What linguists think everyone should know about language, according to a 2025 survey: 1) How and why do languages change? (61) 2) How do children acquire language? (58) 3) How many languages are there in the world? (51) 4) What are the benefits of learning foreign languages? (45) 5) What is language? (45) 6) What are the functions of language? (44) 7) What is the difference between spoken and written language? (43) 8) Does language influence thought? (42) 9) What is the difference between a language and a dialect? (41) 10) What differentiates human language from animal communication? (36) 11) What is linguistics? (34) 12) Are some languages better than others? (32) 13) Are sign languages real languages? (30) 14) How common is multilingualism in the world? (29) 15) What are the different parts of speech? (28) 16) What are the basic components of language? (28) 17) What is the relation between language and culture? (26) 18) What other languages are related to your first language(s)? (26) 19) What does it mean to know a language? (23) 20) What is the difference between learning first and subsequent languages? (21) 21) Is language change equal to language decay? (21) 22) In what ways do languages differ? (21) 23) What features are shared by (almost) all human languages? (21) 24) What does it mean to speak “correctly”? (20) 25) What is syntax? (20)

My general advice for using AI for any purpose is to ask yourself: "how do I make what I am doing even more ambitious?" You do this until you either hit a wall and then you know the shape of the frontier (useful!) or you find out that the AI can do more than you think (common!)

Our release process continues to evolve in response to feedback from the associated academic communities. In our view, these manuscripts are best understood as a form of raw research data. To convert this raw research data into scientific understanding, in this release we are experimenting with a new approach. We reached out to external experts and asked them to act as shepherds, organizing groups to understand the new results. They advised three different release strategies: For the math result, release the AI output, leaving researchers free to organize their own efforts to explain the result. For the computer science result, form an external group of collaborators to explain the result, while making the AI output publicly available for anyone to examine and build on. For the physics result, form an external group of collaborators and defer to them on the release of materials.

The odd situation with AI is that there is clearly a financial bubble, which may pop any day now, but the underlying technology is if anything *under* hyped. People just aren’t using it to its full potential. How many people do you know using Grok Bot or equivalent? Outside coding who uses agents to their full potential? How many cities have actually adopted autonomous vehicles to the extent SF and Austin have?

GPT-6.1 Terra: delivering fast and capable computer use at a lower price. GPT-6.1 Terra understands what’s on screen, takes action across apps, writes naturally, and is a more capable model for your whole workday with the efficiency to take on more of it.

it blows my mind how big part of the economy is: - "there was no software for this part so we patched it with more people" - "our customer doesn't know what to do because our software is bad so we patched it by hiring more people" it's probably like 10-40% of white collar?

An internet where every company has to pay big bucks to Anthropic and OpenAI every month just to avoid destruction by ever-present swarms of rogue AI agents is not nearly as economically valuable as the free internet we've enjoyed up to now.

An alternate data point. This has been my best year selling self-paced video courses, which has been 98% of my income since 2022. A great course reorients the learner in the world. It doesn't just give them new knowledge, it makes new things possible. Also, course creation is curation. It takes a spread-out graph of information, finds the most impactful nodes, and linearises it. This means courses are fast-tracks - they get you to where you need to go faster. So, in the AI world, what is worth learning? For me, that has meant finding techniques which are: - Harness-agnostic - Model-agnostic - Language-agnostic ...that benefit the agent and increase output quality. Because they're everything-agnostic, you can drop into any codebase and succeed. This is a huge space. I've explored a small part of it: software fundamentals, module design, increasing test quality. But there is SO much more there. User interfaces, company process, leadership, observability, product design. But you've got to make the right course. I wouldn't make my "Total TypeScript" course now (sales have dropped off a cliff). That one focuses too much on API's, learning language features, aspects which are easy to ask agents about. Instead, I'd make "TypeScript for AI", focusing on the actual things you need to know to make a productive TS codebase for agents (and it would rock, too). It boggles my mind when people say courses are dead. The hunger for information is very much alive. Go feed it.

Why big companies are struggling to move faster with AI. 1. Let’s have a meeting to get everyone aligned. 2. Simple solution but we need a PRD before anyone can do anything. 3. Someone uses AI to generate said PRD that is mostly nonsense. Most people never read it. 4. 2 weeks later someone says they are confused on what needs to be done. So another meeting takes place. 5. A new document is generated that inevitably no one will read. 6. I get escalated too, and try to figure out why this is complicated. Then I realize some people have opinions that make no sense. 7. I call this out, and end up still being told we need another meeting and a document. 8. If I was high enough up, I’d not allow any of this. When in fact I could have just built the thing in 2 hours myself. Some days I get so frustrated, that I wish I could just be okay with this so I’d live a much less stressful life.

We have the job. It's called "professor" or "lecturer". Historically these jobs were geared towards engaging in scholarship and passing on that scholarship,tradition,and the fruits that activity through teaching, lecture, and writing. The job predates the PhD, and there's a good argument that the research centricity of the PhD made a damn mess of it.

Wow, over 1,000 mathematicians have now joined the Association for Human Mathematics (AHM), committing to: > NOT publish AI generated math papers > NOT work with frontier labs > 654 of them went even further to NOT even use AI in their research, at all

Just realized we've pretty much maintained pace with the "28 days of improvements to Codex". We've had at least one solid improvement every day of the month. Let's make it official. Gonna ship one meaningful improvement to T3 Code every day for however long I'm capable :)

To be fair, I haven't shipped much since Opus 5.5 dropped. Short list of what I've done in that time: - ts-rust, a complete rewrite of Typescript compiler in Rust that is 3x faster than the Go version - lust-lang, an alternative toolchain for a subset of rust that compiles 100x faster - Slopalytics, the best place to see how models compare against popular benchmarks and usage - Slophouse (private), my personal analytics pipeline and platform for T3 Code usage - fleethouse (private), my internal Clickhouse that gets all of my data from my cliproxy - t3os (private for now), the start of my "perfect os for agents". Got 5 boxes running it already 😈 Got a couple small changes into T3 Code as well, such as: - Rewrite search to be 4x more accurate two-click repo clones to other machines (including .envs) - "Working" sidebar section to hide working threads - live guest thread sharing so others can watch and queue prompts 🤫 - watch_pr tool so agents sleep on CI/reviews and cut GitHub API calls by 90% while babysitting - interactive in-app visualizations rendered straight in chat threads - a rebuilt usage dashboard breaking down spend by cache tier, speed, and model - one-click "Update everything" across local and remote boxes - early compaction for idle Claude threads before prompt cache expires t3 thread CLI to start, read, and message threads from the terminal - PR diff badges that split line counts into code, tests, and docs - migrating T3 Code's own typechecking onto tsc-rs - purged over 50% of our test suite - (and roughly 370 other merged PRs) Pretty slow month. Hope I can pick up the pace for the rest of Q4!

From Jill Lepore's interview with Vanity Fair, in which she argues that Sam Altman is more powerful than Donald Trump, and that The Artificial State controls the United States, and the world.

We should rethink the role of human judgment in the review process. Human expert should have a role, but it's no longer about writing 1,500-word essays to summarize findings, check proofs, and so on. On top of that, any journal that forbids AI usage in the review process should reconsider their position.