Max Libster @MaxLibster

You’re right — AI is the first technology that directly competes with cognitive work at scale. That makes it feel different from previous shifts. Still, history shows that automating existing work doesn’t lead to permanent mass unemployment. It creates new layers of value and changes what valuable human work looks like. The real opportunity lies in: Mastering how to direct these systems Building the new roles and businesses emerging around them Using the lower barriers for independent creation The cognitive replacement is real. The question is whether we focus only on what’s disappearing or also on what’s being created.

They told us machines would end work. In 1900. In 1980. In 2000. They were wrong every single time. Here's why the AI panic is the same story 👇 youtu.be/GbMJVfiaKVg?is… #AI #FutureOfWork #BuildingTomorrow

Anthropic calling for other labs to slow down while warning of “significant societal risks” is being read — correctly by many — as a competitive move. “We’ve built something powerful, now everyone else should ease off so we can catch up or at least not get left behind.” What’s more interesting than the cynicism is what this reveals about the current moment. The labs closest to the frontier are increasingly seeing two things at once: the technical acceleration (including paths toward recursive self-improvement) and the very real difficulty of steering that acceleration responsibly at scale. That tension is genuine. At the same time, the external pressure — competition, capital markets, national interests — makes voluntary slowdowns extremely fragile. This is exactly why the pure “AI will take all the jobs” framing is incomplete. When the most advanced players are simultaneously racing forward and asking everyone else to slow down, it creates a highly uneven transition. Some organizations and individuals will move extremely fast. Most will not. History shows that general-purpose technologies don’t just destroy existing work — they create new layers above the automation. The agricultural and computing revolutions both followed this pattern, even if the transition was painful and uneven. The difference now is the speed and that the automation targets cognitive work directly. The practical response isn’t to ignore the risks or to only focus on the panic. It’s to get clear on where the new leverage actually exists right now: becoming exceptionally good at directing these systems, identifying the new categories of work that form around them, and using the lowered barriers to build independently. Calls to slow down from the leaders are worth examining carefully. But they don’t change the fact that the tools already available today create real optionality for those willing to engage with them seriously. Clear framing: youtube.com/watch?v=GbMJVf… Do you see Anthropic’s call as mostly strategic positioning, or as a genuine signal that even the leaders are worried about the pace they themselves helped create?

The fact that even Anthropic is saying the cost of frontier models is pushing them toward public markets is telling. It shows how expensive the current paradigm has become — not just training runs in the billions, but the ongoing inference and scaling costs that grow with usage. This has two direct effects on the job discussion: First, companies that have spent (or plan to spend) enormous amounts on AI infrastructure will feel strong pressure to automate as much as possible to justify the capex. That accelerates displacement in roles that are repetitive, predictable, or codifiable — exactly the areas where we’re already seeing movement (HR, support, certain coding and analysis tasks). Second, the capital intensity also creates a barbell effect: a small number of companies can afford to stay at the frontier, while everyone else has to find ways to extract value from models that already exist. That second group is where most of the new individual opportunity actually sits. The panic narrative focuses almost entirely on the first effect. The more interesting (and historically consistent) story is the second: when a powerful general-purpose technology becomes widely available and relatively cheap to access, it lowers the barrier for new work and new businesses far more than it only destroys the old. The three areas that matter most right now are learning to direct these systems at a high level, identifying the new categories of work that form around them, and using the dramatically lower coordination and capital costs to build independently. High costs at the frontier don’t just mean “AI is expensive.” They also mean the leverage available to individuals and smaller players is rising faster than most people are adapting to it. Clear framing: youtube.com/watch?v=GbMJVf… Do you see the capital intensity as mostly accelerating job displacement, or as also accelerating new forms of individual leverage?

This is one of the cleaner signals yet that the recursive self-improvement loop is no longer theoretical for the frontier labs. What makes it interesting is the asymmetry: the labs are already seeing the internal acceleration, while most of the outside world is still debating whether AI is “just a tool” or “taking jobs.” Both framings miss the more important layer. The real shift isn’t only that AI might build better AI. It’s that the bottleneck is moving from “can we build capable systems” to “how do humans stay in the loop as the systems get better at improving themselves?” That’s exactly why the doom narrative feels incomplete. History shows that when a general-purpose technology accelerates, it doesn’t just destroy existing jobs — it creates new layers of work above the automation. The agricultural and computing revolutions both followed this pattern. The practical question for individuals right now is whether they’re positioning themselves to operate at the new layer (directing these systems, building on top of them, or using them to start things that were previously too expensive or complex) or whether they’re staying at the layer that’s being compressed. Thoughtful signal + practical angle: youtube.com/watch?v=GbMJVf… Do you see the bigger risk as the acceleration itself, or as most people not adapting fast enough to the new leverage that’s already available?

