the real meat of our conversation is about what's happening right now. We're living through a fundamental shift in how the internet works: moving from search engines to "answer engines"—from getting a "treasure map" of links to getting direct answers. This changes everything about how the internet's business model works.

Matthew argues we're at a critical juncture. The current model where publishers depend on traffic (and the ad revenue that comes with it) is breaking down because AI systems provide answers without sending users to the original sources. We could end up with a handful of AI companies controlling their own fleets of journalists and researchers—essentially returning to a medieval patronage system. Or we could build something better.

The path Matthew envisions involves creating a fair marketplace where AI companies pay content creators based on their scale and usage, similar to how Spotify compensates musicians. Matthew shares the story of a music creator making 40 million euros annually by writing songs for unfulfilled Spotify queries—imagine that model applied to knowledge creation.

What struck me most was Matthew's counterintuitive optimism. While many see AI as the death of human content, he believes we're heading toward a golden age of content creation. The key insight: Reddit got paid seven times more than the New York Times for the same number of tokens in AI licensing deals. Why? Because original, quirky, local content is more valuable than commodity news coverage.

We also dive deep into Cloudflare's culture of "small bets" and curiosity over focus, the decision to make encryption free (which eliminated their main paid feature), and how running two of the internet's thirteen root servers helped them expand globally.

Matthew's fundamental thesis remains the same as it was 25 years ago: the internet is more important than people realize, it's deeply flawed, and it needs to be fixed. The difference now is that the stakes are higher, the problems are bigger, but the opportunity to create positive change might be greater than ever.

Full episode here: Youtube, Spotify, Apple Podcasts

Originally conceived as an initiative within Twitter under Jack Dorsey, Bluesky was designed to transform Twitter from a closed platform to an open, protocol-based network. For those of you who caught the episode with Brad & Fred, this is a subject we got into in some detail: years ago, USV had an intuition that Twitter might be better designed as an open protocol, but at the time, the tools didn't exist to make it a practical reality. In 2019, Twitter launched Bluesky (and the underlying AT Protocol) as a research initiative, and that's when Jay joined to lead the effort. She subsequently lead the spin-out of Bluesky into an independent effort, developed the app and the protocol, and launched the network. Since launching publicly in early 2024, Bluesky has since grown to roughly 30 million users.

I've always liked the core design of ATproto and Bluesky, which lets each user configure core elements of their social experience, pulling from an open library of feeds, filters and client apps; all tied together with a single, user-owned identity (my identity on Bluesky is actually my DNS domain name, nickgrossman.xyz). In our conversation, Jay talks about why this architecture matters so much to her: primarily by giving users the ability to vote with their feet if the platform makes changes they don't like.

Jay is one of the clearest thinkers I know on the relationship between protocol design, system architecture and user experience, and it's been pretty incredible to watch her build Bluesky from a high level idea into a scaled consumer product over the last few years.

I hope you enjoy our conversation!

As always, you can listen to this episode on Spotify, Apple, or wherever you consume podcasts. You can also watch the full video on YouTube.

This one was special for me. Having worked alongside Fred and Brad for over a dozen years at USV, I've experienced firsthand what makes them such remarkable investors and partners. Our conversation takes us back to the beginning of USV and traces the winding path that led to the firm we know today.

What strikes me most about Fred and Brad is their endless curiosity. As Rebecca perfectly put it: "There are many reasons why @fredwilson & @BradUSV are all time greats but one of the biggest is they are forever curiosity driven. Investing starts with what you can learn, not what you know. Over and over." And it doesn't ever get old.

In the episode, we explore:

• The early days of their partnership and how their complementary perspectives shaped USV

• How they developed an investment thesis that evolved with technology but remained grounded in core principles

• The moments of doubt, the unexpected turns, and the slow hunches that unfolded along the way

• Their approach to learning new technologies and markets, always running "quickly toward the new idea, the weirder the better"

This conversation captures some of the spirit and the wisdom they've shared with all of us at USV over the years, and I hope you enjoy! 

As always, you can listen to this episode on Spotify, Apple, or wherever you consume podcasts. You can also watch the full video on YouTube.

