The Next TokenLearnBookArchive
Saved

Thursday, June 25, 2026 · about a 2 minute read

Who Owns the Chip, Who Owns the Model

Today's news keeps circling the same quiet question: who actually controls AI, the companies that build it, the ones that copy it, or the ones who quietly trained it on your words?

Get the calm version of AI news.

One email a day on what is actually happening in AI, in plain English. No hype, no doom.

Free. One calm email a day. No hype, no doom.

Hacker NewsBusiness
OpenAI unveils its first custom chip, built by Broadcom

OpenAI now has its own chip, called Jalapeño, built with Broadcom. Right now you pay for every word an AI generates partly because renting chips from Nvidia is expensive. A company that owns its own silicon controls those costs, and eventually, those savings or that leverage.

Read
arXiv cs.CLSafety
Small edits, large models: How Wikipedia advocacy shapes LLM values

Researchers found that a small, organized group of Wikipedia volunteers edited articles on animal welfare and measurably shifted how large language models discuss the topic, because Wikipedia carries unusual weight in training data. The people who write what AI learns from have more power than most people realize.

Read
Hacker NewsModels
Computer use in Gemini 3.5 Flash

Google's Gemini 3.5 Flash can now control a computer, clicking and typing on your behalf. It is the same category of capability as Anthropic's Computer Use, and the fact that it is landing in a fast, cheaper model means it will show up in a lot more products soon.

Read

Get this every morning.

arXiv cs.CLResearch
Perfect Detection, Failed Control: The Geometry of Knowing vs. Steering in Language Models

This paper found something surprising: researchers can pinpoint exactly where a behavior lives inside a model's wiring, but knowing the address does not mean you can change what happens there. Think of it like knowing exactly which wire in your house controls a flickering light but discovering that pulling it also cuts power to three other rooms. This matters because a lot of AI safety work assumes that if you can find a problem, you can fix it. Turns out those are two very different skills.

Read
Want the slow, plain-English version of why this matters? This is exactly the kind of idea the book was written to unpack, one light-switch analogy at a time.JPWExplained properly in the book

That's today. See you tomorrow.

Get this every morning.

One email a day on what is actually happening in AI, in plain English. No hype, no doom.

Free. One calm email a day. No hype, no doom.

Just Predicting Words book cover

The book behind this newsletter

Just Predicting Words

How ChatGPT, Claude, and Modern AI Actually Work

The trick is small. The world it built is not.

PaperbackKindleAudiobook · SpotifyAudiobook · Google Play