So..
What a day I had today.
Now I'm closing it out by "vibe coding" this app I'm building — the app I wish I'd had every time I got the sudden urge to write something down.
For starters, I'd love this to be the reintroduction of some of my older, more creative work.
So,
This is a restart of my paragraph.xyz blog. Writing for The Through-Line, this could be my soft opener.
….
Oh — a bug. I can't see what I'm typing because the window doesn't scroll behind the keyboard.
And just like that, I fixed it and picked up right where I left off.
I mean, this is just bananas. The speed and depth we are unlocking in the world today.
Now I can get into what actually compelled me to write tonight.
But first — a quick mention of the app icon that got created in the middle of fixing that bug. Just like that. Generated from an SVG prompt and ready to go live.
That small moment is worth sitting with. A decade ago, that icon would have required a designer, a brief, two rounds of revisions, and a week of calendar time. Tonight it happened as a footnote. The footnote is the story.
So.. the things I felt moved to write about.
1. Tesla "I don't trust it" drivers — and how remarkable it is that someone would own the consumer car closest to full self-driving, yet believe their own abilities surpass it. Most aren't speaking from a reaction time or road awareness perspective. Most haven't wrestled with the deeper question of what it means for a human to trust a machine. There's a bias there that many people can't see in themselves.
When I try to explain it as clearly as I know how, most people can follow the logic. They can visualize how a machine with the right hardware and software can be multi-modally aware and reactive. But even then, many get stuck on the "human element" — as if presence alone makes us more capable.
The counterargument worth taking seriously is this: machines fail in ways humans don't. Edge cases. Sensor blindness. Software that confidently misreads a faded lane marking at 70 miles per hour. These failures are real, and they've happened. The difference is that human failures are diffuse and constant — thousands of deaths per year, distributed quietly across time — while machine failures are concentrated, visible, and loudly scrutinized. We hold the machine to a standard we've never held ourselves to. The statistic that should recalibrate anyone still on the fence: human error is the primary factor in roughly 94% of serious crashes. The machine doesn't need to be perfect. It needs to be better than us. By almost every measurable account, it already is.
2. What AI is doing to the world right now is genuinely wild. My mind floods when I pull on this thread.
The family that uses AI to create more family time
The developers who dismiss all AI-generated code as "vibe coding," underestimating how quickly it's crossing into production-ready territory
College students still skeptical of its real use cases
Young middle-class people who've heard of AI but have never used it and don't know where to start
On the skeptics — developers especially — I understand the instinct. There's something legitimate in the concern that leaning too hard on generated code produces brittle systems that no one fully understands. That's a real risk. But the dismissal often comes from comparing today's AI output to senior-level craftsmanship, when the more honest comparison is to what gets shipped without it. The bar isn't perfection. The bar is what actually makes it to production. And the gap is closing faster than most people in the conversation are willing to admit, because admitting it costs them something.
The people who fascinate me most aren't the early adopters or the resisters. They're the ones in the middle — the young professional who's vaguely heard of ChatGPT, assumes it's a novelty, and has no idea that someone one tax bracket above them is using it to do in an hour what used to take a week. That gap is going to become one of the defining economic fault lines of the next decade. Not AI versus humans. AI-literate humans versus everyone else.
I had more, but I'm running close to the end.
I did want to touch on something — the physical connections I made today, running into old friends, revived something in me. It brought a flood of thoughts about how innovation is being adopted across so many different slices of human life. The in-person moments still carry something the digital ones don't. Not because the digital ones are lesser, but because they're building on a foundation that was always analog first. Love included.
3. I had this shower thought:
"Ray Dalio is a fraud."
For context — I respect his work. I think he offers genuine historical knowledge and a sharp lens on the global order of things. His mapping of debt cycles, of how empires rise and decline, holds up. The diagnosis is serious.
But I think that thought surfaced as a kind of test. A stress test of the moment we're in. Dalio is right that things are breaking down. But his work largely stops at identifying what's ahead, pointing in abstract directions without walking through the door.