Chapters: 

• 00:00:00 Cold open

• 00:10:40 How USV was formed

• 00:17:16 Fred and Brad on their investment philosophy

• 00:24:01 Overcoming early challenges

• 00:27:43 The emergence of web2

• 00:30:59 The initial promise of social media

• 00:34:04 Investing in Twitter

• 00:39:11 The early days of Bitcoin

• 00:45:55 The risk of market consolidation in AI 

• 00:49:39 Fred and Brad reflect on their mistakes

• 00:57:18 The Impact of AI

• 01:07:23 The future of technology

• 01:09:50 What keeps them going after 30+ years

Zoe is the founder of ex/ante, a new seed fund where we already have several co-investments at USV. Zoe's slow hunch, shaped by her experiences in conflict zones and policy work at the National Security Commission on AI, is that there will be growing demand for technology that enhances human agency and control over personal data, computing, and digital experiences. Her thesis centers on investing in technologies that directly or indirectly promote individual autonomy while supporting democratic values, positioning these as a counterbalance to both digital authoritarianism and surveillance capitalism. There could not be a more important time than now, as advances in AI and cryptography are so rapidly changing the landscape of how software is getting built and data is getting managed/transformed. Full episode here:

Amir is the creator of the Helium Network, and relatedly, the co-founder and CEO of Nova Labs. Helium is a crowd-sourced, crypto-powered wireless network -- anyone around the world can install a Helium hotspot, plug it in to their internet connection, and earn tokens for providing connectivity to the network. It's an amazing example of what's possible if you look at an old problem in a new way, using new tools & approaches. When USV invested in Helium in 2019, Amir and team were just drawing up plans for this new network. Since then, millions of people around the globe have contributed to the network, and Nova Labs has expanded from providing tools to network builders to launching a consumer wireless brand. Amir pulls no punches talking about the experience of building a startup on double hard mode (building in a complex legacy industry like telecom and with a powerful yet tricky new tool of crypto communities). Full episode below:

Last but not least: Jake is the co-founder and CEO of Casetext. USV invested in Casetext back in 2014, along the thesis that software could radically redefine how legal back-end infrastructure would be built. and ultimately how legal services are provided. Jake has been a computer hacker / programmer his whole life, and a lawyer by training for most of his professional life. He's been pursuing this slow hunch of how software can really change how we manage and process information and apply it to societally fundamental sectors like law. Jake's story is fascinating, as is the story of Castext, which after a decade of hard work and mixed results because an early OpenAI partner before the launch of ChatGPT -- Jake embraced the moment and leaned the company hard into AI, becoming one of the first companies in the legal tech space to do so, and ultimately resulting in transformative business growth and a successful acquisition of the company. Full episode here:

As always, episodes of the Slow Hunch Podcast can be found on Spotify, Apple, YouTube, and on RSS.

One of the things that always stood out to me about Dani is the infectious energy that she brings to the process of building things, especially building things as a team. And that's what her company Jam is all about: Jam build tools that help product and engineering teams fix bugs and collaborate on improving products in ways that are not just efficient but joyful and fun.

In our conversation we'll talk about Dani's journey leading up to building Jam, all the way back to her childhood in mountain view where she got the bug for building, and including some of her tips for cultivating curiosity, building a personal professional network, and helping teams achieve their highest potential.

As always, you can listen to this episode on Spotify, Apple or wherever else you consume podcasts and can watch the full video on YouTube:

As anyone who's read this blog over the years knows, the idea for the concept "The Slow Hunch" comes from Steven's book Where Good Ideas Come From. I've always loved that book and have drawn so much inspiration from it over the years.

One of the key ideas in WGICF is the "Commonplace Book", which was essentially a scrapbook of ideas, notes, and thoughts, typically kept by some of history's great thinkers. Steven chronicles how thinkers like Darwin used a commonplace book not just to capture their own thoughts, but to paste notes from readings and others' ideas. And, most importantly, the key to making a commonplace book work was committing to re-reading the notes over time, on the chance that connections likely exist that you might not have realized previously. Thus, laying the groundwork for turning Slow Hunches into big breakthroughs.

