Companies Are Becoming a Series of Loops: AI Can Reach the Peak, but Choosing the Next Mountain Still Takes a Human
Companies are being rebuilt as a chain of interlocking loops; loops can push the current business to a local maximum, but after the plateau, choosing the next mountain still depends on human intuition.
The video won't play here. Listen to the audio instead:
The argument · tap a timestamp to hear it
The AI era will not replicate winner-take-all
The AI era will not play out the way mobile internet did, with winner-take-all dynamics: network effects once pushed search and social toward one or two players, whereas every layer of the AI stack now has roughly 20 relevant participants. Two years ago everyone thought coding agents would end up with a single winner; today Claude Code, Codex, Lovable and Replit are all moving forward at once. Anish says the people at the frontier inside the labs believe what is actually happening is autocatalytic effects — using new technology to improve your own processes — rather than recursive self-improvement, so there will be no runaway winner that swallows everyone. The takeaway for founders: don't bet on a single full-stack giant; every layer (models, agents, interfaces) still has a long-term place.
— Anish AcharyaCEOs want to build bigger companies, not fewer employees
Employees probably don't need to worry too much about being replaced by AI — not because the technology isn't strong enough, but because a CEO's incentive is growth, not shrinking. Sundar doesn't want to run a more efficient $4 trillion company; he wants to build a $40 trillion company. A Google executive friend told Anish the company isn't laying people off but using AI to compress a two-year roadmap into three months, and the hard problem has become ‘what should we add to the roadmap’. When execution is no longer scarce, attention shifts from saving labor to finding new things to do for a bigger goal.
— Anish AcharyaCompanies will become loops, and the next step needs a human
The shape of a company is shifting from a list of roles and departments into a cascade of loops: it starts with automating an individual's workflow, then functional loops, then loops that cross business units, until eventually loops run most of the company's operations. The loops themselves keep producing output, and that output becomes the input to the next layer of loops; some mechanism sends the signal back to the CEO — when to adjust the business entity, when the strategy should change. This machinery can help a company climb to a local maximum, but after the plateau it must rely on human intuition to do out-of-distribution thinking and put the organization at the foot of the next mountain. Asking AI to directly come up with a million-dollar idea often fails, because humans can set the right direction while AI only sweeps along a given slope.
— Anish AcharyaUsers want to spend time, not save it
Silicon Valley defaults to AI's value being efficiency; Anish thinks that's a misjudgment: most people don't really want to save time, they want to spend it where it brings connection, love, progress and fun. The biggest products in history have always been entertainment and social, so the consumer opportunity in AI isn't ‘a few more minutes saved’ but building loops that make people feel more connected, more loved, and feel progress or fun. He criticizes the past 40 years of technology for extending intelligence while barely extending the soul; fitting AI into these desires is a product design challenge, not a model capability challenge.
— Anish AcharyaAI is decoupling skill from desire
Anish sees AI not just as a productivity tool but more like an emotional and spiritual interface — it can split ‘skill’ and ‘desire’ apart. People who want to make music are no longer blocked by ‘I can't play piano’; people who want to write code don't have to first learn enough to hold a job; personal identity is amplified as a result. He looks back at the high-risk environment of the 1950s-60s, which greatly unleashed human potential, and argues today's world is in a low-risk state; AI has a chance to pull GDP growth from 2% to 10-15% — provided what it raises is not the speed of a single step but the scale of what a group dares to imagine.
— Anish AcharyaIdeas that are too small now fail to raise money
Investment criteria have reversed within three years. In the past, if a project was too ambitious, funds would often say a $100 million seed round was absurd; today, looking at deals at a16z, the standard has flipped — if an idea can still only be that small even with AI's help, they simply don't invest. ‘Unbounded ambition’ has become the core value for screening companies; Marc Andreessen's version is: we either go to the moon, or we leave a moon-sized crater.
— Anish AcharyaShipping something small every week is how you build product intuition
The most direct advice for product people is the old line: build more things. Find a project you tell no one about, the smaller the better, treat it as a ‘chassis’, run each new model through it, ship the finished thing, discuss it with people — that's the path to building AI intuition. The cadence only needs to be one small release a week; going crazy isn't sustainable. Anish himself used Codex to make a Mother's Day slideshow for his wife: pulling material from texts and the photo album, auto-scoring it, and finally generating 20 pages reviewing their relationship, even digging up old texts from their first date. It's small, but work like this lets you taste AI's ‘moments of delight’ over and over — once you've tasted it, you can't go back.
— AnishIn their own words · checked verbatim
The loop will help you climb to the local maxima, but then it plateaus. You need human intuition. You need somebody to actually help you land at the base of the next hill.
Anish Acharya15:12
We believe that people want to be more productive, but they don't. I think more people want to spend time than save time.
Anish Acharya32:24
It kind of unbundles skill from desire. For example, if you want to make music, you can make music now. You don't have to know how to play the piano.
Anish Acharya37:29
It can be totally unimportant. Um but use it as a chassis to use all the new models, ship things, talk about them, build your own intuition.
Anish1:09:48
Figures
| Relevant players at each layer of the AI stack | about 20 | 4:05 |
| Comparison of CEO target company size | $4 trillion → $40 trillion | 10:08 |
| Roadmap execution time compression | 2-year roadmap completed in 3 months | 10:08 |
| Administrative costs as a share of healthcare costs | 45% | 40:31 |
| Potential GDP growth driven by AI | 2% → 10-15% | 37:29 |
Glossary
- loop
- A workflow that uses its own output as feedback and keeps iterating automatically; Anish uses it to describe the smallest operating unit of the future company.
- autocatalytic effects
- Using new technology to improve one's own processes, creating acceleration rather than runaway amplification.
- recursive self-improvement
- AI repeatedly improving itself, leading to exponential runaway; Anish argues this is not the dominant mechanism today.
How to listen
Founders and product leads using AI to change their own product or restructure company processes; also investors weighing AI's impact on jobs and consumer opportunities.
If you're short on time, skip the book list, life maxims and DJ anecdotes at the end — they don't affect the AI judgment.