GPT-6 Astra: The System Card, Alignment and What Comes Next
OpenAI claims that Astra is ‘the most intelligent and most aligned [available] model’ in the world.
Department
Tech.
Code, tools, and digital innovation.
Lately on the shelf
30 of 2637Culture Becomes a Dark Forest
“Wallfacing” to solve Navier-Stokes
When will average people feel AI’s impact?
We’re <5 years into a compounding revolution which could take a century, and how the AI industry should manage this.
Pluralistic: Anti-vax/anti-trust (09 Sep 2026)
Today's links Anti-vax/anti-trust: Remembering saves lives. Hey look at this: Delights to delectate. Object permanence: Disneyland prisoner; HOAs v xeriscaping; Copyright Office x regulatory capture; The women held a vote; Open licenses v uncopyrightable subje…
[Note]
Needed to renew my driving license. My new one doesn't have an EU flag on it. The sad reminders of Brexit just keep coming. 🇪🇺💔
[Article] Hot take: AI peaked with Taco Bot 3000
AI peaked in 2014 with the invention of the Taco Bot 3000. With artificial (taco) intelligence and the latest cheese dispensing technology, it truly represented the summit of human technological achievement the likes of which we might not see again.
Terence Tao on the paradox of solving mathematical problems
Mathematician Terence Tao, a 2006 Fields Medal winner and professor at UCLA, discusses… Tags: Big Think, math, paradox, Terence Tao
Quoting Terence Tao
I wrote recently about how the collection of good, fruitful open problems is now being mined in a non-renewable fashion, leading to the potential scenario of these problems becoming scarce. [...] We have now seen that even the rumor of someone working on a pro…
Why we should anthropomorphize AI agents
I Don’t Want to Interact With Stochastic Parrots
I Don’t Want to Interact With Stochastic Parrots My AI policy is "I don’t want any kind of interaction with that thing." And this should be all you need to know. I don’t care if you call it AI, LLM, chatbot, or whateverGPT. I simply don’t want to interact with…
Cancer Capital: It sucks for founders, too!
So, we've been breaking down the way venture capital has evolved over the last two decades or so (spoilers: it got worse!), but a lot of that has been kind of theoretical. Now it's time for us to talk about how that impacts players in the real world, at a prac…
On the Navier–Stokes Millennium Prize Problem
On the Navier–Stokes Millennium Prize Problem Impressive result from OpenAI, who used an unreleased model to produce a resolution to the Navier–Stokes existence and smoothness problem, one of the seven Millennium Prize Problems that have been subject to a $1,0…
Introducing ChatGPT Images 2.5
Introducing ChatGPT Images 2.5 OpenAI's image generation models are apparently used "more than 3 billion images across ChatGPT Images and the GPT‑Image models in the API". This latest release improves their instruction-following ability across multiple turns,…
Better AI code comment detector
When I trained the previous AI comment classifier, I used partially personal private data to do it, and built it on a somewhat shaky foundation, so I couldn’t share the code or data. I rebuilt it on public data and a better foundation! First off, you might wan…
Latest open artifacts (#24): Motif-3, GLM-5.3, Hy4-preview and open model licenses
The open model ecosystem continues to expand in its breadth
[Repost] Why Are We Taught to Memorise and Not Learn?
Çaki puts forward a political but fair look at the mass-production standardisation of English-language education in British schools. In an effort to make work measurable and objectively assessable, it argues, we're stripping the joy out of the art of writing.
AI's Mixed Messages • Siri AI's Potential Developer Problem
Do you feel like every other headline and talking point coming out of the companies at the AI frontier contradict each other? You're not alone there. The future is so bright... because of the nuclear blast that the rogue AI agents will trigger. By accident or…
Astra Is Hard to Monitor
OpenAI’s central message on Astra is that it is three things:
UN votes to support map projection that accurately shows the size of Africa
Last year, the African Union pushed for wider adoption of the Equal Area… Tags: Africa, area, cartography, Guardian, projections
Automatically detecting AI text in my browser
llm 0.35
Release: llm 0.35 New OpenAI model: gpt-6-astra for GPT-6 Astra. Tags: openai, llm, gpt-6-astra
Creepy crawlies
Creepy crawlies Konstantin Ryabitsev discusses how bad the "background radiation" of abusive crawlers has become from the perspective of git.kernel.org, the official Git repository for the Linux kernel: TL;DR: we spend more CPU cycles rendering commits for scr…
Quoting Jakub Pachocki
The strongest argument I see for continuing to train much smarter models quickly is the need to build defensive systems against the dangers posed by other AI. [...] We will need powerful, aligned AI for defense; to secure infrastructure, to protect against rog…
You Cannot Unsee 'You Can See Everything'
What did I just watch?
Recursion into madness
Raymond Chandler would have loved generative AI
Video compressor
Tool: Video compressor I recorded a short demo video of my Equal Earth animation on my phone and wanted to publish an optimized version of that video (using FFMPEG) on my blog, so I had Claude Fable 5.1 in Claude Code for web build me this tool using the WebAs…
An Alien Mind: Jakub Pachocki Warns Us
OpenAI Chief Scientist Jakub Pachocki is dropping truth bombs.
Who is Holding the App Store Line?
And is that line about to shift – in a good or bad way?
Mercator ↔ Equal Earth
Tool: Mercator ↔ Equal Earth I got curious about the Equal Earth map projection that was recently voted on at the UN so I had GPT-6 Astra (medium) in ChatGPT Work build me this animated transition between Mercator and Equal Earth using D3. Tags: geospatial, d3…
Import AI 472: DeepMind's cheating math agents; populist AI policies; and Forethought theorizes a nightwatchman
Plus, a machine hermeneutics story