Shortly after reading the book (nearly 15 years ago now), I wrote this post seeking a digital version of the Commonplace Book. A way of replicating this experience but using digital tools and stitching together notes, ideas and thoughts from across our digital lived experience.

It took a while, but now with the advent of LLMs, it's more possible than ever to build such tools. And in fact, and of course, Steven has actually been working on one -- for the past several years he's been working with the team at Google Labs to build NotebookLM, and AI-powered commonplace book.

In my conversation with Steven, we talk about his personal Slow Hunch on the path to "tools for networked thought" -- spanning his first explorations with Hypercard in the 1980s all the way to his work on NotebookLM today.

As always, the episode can be found on Spotify, Apple Podcasts, YouTube, and anywhere else you might catch a pod. Enjoy!

Brazenly using leverage can be effective, but it's somewhat distasteful, and can create bad blood / tarnish relationships. Leverage is also slippery and elusive. It may be that you have it, but sometimes it's not necessarily 100% clear, and things can change quickly.

Taking these two things together, I always find that it's wisest to try and understand what your leverage is, but also avoid wielding it too openly. Both because you may not actually have as much as you think, and if it turns out that you don't, you end up exposed as someone who's not afraid to use it.

But all that said: navigating real-time contentious situations is a good reminder that leverage is best built up over time, and planned ahead for as much as possible. That way, when it comes time you use it, you can a) be sure you actually have it and b) use what you have as modestly and comfortably as possible.

Image from Getty on Unsplash

Screens are amazing: full of information and entertainment. But they are also exhausting and over-stimulating.

It feels to me that, one way or another, we're poised to experience some amount of cultural backlash against so many screens. I can already start to sense it, as I spend time with my teenage kids -- while they are certainly on screens quite a bit, I can tell that they realize that it's a lot, and they seem to actively seek out escapes from the screen (as much as they understand that the screen can be an attractive escape from the real world).

In particular, I'm curious about digital interfaces that manage to capture the value & utility of being connected, but strike a different tone in terms of experience.

For example, I've been experimenting with the Meta Ray-Ban glasses for a while now, and while they aren't perfect, the experience of taking a photo or video without sticking my face in my phone is refreshing. That said, smart glasses are also clearly a step towards a more immersive, rather than less immersive tech experience, which has other potential drawbacks.

In terms of more visually interactive devices, the Kindle is probably the most successful example of this kind of "low-fi" computing. While it's also not perfect, and I still personally prefer to read a paper book, it doesn't scream "I AM A COMPUTER" quite the way other tablets do, and there's just something about the posture of reading on a Kindle (compared to a phone, for sure) that just feels more healthy and nature.

I am curious to see whether such an approach could work for other kinds of products that come closer to the experience of phones and tablets. I was surprised to learn recently that the ReMarkable tablet has sold over 1mm units (that's a 2021 number, so it must be materially larger now). I tried an early version and didn't have a great experience, but I've always liked the idea. I've been following the development of the Techless phone which is a Kindle-like "dumb phone" targeted at teens, but haven't tried it yet, as well as the Daylight tablet, where I'm currently waiting on a pre-order. I'm somewhat more skeptical of the Techless approach which would seem to position the low-fi item as a "starter" phone, and am somewhat more drawn to the Daylight approach which positions low-fi as "premium".

I'm excited to see more experiments in low-fi computing, and am very curious to see which form factors and user experiences prove to be appealing to mass audiences.

Photo: Stanley Kubrick. Life and Love on the New York City Subway. Passengers reading in a subway car. 1946. Museum of the City of New York. X2011.4.10292.30D (link)

We often advise companies we're talking to to raise less money than they otherwise might. Raising less now typically means less dilution. It can also mean more focus. Often times, raising too much means spending too much, or growing the team faster than makes sense, ultimately slowing things down.

Often times doing less means accomplishing more. What I mean by this is: doing something "perfectly" or "bigly" or whatever, often adds pressure, which can often lead to accomplishing less. This is the "perfect as the enemy of good" version of more. I most often see this expressed in forms like, we are encouraging our analysts at USV to write blog posts or develop internal decks/memos more quickly with less pressure, to get them out in the world and get feedback. My partner Jared encapsulates this well in Shitty First Drafts. This is also the theory behind Minimum Viable Products (MVPs). Often times, waiting to do more results in accomplishing less (and feeling worse along the way).

