During my first AI workshop of the year, I worked with a team of account executives at a B2B data aggregator to create their first mini-apps and AI-powered bots. We spoke about a variety of ways that AI use can rapidly augment the work of any business professional today, and took a peek into the future of a fully agentic world where large parts of the pipeline are wholly owned by AI.

At the end of the session, one person asked the question many of us were thinking:

“Wait. So, AI is writing our emails and AI is managing their inboxes…then who will read the emails?”

Well…

…they might not be humans.

Where gen-1 software engineers started out learning in textbooks, gen-2 engineers learned from the collective hive mind of the peer-to-peer internet. Sites like Stack Overflow boasted the greatest aggregated collection of user-generated content about programming questions and answers.

For a long time, it was the only place in the world to get an answer to that thorny bug.

Until…it wasn’t.

Sam Rose@samwhoo

StackOverflow graph of questions asked per month. Holy shit.

11:12 PM · Jan 3, 2026 · 4.55M Views

887 Replies · 1.94K Reposts · 28.2K Likes

Today of course, there’s no need for human-generated questions and answers when an AI will do just as well. Increasingly, due to coding agents embedded in tools like Claude Code, we practically code without asking any questions at all.

Not only are the large language models trained on all our collective digital exhaust, but they can self-correct and answer their own questions too, often completely bypassing the need for human intervention.

This is why even popular open source coding frameworks like Tailwind can’t even afford to sustain themselves.

Yash Bhardwaj@ybhrdwj

Tailwind lays of 75% of their team. the reason is so ironic: > their css framework became extremely popular w AI coding agents, 75m downloads/mo > that meant nobody would visit their docs where they promoted paid offerings > resulting in 40% drop in traffic & 80% revenue loss

5:18 AM · Jan 8, 2026 · 944K Views

429 Replies · 967 Reposts · 13.8K Likes

While their documentation is more popular than ever and used constantly by AI coding agents, their business model depends on human users discovering and paying for commercial services. In a world where agents are the primary users, that model breaks.

The result? They had to lay off 75% of their workers.

This isn’t a productivity crisis. It’s a financing crisis.

I started Build First because I believe the digital transformation from software user to maker is the most powerful paradigm shift of our generation.

AI gives us agency to express ourselves through software in a way that was only possible before among technologists with engineering degrees.

It’s exhilarating to see friends who have never considered themselves to be engineers come out of the woodwork over the past year and use technology to solve real-world, relatable challenges.

But being a high-agency builder comes at a cost. While it’s easier than ever to create, it’s harder than ever to get attention and stay in the game.

As Tina He pointed out in her seminal essay, “Jevon’s Paradox: A Personal Perspective,” when AI makes it feel like you can do the impossible day after day, how do you know where to draw the line and actually stop working?

“We’re witnessing what I call the “labor rebound effect“—productivity doesn’t eliminate work; it transforms it, multiplies it, elevates its complexity. The time saved becomes time reinvested, often with compound interest.

When your productivity increases, several mechanisms kick in simultaneously:

Leisure’s opportunity cost skyrockets. When an hour of work generates what once took days, rest becomes luxury taxed by your own conscience. Every pause carries an invisible price tag that flickers in your peripheral vision.”

That’s why even the world’s top AI experts like Andrej Karpathy admit that they can’t keep up. The elusive carrot of 10x’ing ourselves dangles precipitously in front of us, with the AI always reinventing itself while we sleep.

Andrej Karpathy@karpathy

I've never felt this much behind as a programmer. The profession is being dramatically refactored as the bits contributed by the programmer are increasingly sparse and between. I have a sense that I could be 10X more powerful if I just properly string together what has become

5:36 PM · Dec 26, 2025 · 16.2M Views

2.61K Replies · 7.44K Reposts · 55.4K Likes

After all, while AI collapses the cost of building, it doesn’t collapse the cost of living.

To break from the 996 work cycle requires thoughtful self-discovery and time, a luxury many can’t afford. This is why people like Christine Ist, who spent six months of steadfast focused effort on an independent career, ultimately paused her indie creative sabbatical lifestyle in exchange for a full-time job (with a steady paycheck).

christine's corner

6 months later: A sabbatical reflection

I quit my job in May 2025. I wanted to go independent. I wanted a spiritually aligned career. I wanted passive income and published assets. I wanted time and space for creative pursuits…

Read more

2 days ago · 28 likes · 8 comments · christineist

When I read Christine’s reflection post, I’ll admit I was quite conflicted. On the one hand, I’m so happy to hear she found a job on a team of people who inspire her and open up the door to new opportunities (like getting an apartment). On the other, I was sad to see the end to what had seemed like a seedling of creative flourishing. (As I’ve personally learned the hard way, it takes more than 6 months to “get there.”)

To me, this mismatch represents a flaw in the system of how ideas get funded and how creators get compensated.

If AI introduces new builder personas to the table, then those people need time to tinker and experiment and play. In a way, this “gen-3 engineer” may become the artisan of our era: A builder who brings human judgment, taste, and lived experience into systems originally hard-coded by engineers with very different assumptions.

But of course, for this to work, we need economic structures that incentivize and support these new-age artists. They don’t just need time; they need financial security to build.

The limiting factor for the next wave of creators may not be skill or grit at all, but whether our funding models can evolve quickly enough to support this new way of building.

It’s too early to say how this will play out. And I’m already seeing interesting new economic models appear (including things like startup accelerators for one, revenue sharing, tipping, personal “creator coins” that fans can buy into, and referral-based partnerships). Still, until those models mature, some very capable people will step back. Not because they couldn’t build, but because waiting has a cost.

In the meantime, the rest of us will keep learning, building, and automating. Not because the path is clear, but because learning how to learn with technology is becoming a basic form of agency.

