Jeremy

The ACX deep dive on the OpenAI/DoW contract is worth your time. The government can buy Americans' data from third parties and analyze it however they want. That's not "surveillance" under the legal definitions the contract uses. AI just removed the only real constraint, which was scale.

Fair. His framework is great at explaining why platforms win in stable technological paradigms but less equipped for paradigm shifts. He dismissed Citrini as doomerism rather than engaging with the structural tension between “AI changes everything” and “incumbents will be fine.” It’s model capture.

Hard to sound the alarm on disruption when your subscribers are the ones being disrupted.

But there's a tension Thompson won't resolve. You can't argue AI is the most transformative technology since the internet AND that its economic disruption will be easily absorbed. His audience is the tech executives who benefit most from the "incumbents with network effects will win" narrative.

Another Viral AI Doomer Article, The Fundamental Error, DoorDash’s AI Advantages Another AI doomer article has gone viral, and like many in the genre, it lacks an appreciation for dynamism and markets. Then, why DoorDash is going to be fine.

Ben Thompson's critique of the Citrini AI report gets some things right. His point about physical-world moats is strong. You can't vibe-code a delivery fleet. And Citrini does treat the economy as too static.

stratechery.com/2026/another...

Cloudflare essentially commoditized Vercel’s core moat (the tight Next.js integration) in a week for $1,100.

A recruiter just shared a role at a large, publicly traded company which explicitly asks for “communication skills not reliant on LLMs.”

Someone on that team absolutely got one too many ✨perfectly worded✨ Slack messages with zero original thought.

The 1000 commits problem | Dave Kiss Software is moving faster than humans can manage. The old ways don't scale to AI-assisted velocity.

davekiss.com/blog/the-100...

My entire industry needs to internalize this. There's a knee-jerk bias to be "contrarian" without getting good at identifying the rare circumstances that actually require it. The skill isn't being contrarian, it's recognizing when conventional wisdom has actually broken down.

Reed Hastings - Building Netflix - [Invest Like the Best, EP.453] Spotify video

Reed Hastings on contrarian thinking: "Most of the time, contrarian thinking is wrong. The conventional thinking is right. Once in a while it's right and that's when you get the big reward."

open.spotify.com/episode/6PrR...

5/ Build better containers for safe experimentation and deviance returns. The appetite is there. The infrastructure isn’t.

4/ What would help: sabbatical insurance, companies with explicit return tickets after 1-2 years of exploration, low stakes incubators focused on experimentation rather than venture scale outcomes, geographic trial programs like Tulsa Remote but broader.

3/ What’s the equivalent for career experimentation? Intellectual risk? Unconventional life paths?

The Thiel Fellowship and YC exist but they’re gated by elite selection. The real problem isn’t that people are less weird. It’s that the cost of reversibility is too high.

2/ My reframe: the demand for deviance hasn’t declined. We’ve lost the safe containers for it.

I hate horror movies, but they serve a function: you get to engage with the terrifying from your couch, knowing it ends in 90 minutes. Deviance with a seatbelt.

The Decline of Deviance Where has all the weirdness gone?

1/ Mastroianni’s data is solid: teens drink less, crime is down, people move less, culture is stagnant. But I think he’s diagnosing the symptom, not the cause.

His explanation: life is safer and richer, so we play it safe. We have more to lose.

www.experimental-history.com/p/the-declin...

Not all questions are truth-seeking—and that’s where generative AI actually fits.

LLMs aren’t search engines or oracles. They’re mirrors; sampling and reflecting the structure of our data back to us.

Their value isn’t consistent answers, but consistent structure across interpretations. Less about truth retrieval, more about sense-making.

arxiv.org/pdf/2509.06120

A worn O’Reilly-style parody book titled ‘Vibe Coding: I’m a Developer Now’ sits on a wooden desk next to a black pen, computer mouse, and keyboard. The cover shows a simple cartoon figure happily coding on a laptop with an Apple logo.

A cartoon scene from The Simpsons shows Lenny on a ladder replacing a sign that tracks “Days Since Cloudflare Outage.” He removes the number 17 and replaces it with a zero, while Carl steadies the ladder below. The joke is that Cloudflare outages reset the counter so frequently that it never stays high for long.

Buyout executive warns private equity push into US savings risks bailouts Move to widen access comes with high fees, average returns and ‘ample dollops of additional risk’, says Olympus Partners founder

Why do we keep outsourcing decision-making and risk-bearing to individuals least equipped for it?

on.ft.com/48GbhPt

Really feels like someone needs to release a model that ends with 6.7 before the opportunity is lost.

We Asked Roblox’s C.E.O. About Child Safety. It Got Tense.

The people with the dashboards and the power are still reluctant to talk about the costs of scale in plain, accountable language.

open.spotify.com/episode/5Q1i...

A two-panel meme labeled “HOW IT STARTED” on top and “HOW IT’S GOING” on the bottom. The top panel shows a news article headline: “Satya Nadella says as much as 30% of Microsoft code is written by AI,” with bullet points noting that Microsoft CEO Satya Nadella claimed 30% of company code is now AI-generated and that he discussed this during a conversation with Mark Zuckerberg. The bottom panel shows a different headline: “Microsoft finally admits almost all major Windows 11 core features are broken,” accompanied by text describing outages, performance issues, emergency hotfixes, and user backlash.

That’s why AI commerce built around answers is so appealing. A clean, truthful result feels almost revolutionary at this point. I just hope AI chats from frontier labs—OpenAI, Anthrophic—don't t fall into the same trap.

It’s wild how normalized that’s become. We treat the noise as discovery, when it’s really just revenue choreography. A search bar that pretends to help but mostly sells.

Search “Alo” on Amazon and you won’t actually get Alo.
You’ll get a dozen cheaper look-alikes, all “sponsored.”
The chaos isn’t bad UX — it’s the ad system doing its job. Every scroll, detour, and duplicate keeps the auction running.

Thoughts by a non-economist on AI and economics Crossposted on lesswrong Modern humans first emerged about 100,000 years ago. For the next 99,800 years or so, nothing happened. Well, not quite nothing. There were wars, political intrigue, the in…

AI progress isn’t linear. It’s doubling its ‘attention span’ every ~6 months. That means going from 2-hour tasks → 8-hour → 32-hour → whole projects in just a few years.

If that curve holds, real white-collar job loss could occur in the 2030s.

windowsontheory.org/2025/11/04/t...

From the LocalLLaMA community on Reddit: basketball players recognition with RF-DETR, SAM2, SigLIP and ResNet Explore this post and more from the LocalLLaMA community

Imagine the overlays you could leverage this for on Youtube TV: dynamic labels for players, refs, possession, shot distance, foul/violation hints, and +2/+3 markers.

www.reddit.com/r/LocalLLaMA...

Are games already a failed experiment for Netflix in your mind?

Yeah I guess, too, that OpenAI is the only company investing across the entire stack. They realized you can’t just pretrain your way to AGI anymore. You also need inference infrastructure, product infrastructure, deployment infrastructure. These all need to advance in parallel.