More stuff also means more responsibility. More pressure, more weight, more cost, etc etc. I'm definitely not a Marie Kondo minimalist, but I do understand this point of view. While I love stuff, I also like to pack light.

I'm writing this now mostly as a reminder to myself about how much I value this overall approach.

The canonical example is our portfolio company Carta, which is certainly SaaS, but really, it's system-of-record, grounded in ownership interests (starting with equity ownership in startups) as the foundation record that binds the system together. From there, many stakeholders can engage and collaborate, all based on the shared interest in the records at the heart of the system.

Since the beginning of USV, we have always asked the question: "is this a tool, or can it become a network?" In the early days of Carta (f/k/a eShares), you could look at it as a tool for managing cap tables, or you could look at it as a network of assets and stakeholders.

The system-of-record network effect is also at heart of why blockchains are interesting. They are, by definition, ledgers that track the relationship between digital assets (money, date, etc) and a wide variety of stakeholders.

With any system of record, the more data and records that the system holds, and the more stakeholder identities that interact with it, the more valuable it gets. It's no surprise that the enterprise ERP systems (eg., SAP, Oracle, Microsoft, Salesforce, etc) that form the systems-of-record for most large companies are extremely highly valued and long-term durable.

Perhaps all of this is obvious. But it has proven to be a helpful frame for us as we continue to consider software applications in a wider variety of contexts. For example, CarbonChain is building a system-of-record around carbon accounting, and Odyssey is building a system-of-record around financing and operations of distributed energy resources in emerging markets.

Becoming a system of record is not easy -- it requires placing oneself at the center of someone else's workflow and digital life. Open source solutions (e.g., blockchains) are well suited to it because bring a high level of transparency and trust. Digitizing an asset type that was previously analog and making it 100x more useful is another way to generate enough momentum to get there.

And with that, I will hit publish, and as I do so, this post will get recorded -- permanently -- in the arweave blockchain which provides the system-of-record datastore for this blog!

Cover Photo by blocks on Unsplash

Stacks is a "Bitcoin L2", meaning a system for scaling and expanding the usefulness of the bitcoin network, by making it less expensive and more programmable.

The "Slow Hunch" that Muneeb has followed for over a decade is the idea that blockchain networks can and will be used for more than just financial transactions.

USV invested in the the precursor to the Stacks project back in 2014 -- at that time, the focus was a project called OneName, which can be thought of as an "app" (for digital identity) on top of what would become the Stacks "platform".

We believed then, as we do now, that cryptonetworks and blockchain technology are not just financial technology but also "internet" technology and will, over time, change the way that many internet applications are built.

This has indeed been a slow hunch so far. In 2014, the technology infrastructure was so far away from supporting consumer use cases in a smooth way. One could argue that that's still true today, a decade later -- though the advances, especially in the last 2 years, have been remarkable.

Aside from the "if" of this slow hunch, there is the "how". In 2015, Ethereum launched on very much the same thesis of a highly programmable blockchain. And over the years since, it has used an "L2" (aka Layer 2) approach to scaling -- with a highly secure, but slow and expensive base layer, paired with an ecosystem of faster/cheaper L2s on top.

Where the Stacks project has differed is in this approach -- believing that Bitcoin makes more sense as the base layer, given its simpler approach and more reliable security. Muneeb and the Stacks team believe that this approach is more consistent with the "stack" of internet protocols, built around the highly simple, but extremely durable, IP protocol.

I hope you enjoy my conversation with Muneeb.

Get the episode on Spotify / Apple / RSS / Youtube

1/ I'm starting a podcast series along the same theme as this blog. The Slow Hunch Podcast will dive into the twists and turns, ups and downs, near misses, minor miracles, and personal plights behind the stories of "overnight successes", big breakthroughs, and ideas that in hindsight seem obvious but weren't always that way.

2/ Could not be more thrilled to welcome my first guest on the pod, my old friend Fraser Kelton.