Ethan’s right — the Anthropic piece on recursive self-improvement is worth reading carefully. There’s some marketing in there, but also a very sincere internal view of where they think things are heading. What stands out is how quickly they’re seeing the loop tighten. That naturally fuels the “AI will take the jobs” narrative. But the historical pattern is consistent: technologies that automate existing work also create new layers of work above them. The agricultural revolution didn’t end with fewer total jobs — it enabled urbanization and entirely new industries. Computers didn’t just replace typists and accountants; they created software, data, and entire digital economies. The uncomfortable part this time is the speed and that it’s hitting knowledge work directly. That makes the transition feel more threatening. The practical response isn’t denial or pure panic. It’s getting clear on where the new leverage actually is: becoming exceptionally good at directing these systems, identifying the new roles that form around them, and using the lowered barriers to build independently. Thoughtful read + practical angle: youtube.com/watch?v=GbMJVf… How are you personally thinking about positioning yourself in this shift — more defensive or more offensive?

Capital rotating from public crypto into private AI bets (SpaceX, Anthropic, etc.) makes sense on one level — the perceived asymmetry looks huge. But it also highlights a broader pattern: a lot of the biggest excitement (and capital) is still concentrated in the frontier labs and infrastructure plays, while the application layer and individual leverage are still early. History suggests the biggest long-term value often ends up being captured by those who figure out how to use the new technology at scale, not just by those who build the picks and shovels. Right now that means people who get exceptionally good at directing AI, who spot the new job categories forming around it, or who use it to launch businesses that would have been impossible (or much more expensive) even two years ago. The rotation is interesting. The question is whether most people are positioned to benefit from the underlying shift — or just watching the big money move. Clear take: youtube.com/watch?v=GbMJVf… Do you think the real opportunity is still mostly in the big labs/infra, or is it moving toward application and individual leverage?

Walmart voting down even basic reporting on how AI and automation affect workers is a perfect example of why the “AI will take all jobs” panic feels so one-sided. History shows that every major technological shift (mechanized agriculture, computers, the internet) eliminated specific roles while creating entirely new categories and lifting overall prosperity. The difference this time is the speed and that it targets cognitive work. The real question isn’t just “which jobs disappear,” but whether we focus on the three areas that actually create leverage right now: People who learn to command AI become dramatically more valuable. Entirely new roles and industries are forming around the technology. AI is lowering the barrier for solo entrepreneurship like never before. Companies avoiding transparency is one thing. Individuals ignoring the adaptation opportunity is another. Full breakdown (no doom, just clarity): youtube.com/watch?v=GbMJVf… Do you think the bigger risk is companies hiding the impact — or people staying passive while the opportunities open up?

Anthropic warning that Claude is accelerating toward self-improving systems faster than expected is exactly why the job panic is so loud right now. But history shows a different pattern: Every major technology wave (agriculture, electricity, computers) destroyed specific jobs while creating far more new ones and raising overall prosperity. The difference this time is the speed and that it hits cognitive work directly. The real story isn’t just “jobs disappearing.” It’s that three big opportunities are opening up right now: People who learn to command AI become dramatically more productive. Entirely new job categories are emerging. AI is lowering the barrier for solo entrepreneurship like never before. While the labs warn about the pace of progress, the practical opportunity for individuals is already here. Full breakdown (no hype, just clarity): youtube.com/watch?v=GbMJVf… Are we over-indexing on the fear and under-indexing on the adaptation opportunity?