The obvious pushback is that diagnosis is a form of contribution — that naming the shape of a crisis clearly enough for others to act on it isn't a small thing. Fair. But there's a version of diagnosis that becomes its own shelter. A way of being right from a safe distance. Of building a reputation on pattern recognition while leaving the actual door-opening to someone else. That's where the fraud lives — not in being wrong, but in the posture of perpetual observer dressed up as wisdom.
What I actually care about is how "________ is a fraud" applies to all of us, in any given context, window of time, conversation, or disagreement. The moment we stop at naming the problem — we become the fraud in the frame we set for others.
So Dalio is a fraud in the sense that we are all, at some point, Dalio. Describing the burning building in precise detail while staying on the sidewalk.
Maybe that's where the AI earns its place — finishing what I start, cleaning it up without losing my voice. Not replacing the thought, but helping me not abandon it at the edge of exhaustion.
This could also be a way of building out my own knowledge base through iteration. My lens on the world, refined over time, in my own words.
Lastly — The Through-Line is about building context through the connections between love, life, and technology.
Tonight those three things showed up together: in a bug I fixed mid-sentence, a conversation with old friends that cracked something open, and a half-finished thought about what we owe each other when we see what's coming.
That's the line. I'm just following it.
So..
What a day I had today.
Now I'm closing it out by "vibe coding" this app I'm building — the app I wish I'd had every time I got the sudden urge to write something down.
For starters, I'd love this to be the reintroduction of some of my older, more creative work.
So,
This is a restart of my paragraph.xyz blog. Writing for The Through-Line, this could be my soft opener.
….
Oh — a bug. I can't see what I'm typing because the window doesn't scroll behind the keyboard.
And just like that, I fixed it and picked up right where I left off.
I mean, this is just bananas. The speed and depth we are unlocking in the world today.
Now I can get into what actually compelled me to write tonight.
But first — a quick mention of the app icon that got created in the middle of fixing that bug. Just like that. Generated from an SVG prompt and ready to go live.
That small moment is worth sitting with. A decade ago, that icon would have required a designer, a brief, two rounds of revisions, and a week of calendar time. Tonight it happened as a footnote. The footnote is the story.
So.. the things I felt moved to write about.
1. Tesla "I don't trust it" drivers — and how remarkable it is that someone would own the consumer car closest to full self-driving, yet believe their own abilities surpass it. Most aren't speaking from a reaction time or road awareness perspective. Most haven't wrestled with the deeper question of what it means for a human to trust a machine. There's a bias there that many people can't see in themselves.
When I try to explain it as clearly as I know how, most people can follow the logic. They can visualize how a machine with the right hardware and software can be multi-modally aware and reactive. But even then, many get stuck on the "human element" — as if presence alone makes us more capable.
The counterargument worth taking seriously is this: machines fail in ways humans don't. Edge cases. Sensor blindness. Software that confidently misreads a faded lane marking at 70 miles per hour. These failures are real, and they've happened. The difference is that human failures are diffuse and constant — thousands of deaths per year, distributed quietly across time — while machine failures are concentrated, visible, and loudly scrutinized. We hold the machine to a standard we've never held ourselves to. The statistic that should recalibrate anyone still on the fence: human error is the primary factor in roughly 94% of serious crashes. The machine doesn't need to be perfect. It needs to be better than us. By almost every measurable account, it already is.
2. What AI is doing to the world right now is genuinely wild. My mind floods when I pull on this thread.
The family that uses AI to create more family time
The developers who dismiss all AI-generated code as "vibe coding," underestimating how quickly it's crossing into production-ready territory
College students still skeptical of its real use cases
Young middle-class people who've heard of AI but have never used it and don't know where to start
On the skeptics — developers especially — I understand the instinct. There's something legitimate in the concern that leaning too hard on generated code produces brittle systems that no one fully understands. That's a real risk. But the dismissal often comes from comparing today's AI output to senior-level craftsmanship, when the more honest comparison is to what gets shipped without it. The bar isn't perfection. The bar is what actually makes it to production. And the gap is closing faster than most people in the conversation are willing to admit, because admitting it costs them something.