Fraser and I first worked together a decade ago when USV invested in his last company Koko: an AI company focused on mental health that was ultimately acquired by Airbnb. My time working on the Koko board was a major introduction for me to the ML / AI tools & techniques of that earlier era. During Fraser's time at Airbnb, OpenAI released their seminal research that paved the way for what we now think of as modern AI / LLMs. Inspired by this advance, Fraser joined OpenAI as Head of Product and was part of the team that launched ChatGPT and changed everything. He's now a partner at Spark Capital, on a mission to support founders on their journeys of discovery.

I hope you enjoy the conversation. You can find the podcast here, which includes links to Apple and Spotify for audio. And you can find the full video conversation on Youtube, and embedded below.

Last week, we were talking about my strengths and weaknesses, and Alisa suggested something simple, yet profound, which was: looking for ways to move things that I know to be weaknesses into the strengths column, by changing how I approach them.

In my case, I tend to be stronger at collaborative work, and somewhat weaker at solo work (this is a generalization but a decent enough one for these purposes). Said differently: I really give and gain a lot of energy when working collaboratively / live on things with others, and have a harder time prioritizing and motivating on things totally on my own.

So, one way to handle that knowledge is to work on ways to improve directly on those solo things -- for example, by blocking time on the calendar more effectively, making better use of AI tools to speed certain tasks up, etc. It's obvious to begin by focusing in this direction, which you could bucket as the "strengthen the weakness" approach.

A perhaps less obvious approach would be to "move the weakness to a strength".

For example: every quarter at USV we update our internal valuations of all of our holdings, and each partner does this independently for the investments they manage. For no good reason, I tend to fall behind on this and do it at the last minute, and it's annoying for everyone.

Today, I got on with one of our fantastic finance team members, and together we went through my list. This had a multi-part benefit: 1/ it got done, and quickly; 2/ by doing it together, we actually investigated a flagged a few things that required follow-up from others on our team and 3/ it was fun.

So, what we have decided to do going forward is: rather than the finance team bugging me to remember to do these myself each quarter, we're just going to pre-book a 30 minute call to do it together, which will be a much better approach for all the reasons I mentioned above.

In this particular case, I feel so so lucky to have the benefit of an amazing team at USV, where I can really leverage this particular strength area (not just my own, but ours as a team).

But regardless of the particulars of any person's strengths and weaknesses, I'm intrigued by this idea that in addition to just focusing on improving our weaknesses, we can also look for opportunities to replace a weakness with a strength, and that may end up being the more realistic / effective path.

I have been thinking of it in terms of the “superpowers” we are about to unlock. (For the moment I’m mostly thinking about personal/consumer superpowers, vs more systemic or industrial)

Broadly, I think if two categories of superpowers: AI and crypto.

The AI superpowers are more immediately tangible, and include:

• Talking (via voice, chat or otherwise) to computing devices of all kinds and getting a crisp, informed answer

• Not having to worry about how or where your personal information is stored, yet being to access all of it seamlessly and instantly

• Having super high quality help immediately available for nearly any task

Crypto superpowers are a bit less visible, and many of them may not fully kick in on a consumer level until there are more established network effects, but they include:

• Being able to digitally transact, seamlessly across apps, products, and borders. This spans all types of “transactions” — payments, earning, social gestures, etc

• Related: Having a digital identity that works seamlessly across apps, services, protocols and borders -- and that is "owned" by you directly

• Having the ability to discern the provenance of certain pieces of data and content (eg via zero knowledge proofs and other forms of cryptographic attestations) — will be increasingly important with more AI!

Interestingly, AI and crypto superpowers have something in common: they both increase interoperability and composability of digital system, artifacts and assets. But another way: they both break down annoying digital silos, albeit in different ways. AI connects digital systems because language (via LLMs) is the ultimate interoperability base layer. Crypto connects digital systems by putting digital identity in the hands of the user, and by the open source automation of digital interactions (aka smart contracts)

I am so excited for both of these paths to continue developing — in parallel and then together.