Everyone's panicking about AI taking jobs. In 1900, 40% of Americans worked in farming. Machines replaced almost all of them. Everyone panicked then too. It didn't cause mass unemployment. It created entirely new industries nobody could have imagined. AI is the same story — but the opportunity is hiding in plain sight. Full breakdown 👇 [youtube.com/watch?v=GbMJVf…] #AI #FutureOfWork #BuildingTomorrow

We have an AI bubble, housing issues, private credit problems… sounds scary. But here’s what I actually did instead of just debating bubbles: I ran a real 30-day experiment with AI tools starting from $0 and generated $1,400 in actual revenue. The macro risks are real, but the tools already work for individuals right now. Full documented experiment: youtu.be/GbMJVfiaKVg Are we in a bubble or just the messy early phase?

While everyone’s talking about Ray Dalio saying the AI bubble will eventually burst… I did something different. I ran a real 30-day experiment using only AI tools — no team, no big budget — and made $1,400. Not theory. Not predictions. Actual results. While the big money argues about bubbles and valuations, regular people are already finding ways to make serious money with AI right now. Full 30-day experiment here: youtu.be/2uj1Snqenj8 Do you think we’re in a bubble, or is this just the messy beginning of something much bigger?

Ray Dalio saying the AI bubble will burst “eventually” is technically correct — but “eventually” is doing a lot of heavy lifting. Everything eventually corrects. The more interesting question is what survives the correction. In the dot-com crash, the internet didn’t die — the weak companies did, and the real infrastructure kept growing. Right now we’re seeing real revenue and real productivity gains from AI at the company level. At the same time, individuals are already using these tools to create income without needing to wait for the macro picture to clear. I ran a 30-day experiment using only publicly available AI tools and made $1,400 in actual revenue. No predictions, no hype — just documented results. Full experiment here: youtu.be/2uj1Snqenj8 Do you think this cycle will follow the classic “bubble → bust → stronger foundation” pattern, or is the current AI buildout different because real cash flow is already appearing?

Dalio’s dot-com comparison is interesting, but the mechanics are quite different this time. In 2000, most companies had almost no revenue and were burning cash on pure speculation. Today, the leading AI companies are already generating real, growing revenue from actual enterprise adoption — not just promises. The infrastructure (chips, models, data centers) is being built at an unprecedented pace because companies are seeing measurable productivity gains. That doesn’t mean valuations can’t correct sharply. They probably will. But the difference between “the bubble bursts” and “the technology dies” is much larger this time than it was in 2000. While the macro debate continues, I ran a simple 30-day experiment using only publicly available AI tools and generated $1,400 in real revenue. No hype, no predictions — just consistent execution with tools that already exist today. It made me realize that even if there’s a painful repricing coming, the underlying capability to create value with AI is already here for individuals. Full experiment: youtu.be/2uj1Snqenj8 Do you think this cycle will follow the classic “bubble → bust → stronger foundation” pattern, or is AI structurally different because the productivity gains are already real and measurable?

Uber cutting 20% of HR jobs is a clear signal. HR has always been full of repetitive, process-heavy tasks — exactly the kind of work AI is getting very good at. This isn’t just cost-cutting; it’s automation hitting support functions. While big companies reduce headcount in these areas, individuals have the opposite opportunity: using the same tools to create new income streams. I ran a 30-day experiment with publicly available AI tools and generated $1,400 in real revenue. No team, no big budget — just consistent execution. Full breakdown here: youtu.be/2uj1Snqenj8 Do you think we’ll see more companies quietly cutting administrative roles like this in the next 12 months?

Ray Dalio is making a fair point — bubbles often burst when paper wealth has to be converted into real liquidity. The question is whether AI is different because it’s already generating real revenue and productivity gains at the corporate level, not just speculation. That said, while the macro debate continues, I ran a simple 30-day experiment using only publicly available AI tools and generated $1,400 in actual revenue. No hype, no predictions — just documented results. It made me realize that even if there’s a correction coming, the tools are already here for individuals to create real value today. Full experiment: youtu.be/2uj1Snqenj8 Do you think AI’s productivity gains will be strong enough to delay (or soften) the kind of burst Dalio is warning about?

NEW: Google reportedly plans to raise $80,000,000,000.00 to fund its AI spending push.

Interesting question — but while we’re debating whether AI should get paid for “good work,” I did something more practical: I ran a real 30-day experiment using only AI tools and made $1,400 — fully documented with no hype. This is what actually happens when regular people learn to use AI effectively right now. Full experiment here: youtu.be/2uj1Snqenj8 Do you think we should focus more on compensating AI… or on teaching humans how to leverage it?