During my first AI workshop of the year, I worked with a team of account executives at a B2B data aggregator to create their first mini-apps and AI-powered bots. We spoke about a variety of ways that AI use can rapidly augment the work of any business professional today, and took a peek into the future of a fully agentic world where large parts of the pipeline are wholly owned by AI.

At the end of the session, one person asked the question many of us were thinking:

“Wait. So, AI is writing our emails and AI is managing their inboxes…then who will read the emails?”

Well…

…they might not be humans.

Where gen-1 software engineers started out learning in textbooks, gen-2 engineers learned from the collective hive mind of the peer-to-peer internet. Sites like Stack Overflow boasted the greatest aggregated collection of user-generated content about programming questions and answers.

For a long time, it was the only place in the world to get an answer to that thorny bug.

Until…it wasn’t.

Sam Rose@samwhoo

StackOverflow graph of questions asked per month. Holy shit.

11:12 PM · Jan 3, 2026 · 4.55M Views

887 Replies · 1.94K Reposts · 28.2K Likes

Today of course, there’s no need for human-generated questions and answers when an AI will do just as well. Increasingly, due to coding agents embedded in tools like Claude Code, we practically code without asking any questions at all.

Not only are the large language models trained on all our collective digital exhaust, but they can self-correct and answer their own questions too, often completely bypassing the need for human intervention.

This is why even popular open source coding frameworks like Tailwind can’t even afford to sustain themselves.

Yash Bhardwaj@ybhrdwj

Tailwind lays of 75% of their team. the reason is so ironic: > their css framework became extremely popular w AI coding agents, 75m downloads/mo > that meant nobody would visit their docs where they promoted paid offerings > resulting in 40% drop in traffic & 80% revenue loss

5:18 AM · Jan 8, 2026 · 944K Views

429 Replies · 967 Reposts · 13.8K Likes

While their documentation is more popular than ever and used constantly by AI coding agents, their business model depends on human users discovering and paying for commercial services. In a world where agents are the primary users, that model breaks.

The result? They had to lay off 75% of their workers.

This isn’t a productivity crisis. It’s a financing crisis.

I started Build First because I believe the digital transformation from software user to maker is the most powerful paradigm shift of our generation.

AI gives us agency to express ourselves through software in a way that was only possible before among technologists with engineering degrees.

It’s exhilarating to see friends who have never considered themselves to be engineers come out of the woodwork over the past year and use technology to solve real-world, relatable challenges.

But being a high-agency builder comes at a cost. While it’s easier than ever to create, it’s harder than ever to get attention and stay in the game.

As Tina He pointed out in her seminal essay, “Jevon’s Paradox: A Personal Perspective,” when AI makes it feel like you can do the impossible day after day, how do you know where to draw the line and actually stop working?

“We’re witnessing what I call the “labor rebound effect“—productivity doesn’t eliminate work; it transforms it, multiplies it, elevates its complexity. The time saved becomes time reinvested, often with compound interest.

When your productivity increases, several mechanisms kick in simultaneously:

Leisure’s opportunity cost skyrockets. When an hour of work generates what once took days, rest becomes luxury taxed by your own conscience. Every pause carries an invisible price tag that flickers in your peripheral vision.”

That’s why even the world’s top AI experts like Andrej Karpathy admit that they can’t keep up. The elusive carrot of 10x’ing ourselves dangles precipitously in front of us, with the AI always reinventing itself while we sleep.

Andrej Karpathy@karpathy

I've never felt this much behind as a programmer. The profession is being dramatically refactored as the bits contributed by the programmer are increasingly sparse and between. I have a sense that I could be 10X more powerful if I just properly string together what has become

5:36 PM · Dec 26, 2025 · 16.2M Views

2.61K Replies · 7.44K Reposts · 55.4K Likes

After all, while AI collapses the cost of building, it doesn’t collapse the cost of living.

To break from the 996 work cycle requires thoughtful self-discovery and time, a luxury many can’t afford. This is why people like Christine Ist, who spent six months of steadfast focused effort on an independent career, ultimately paused her indie creative sabbatical lifestyle in exchange for a full-time job (with a steady paycheck).

christine's corner

6 months later: A sabbatical reflection

I quit my job in May 2025. I wanted to go independent. I wanted a spiritually aligned career. I wanted passive income and published assets. I wanted time and space for creative pursuits…

Read more

2 days ago · 28 likes · 8 comments · christineist

When I read Christine’s reflection post, I’ll admit I was quite conflicted. On the one hand, I’m so happy to hear she found a job on a team of people who inspire her and open up the door to new opportunities (like getting an apartment). On the other, I was sad to see the end to what had seemed like a seedling of creative flourishing. (As I’ve personally learned the hard way, it takes more than 6 months to “get there.”)

To me, this mismatch represents a flaw in the system of how ideas get funded and how creators get compensated.

If AI introduces new builder personas to the table, then those people need time to tinker and experiment and play. In a way, this “gen-3 engineer” may become the artisan of our era: A builder who brings human judgment, taste, and lived experience into systems originally hard-coded by engineers with very different assumptions.

But of course, for this to work, we need economic structures that incentivize and support these new-age artists. They don’t just need time; they need financial security to build.

The limiting factor for the next wave of creators may not be skill or grit at all, but whether our funding models can evolve quickly enough to support this new way of building.

It’s too early to say how this will play out. And I’m already seeing interesting new economic models appear (including things like startup accelerators for one, revenue sharing, tipping, personal “creator coins” that fans can buy into, and referral-based partnerships). Still, until those models mature, some very capable people will step back. Not because they couldn’t build, but because waiting has a cost.

In the meantime, the rest of us will keep learning, building, and automating. Not because the path is clear, but because learning how to learn with technology is becoming a basic form of agency.