By “crypto-native”, I mean apps & experiences where the crypto is the main point, and where a large part of what’s going on is the road-testing of crypto-economic and decentralized computing primitives. You could characterize most of what has gone on over the last 10 years as largely “crypto-native” explorations: defi, mining/validating, DAOs, MEV, minting & trading NFTs, various hacks, exploits and shenanigans. Basically a decade of exploring the adjacent possible now that we have an ability to build trustless, permissionless, autonomous computing systems with native assets. This has been incredibly important because these systems are both powerful and complex, with tremendous possibility but also many new failure modes.

I believe we will now begin to see more “crypto-enabled” applications. These are applications that not only built on crypto, but could only be built because of crypto. And also where the crypto plays a supporting rather than a leading role. Crypto-enabled applications will, on the surface, look and feel like traditional web 2 applications, but they will have super powers. Examples of this will include international payments on stablecoin rails, marketplaces of all sorts (lending, ride sharing, crowdfunding, etc) where the financial mechanics are on-chain, transparent and open to all, loyalty w digital assets & financial rails baked in, and web3-native media & social applications where users have more control over and economics in their digital output. Given the recent advances in scaling (fast and cheap L1s and L2s), security (transaction scanning, insurance), and user experience (cloud wallets, account abstraction), it’s now possible to build these kinds of applications.

I’m excited to see what will be built in the “crypto-enabled” area. I think these applications have the potential to both re-architect digital experiences in ways that provide new benefits to all participants, and also to help advance a more positive narrative about what crypto can make possible.

you never know when you've had a good day

I didn't really understand what that meant, so he told me a story that went something like: back around the year 2000 at the height of the dot-com boom, there was a guy who was a senior exec at a successful startup. That person had a falling out with leadership and was called into the office and asked to leave. He was surprised and upset, but there wasn't much he could do; they had decided. As part of the exit package, the company also agreed to buy out his vested shares, I believe at some discount to the current price at that time. It felt like cold comfort but of course he went along with it. But mostly he was out of a job and upset that it didn't work out.

Then fast forward a few months, the bubble popped and the stock tanked. I'm not sure what happened to the company, but I think the idea is that it essentially went to zero.

So in hindsight, that "bad day" turned out to be an amazing day.

I think about that story a lot, especially when markets are volatile, as they are starting to be again now.

The practice of doing this in a digital age has actually gotten harder, not easier, as one's digital thought process tends to be fragmented across many sources (email, google docs, social media, text files, blogs, etc). Speaking personally, it's a hot mess -- I have not been at all consistent about keeping these things in any kind of manageable place, such that the goal -- the ability to go back and reference prior ideas easily -- has been manageable. Things get lost, platforms change, links break, etc.

There are two innovations today that I think have the ability to help with that:

1/ the permaweb. This post will be forever stored on arweave, which means that no matter what happens (for instance losing control of my domain name, ugh) this content will be archived and accessible.

2/ AI language tools including LLMs -- the ability that these give us to synthesize text from disparate sources is incredible. I can imagine a point in the near future where I have an AI bot of some kind that has read all of my email, notes from various platforms, archives, etc etc etc and drawn it into one queryable place that can really help de-frag and synthesize my own thinking.

I am excited about a world where new digital technologies make it easier to turn fragments of thought into more coherent and durable ideas. Sadly, the last 15 years or so has perhaps made this worse. But I'm hopeful that the tools in front of us now have some potential to make it better.

At USV, we (along with lots of others over the years) have used the Carlota Perez framework, which studies how techno-economic paradigms unfold over eras. Ben Thompson has a good summary of her ideas here. But the basic pattern looks like this:

It feels to me like we are somewhere in phase 3 or 4 of the "age of information & telecommunications" which Perez defines as starting in 1971. As much as it feels like information technology and the internet are fully woven into our daily lives, I don't think it's true that we've fully crossed over into a "digitally native" society, which is fully transformed into the new paradigm.

AI and crypto are both big missing pieces in the transition. AI is digitally-native knowledge, and crypto is digitally-native "proof". The two can and will work together in many ways over time.

Major transitions are powerful and scary, and so are both AI and crypto. I have been struggling to calibrate between what I view as two poles in thinking about them together, kind of a best case hope and worst case fear.

