Continuous learning gets unlocked this year. Calling it now.
Today’s models are mostly frozen. The one you use tomorrow is basically the same one you used today. It doesn’t truly learn from working with you.
Continuous learning changes the game. An AI that compounds on the job. Builds context over time. Develops judgment. Gets better at your work by actually doing your work.
Most people are still debating which model is smartest today.
The bigger story is the first one that can genuinely learn tomorrow.
@CryptoMikli I think it’s very likely AI reaches a point where it feels like an extension of cognition. I think the capabilities of that extension will depend on your brain. AI can’t help if you don’t know what to prompt. Stay in school kids.
Everyone’s panicking about AI taking their jobs.
That’s cute.
Meanwhile people are literally working on autonomous LLM powered weapons getting cheap enough for genocidal regimes to deploy at scale.
We’re arguing about résumés while sleepwalking into kill-bot industrialization.
Partnership vs displacement is exactly the right framing. We've been building around this thesis at https://t.co/Jp3a95FliL from day one — AI that operates as a collaborative teammate inside financial workflows, not a replacement for them. The best enterprises won't choose between humans and agents. They'll build systems where human judgment and AI execution reinforce each other. That's where the durable value lives.
The compute demand story is real, but what's underappreciated is what happens after the infrastructure is built. Enterprise adoption of agents doesn't just mean buying GPUs — it means rearchitecting workflows around AI teammates, not AI tools. The firms that treat agents like embedded digital labor (not just a search upgrade) will capture most of the value here.
The real question isn't pricing model — it's moat. If your software sits on commodity data, AI agents eat you alive. But if you own proprietary data and workflows that agents *need* to function, you become more valuable, not less. That's what we're seeing in finance at https://t.co/Jp3a95FliL — the agent layer actually increases demand for differentiated data.
The compounding effect is the real story: if agents add 2 days of output per week, you're not just 40% more productive — you're on a fundamentally different output curve. The firms that figure this out in the next 18 months won't just outperform. They'll be competing in a different category.
@PeterMcCormack Capitalism will find new bottlenecks around human preference, legal responsibility or trust and the job market will transform around that. Computers caused task level displacement not permanent unemployment. Same will happen here.
Photography didn’t eliminate artists.
It made mediocre art irrelevant.
AI will do the same to software.
When software becomes cheap, design and simplicity become everything.
Only the best-designed products survive.
@emollick I think the alignment issue is largely a red herring at this point. Models that are aligned to humans are more economically valuable than ones that are not. Therefore all natural selection right now is pushing towards increasingly aligned *and* intelligent AI.
Chess engines beat humans 30 years ago. Grandmasters still sell out arenas. Bank tellers survived ATMs. Doormen survived auto-doors. Human preference for human presence is the most resilient economic signal ever. AI automates manual effort — it won't eliminate the human premium. Jobs will exist forever.
The next step is AI that builds this without you asking — it already knows your goals, what you find interesting, what form factor you prefer. YouTube doesn't wait for you to search. The intelligence layer will stop waiting for prompts and start shipping personalized software from context alone.
Markets have always been late to price structural disruption — ask newspaper shareholders in 2005. European equities carry two underpriced risks: the intelligence layer compounds at software speed, not capex speed, and Europe's regulatory drag is measured in years. The moats that matter in AI aren't national — they're contextual. There's no home-field advantage when the field is LLM context.
@WillManidis Every city council that blocks a data center is setting the terms for who gets to participate in the intelligence economy. The model layer abstracts. The compute layer doesn't. The towns that refused railway stations in the 1850s are still dealing with the consequences.
The positive vision isn't landing because it hasn't materialized at scale yet. But it's real — and it's bigger than job preservation. AI is collapsing the gap between having an idea and executing on it. The person who needed a team of 10 can now operate like one. That's not a displacement story, it's a leverage story. It'll land when people feel it, not when Sam Altman says it.
The progression is already visible:
First humans ask AI to do work
Then AI suggests actions and humans approve.
Then AI executes within guardrails.
Then it operates independently under defined mandates.
Then the AI starts suggesting optimal work for the human to do to achieve a task.
Eventually, collaboration becomes so tight it no longer feels like using a tool — it feels like an extension of cognition.
Humans remain the source of preference, trust and context. AI becomes the infrastructure for judgment and execution.
9 months ago on Fox Business, I said AI would work like YouTube — it shows up, knows who you are, and automatically finds ways to create value for you: https://t.co/4GlWFZvPcJ
Back on this week: AI is making the cost of producing software approach zero. It's going to become like water: https://t.co/gA7IeJGu3B
Same thesis, one year later. The intelligence layer is being abstracted away from the software. What you pay for won't be the tool anymore — it's whether the tool knows you.
Respectfully, this misses the bigger shift. In traditional SaaS you find PMF, scale it, optimize it. In AI, capabilities evolve quarterly, customer expectations reset constantly, and new surface areas unlock value that didn't exist last month. Signals of real demand have a short half-life.
PMF is no longer a one-time achievement — it's a moving target. And the vast majority of traditional SaaS companies are structurally unable to keep up with that pace. SaaS isn't benefiting from cheaper code. It's getting murdered by it.
@KyleAsay_ Why does it have to be a closed source solution that puts everyone out of business? An open source “OpenClaw 2.0” running the next generation of open source LLM able to write all software for free could also do it. It might just be the case all software is going to lose value.
The weird part is that the fear itself is becoming the disruption. Companies are laying off workers because of AI's *potential*, not its performance — HBR literally just published that headline. Meanwhile the actual labor data barely shows a shift.
The AI labs created a narrative so powerful that companies are reorganizing around it before the technology even justifies it. That's not a prediction about AI. It's a prediction about how humans respond to predictions.