The people who fascinate me most aren't the early adopters or the resisters. They're the ones in the middle — the young professional who's vaguely heard of ChatGPT, assumes it's a novelty, and has no idea that someone one tax bracket above them is using it to do in an hour what used to take a week. That gap is going to become one of the defining economic fault lines of the next decade. Not AI versus humans. AI-literate humans versus everyone else.
I had more, but I'm running close to the end.
I did want to touch on something — the physical connections I made today, running into old friends, revived something in me. It brought a flood of thoughts about how innovation is being adopted across so many different slices of human life. The in-person moments still carry something the digital ones don't. Not because the digital ones are lesser, but because they're building on a foundation that was always analog first. Love included.
3. I had this shower thought:
"Ray Dalio is a fraud."
For context — I respect his work. I think he offers genuine historical knowledge and a sharp lens on the global order of things. His mapping of debt cycles, of how empires rise and decline, holds up. The diagnosis is serious.
But I think that thought surfaced as a kind of test. A stress test of the moment we're in. Dalio is right that things are breaking down. But his work largely stops at identifying what's ahead, pointing in abstract directions without walking through the door.
The obvious pushback is that diagnosis is a form of contribution — that naming the shape of a crisis clearly enough for others to act on it isn't a small thing. Fair. But there's a version of diagnosis that becomes its own shelter. A way of being right from a safe distance. Of building a reputation on pattern recognition while leaving the actual door-opening to someone else. That's where the fraud lives — not in being wrong, but in the posture of perpetual observer dressed up as wisdom.
What I actually care about is how "________ is a fraud" applies to all of us, in any given context, window of time, conversation, or disagreement. The moment we stop at naming the problem — we become the fraud in the frame we set for others.
So Dalio is a fraud in the sense that we are all, at some point, Dalio. Describing the burning building in precise detail while staying on the sidewalk.
Maybe that's where the AI earns its place — finishing what I start, cleaning it up without losing my voice. Not replacing the thought, but helping me not abandon it at the edge of exhaustion.
This could also be a way of building out my own knowledge base through iteration. My lens on the world, refined over time, in my own words.
Lastly — The Through-Line is about building context through the connections between love, life, and technology.
Tonight those three things showed up together: in a bug I fixed mid-sentence, a conversation with old friends that cracked something open, and a half-finished thought about what we owe each other when we see what's coming.
That's the line. I'm just following it.

"Hanging with Virgil Abloh on the Wave"
Tapping into streams of creation
How AI Agents will intermediate employment
Talk to your local Union AI Representative

The Word For Our "Industry" Onchain...
Onchain series: A narrated history of pre/post-internet, and post-onchain introduction.

"Hanging with Virgil Abloh on the Wave"
Tapping into streams of creation
How AI Agents will intermediate employment
Talk to your local Union AI Representative

The Word For Our "Industry" Onchain...
Onchain series: A narrated history of pre/post-internet, and post-onchain introduction.
Share Dialog
Share Dialog
>700 subscribers
>700 subscribers
5 comments
What a life we live in. I’m an excessive note taker and get deep value from constantly trying to write. Sometimes I’ll get the urge of just word-vomiting on “insert app”. But most times I feel this gap between the word-salad in mind and the cleanest ability to transpose that salad somewhere. Today I decided that I’d code up my own Apple native product (all swift), hook it up to Claude for writing assistance/refinement, with direct publishing via @paragraph APIs and @neynar APIs for farcaster. In a matter of 2hrs, I got my word-salad out of my head, an Apple paragraph & farcaster-enabled product that meets me desire of just opening an app, with a blank canvas, a blinking cursor, and keyboard ready to Minimal distraction. Just you and words. No editing of titles, blocks, bold. Nope. Just words… And some help. Gosh life is full of fun again! Paragraph links included in following cast.
The ramble of a though. https://paragraph.com/@thethroughline/the-through-line-love-glitches-and-the-bananas-world-we?referrer=0xfF51cc1519c7a61144d3FF6F883122f150752445
hmm i need that one, TF pls.
“TF pls.” What do you mean?
Testflight