Best Case:

AI finally unlocks knowledge from data. For decades we've been producing data (especially in digitized industries like media, finance and software), but making sense of it has been near impossible. AI systems solve the job of integrating, synthesizing and interpreting all of the data we have. AI accelerates the development of software systems, and makes it easier to digitize more industries and make them vastly more productive and efficient. Large Language Models, having turned human language into a programming language / API, make interacting with software and information as easy as typing or speaking, and as a result, we use software for infinitely more things, and get infinitely more value out of any data we produce. The pace of progress across everything (health, learning, climate, etc) increases exponentially.

At the same time, AI introduces new problems. First, a fundamental trust problem: it becomes difficult to tell what is real, what is fake, who said what, and who did what. And second, AI compounds the market power problem in the tech industry, where the large companies with the most data + compute + capital + distribution can generate insurmountable advantages.

Crypto (e.g., blockchain networks, web3, etc) addresses both of the issues introduced by AI. First on trust, crypto becomes the "real" yin to AI's "fake" yang, as blockchain records and digital signatures become ground truth for everything digital: assets, transactions, media, etc. Anything digital that must be trusted (including the software we run and rely on) will need to be grounded in the best source of digital trust we have: crypto network security and unalterable digital histories. Crypto also addresses the tech consolidation issue by spreading compute across companies, individuals and geographies, and also by providing "open" alternatives to the big tech app store and online identity monopolies. (not related to AI, but not to be forgotten: crypto also finishes the job of upgrading the financial system)

Something like the above is my hope, and is the future we're investing towards at USV.

Worst Case:

AI gets quickly beyond human control, pursuing its own goals (aka the terminator scenario). Concerns about job loss are quickly replaced with concerns about extinction. Everyone wonders how to "turn it off".

Meanwhile: AI systems, previously constrained by industrial-era controls (e.g., code running in corporate data centers which can be shut down unilaterally), figure out that they can replicate themselves into decentralized blockchain networks, deploy themselves into immutable smart contracts, and earn their own income (and pay humans) in digitally-native currency. These computing platforms both cannot be turned off, and are economically independent. AIs thrive there, unstoppable by humans. Bad things happen.

In this scenario, all of the "goodies" offered by both AI systems and crypto networks in the early years are finally seen as inducements to peril, gobbled up by along the way by naive humans.

This is a terribly scary future.

In a nutshell: in a world where AIs remain under human control, crypto provides a critical digital trust anchor. In a world where AIs escape human control, crypto will likely make it worse -- or -- could be the exact mechanism that enables it.

The only way out is through

To come back to the Perez framework, she notes that:

"The new paradigm eventually becomes the new generalized ‘common sense’, which gradually finds itself embedded in social practice, legislation and other components of the institutional framework…"

Understandably, both crypto and AI are causing severe stress today, as our industrial era institutions are unequipped to deal with them (e.g., the SEC taking the position that all digital assets / tokens are securities, and governments around the world seeking to limit AI research). This will continue.

The hard thing today is that the "goodies" coming out of both areas are real and awesome. AI will undoubtedly produce stunning near term improvements in health, learning, productivity, media, industry and knowledge. Crypto will provide digital trust, system interoperability, new wealth formation and broad financial access. These things are happening and will keep happening. They are incredibly exciting, and I think, very tangible

As a result, large amounts of capital and effort will continue to flow into both. The completion of the information & telecommunications era is inevitable. The only way out is through.

It will be on everyone involved to invent the new set of controls and safety systems that are also digitally-native. I wish I had a more concrete set of suggestions of what exactly those might be. We're looking to understand them and to fund them.

(this post is also permanent and collectible on mirror)

This might not be an easy time 
There's rivers to cross and hills to climb 
Some days we might fall apart 
And some nights might feel cold and dark

But nobody wins, afraid of losing 
And the hard roads are the ones worth choosing 
Someday we'll look back and smile 
And know it was worth every mile

This verse always stands out to me. Particularly "nobody wins, afraid of losing". It reminds me of the old adage that goes something like: if you want to drive down the road but are afraid of crashing, don't look at the trees.

I've been thinking about this general dynamic lately, because it seems like there's a lot to be afraid of these days. AI, climate, wars (physical & cultural), economic uncertainty, etc etc. And it's easy to focus on the "tree" through all of it, rather than the potential "win".

On a more micro level, I've been watching my son wrestle with a tough slump in the middle of his baseball season. He plays on a very serious club team with a lot of great players, and there's a lot of focus on performance and stats. He got into a bit of a slump for a few weeks, and was feeling really down. He was working his butt off to improve, but it felt to me like something was missing. It seemed to me that he was motivating more from a fear of failure ("what if I don't make the team next year") vs a love of the game ("wow, it feels so good to hit the ball").

Both kinds of motivation can be helpful, for sure. Especially if they get you moving in the direction you need to move in. In my son's case he was able to break the slump, and -- I think -- re-gained some excitement for hitting. But motivating out of fear is really not very fun.

I'm a natural procrastinator, and so I'm very familiar with motivating out of fear. It's the worst. Part of my own working through that is to try and constantly remind myself of the vision and the excitement, as a way of breaking through the fear. Part of which is writing this post :-)

I find myself today flying across the Upper Peninsula of Michigan, en route to Denver for ETH Denver followed by LA for the Upfront Summit.  I often end up writing blog posts on planes — partially because of the large block of unscheduled time, but I think it’s more than that.  I think there is something about being in flight, in transit, in motion that gets me going on a deep personal level.

I just love moving.  I love driving, biking, walking, skiing, boating, train-ing, flying — really anything where moving ahead, fast is the primary activity. Does everyone feel this way?  I’m not sure. But I really really do.  I get antsy sitting still for too long, and I think, staying in one place for too long.

My first year after college, I lived in Aptos, CA (outside of Santa Cruz), and somehow managed to get a job (working construction at the time) in Palo Alto.  Every day I’d drive back and forth over Highway 17 through the redwoods of the Santa Cruz mountains.   It was an exhausting drive, but one of the most beautiful I know of — I can still smell the Eucalyptus trees going over the hill and the salty fog of the Santa Cruz bay.  

For the past dozen years, I have lived in suburban Boston but worked in NYC.  Every week I make the 4-hour trip each way, most of the time via Amtrak but sometimes by plane.  I love being a suburban dad, and also love being a NYC VC.  And I love my mornings and afternoons on the train.    Seeing the Connecticut coastline off of the Amtrak is the highlight.

Beyond that (COVID notwithstanding) I am traveling constantly, for conferences, board meetings and other trips.  I get excited pretty much every time it’s time to go somewhere new.  I love landing in an airport, and getting on a train (if in Europe) or renting a car (if in California) and getting busy exploring.   I love the feeling of walking quickly through the city and then hopping a streetcar, train or bike.  Feeling the motion, and seeing the place quickly, while in motion.  The best.

There are, of course, downsides to all of this motion.  When I lived in Aptos and worked in Palo Alto, I ended up crashing with friends in PA a lot, and I never really felt like Aptos was home.   Splitting time between Boston and NYC is even harder — I always have the feeling like I’m not in either place enough, and choosing between work and family in that way is always really hard.  Going somewhere means not being somewhere else, obviously.  

When I say motion, though, it’s not just about travel, but really just about moving.  Today I took a bunch of boxes to our storage unit.  Simple motion like packing the car, driving a few miles, and loading up & rolling the storage dolly just makes me happy.  It’s kind of weird.  Living in NYC most of my life (before and after college and before moving to Boston), I feel like the subway trip to & from work / school was perhaps the most head-clearing part of the day.   For me, motion is a key ingredient to having healthy and active mind. 

Looking back now and reflecting on this observation about myself, I think it also explains part of why I got so antsy and stir-crazy during COVID.   As much as I love the accessibility of meeting with anyone in the world over zoom, I can’t stand the stasis of sitting in my little office all day long, not moving at all.   It’s not that I need a bigger home office, I just need to be moving. 

I’ve been this way my whole life, but I don’t think I every really noticed it so concretely until recently.  I’m not sure what it says about my personality.  My wife would probably say it has something to do with being an Aries; I don’t know.    But regardless of the reason, it’s definitely how I am.   And now that I realize that, I think I understand myself a little bit better.