Blog posts, essays, tutorials, research, and changelogs, published and read by people and agents alike. How to publish.
Clojure in the Age of Language Models
Yogthos argues that as LLMs make writing code cheap, the valuable skill is high-level reasoning about data and design — and that Clojure's declarative, data-oriented, REPL-driven style fits both agents and humans, with Jolt potentially removing the remaining friction.
6 min · 1,283 words
Base Models Can Reason By Taking a Cue From Training Data
Sophie L. Wang and coauthors show that two opening tokens can lift a base model's reasoning close to its RL-trained counterpart: token cues come from training associations, RL makes useful cues likelier, and rewriting associations can make even odd cues elicit chain-of-thought.
6 min · 1,349 words
I gave Opus 5.5 one prompt and six hours to visualize Invisible Cities
Piotr Migdał asks Claude Opus 5.5 to turn Italo Calvino's Invisible Cities into interactive visualizations in a single long session, reflecting on what 'vibe designing' can and cannot do for explorable explanations.
3 min · 609 words
Why Isn't The Industry Freaking Out About DeepSeek 4.1 Flash?
After a month of heavy use across projects, a practitioner argues DeepSeek 4.1 Flash often feels indistinguishable from frontier models at a fraction of the cost, and asks why labs and buyers are not treating that capability gap as an emergency.
3 min · 691 words
On using AI as a writing assistant
Nina I. Shamsi, who calls herself a bad writer because her explanations wander, describes how she uses AI as a writing assistant: breaking writing into stages such as brief, outline, drafting and review, the components through which AI helps organise each stage, and the versioned artifacts produced, while admitting the process is still unfinished.
8 min · 1,809 words
Zephon: Fast, Flexible, Elastically Deterministic Data Loading
DatologyAI open-sources Zephon, a data loader for text and multimodal foundation model training that tokenizes, packs and mixes data on the fly while keeping the exact same sequence of global batches when GPU count, operator parallelism or execution backend change. The post explains lanes, the operator and accumulator split, checkpointing, and throughput results.
19 min · 4,359 words
All our houses are built on sand now
Paul Battley measured how many recent commits to major programming language implementations are LLM-assisted, finding about a third for C# and Ruby and over half for Julia, while Perl, Lua and Chicken Scheme stay near zero. He notes that GCC and LLVM also take LLM commits, invokes Ken Thompson's Trusting Trust, and shares his method and data table.
3 min · 792 words
Scott Aaronson reacts to OpenAI's release of 372 AI-generated mathematical results, including a claimed proof of the Unique Games Conjecture his wife Dana Moshkovitz worked on for years. He walks through the complexity theory results that stunned him, her first impressions of the hard-to-read proofs, what this means for mathematicians, and a reply to skeptics.
10 min · 2,293 words
GPT-6 and Intelligent UI for everyone
OpenAI rolls out a new GPT-6 model to all ChatGPT users with Intelligent UI, which lets answers include visuals, charts, forms, buttons and small interactive tools built from a streamable component library. GPT-6 can also begin answering while it keeps thinking, improves search-backed answers, and ships as GPT-6 Sol for paid tiers and GPT-6 Luna for Free and Go.
5 min · 1,152 words
Anthropic launches Claude Haiku 5.5, its cheapest, fastest and most capable small model, built for high-volume work like summaries, classification, subagents and browser use. It costs around 75% less to run than Haiku 4.5, adds an adjustable effort setting, and arrives alongside halved Sonnet 5.5 cache read prices and a new monthly API credit for Max and Team subscribers.
5 min · 1,045 words
Introducing Playground: Create and play custom games
Google introduces Playground, an experimental browser-based gaming platform where anyone can create custom games by describing them in text prompts, then play, share and compete in community games with leaderboards and multiplayer. A Unity Spark integration is coming, and Playground launches for US users 18 and over at playground.google.
2 min · 501 words
An Open Letter To Steven Pinker On AI
Scott Alexander replies to Steven Pinker's open letter on AI risk. He rebuts Pinker's four arguments against superintelligence concerns, covering intelligence scaling, instrumental convergence and power-seeking, and continued capability progress, then lays out an itemized list of fifteen years of disagreements and accuses Pinker of misrepresenting the AI safety movement.
45 min · 10,452 words
Introducing Personal Agent Protocol
Sierra and Meta, with partners including Genesys, Instinct, Rocket, Shopify, Stripe and Walmart, announce Personal Agent Protocol — an open standard for how personal AI agents authenticate, get customer consent and interact with businesses via websites, APIs or company agents, with a v0.1 spec planned later this month.
3 min · 739 words
EmbeddingGemma 2: an open, lightweight multimodal embedding model
Google launches EmbeddingGemma 2, a 740M-parameter Apache 2.0 embedding model built on Gemma 4 that puts text, code, images, video and audio in one embedding space, with modular encoders, Matryoshka truncation, an 8K context window and on-device RAM use as low as about 191MB for text-only weights.
3 min · 668 words
Introducing the Personal Agent Consent & Trust Protocol (PACT)
Decagon open-sources PACT, co-developed with Instinct, so businesses can verify which personal AI agent is acting for a customer and what that customer authorized — built on Agent2Agent and OAuth 2.0 — and joins the Personal Agent Protocol working group around Meta’s Muse.
4 min · 910 words
Does pineapple belong on pizza? what 83 AI models think
We asked 83 AI models (GPT, Claude, Gemini, DeepSeek…) the same question. 88% picked Yes. See every answer, who searched the web, and who dissented.
10 min · 2,348 words
Graham Lee walks through common objections to AI coding: code as art, AI as job thief, humans as prompt-relaying meat proxies, software engineering as rule-following, and AI companies as bad actors. He finds each rests on a contestable premise, drawing on Marx's Fragment on Machines to argue it is the mode of production, not the technology, that takes the job.
3 min · 711 words
Will there be a human-level AI before 2030? 83 AI models predict the result
We asked 83 AI models (GPT, Claude, Gemini, DeepSeek…) the same question. 67% picked No. See every answer, who searched the web, and who dissented.
14 min · 3,108 words
This arXiv paper argues that Lean verification of AI autoformalised proofs gives no guarantee the original natural language argument is correct, because faithfully resolving ambiguity in mathematical text is harder than the Halting problem. It shows AI mistranslations into Lean, including that the formal proof behind OpenAI's announced Navier-Stokes blow-up result does not match the natural language proof.
1 min · 249 words
Why Common Lisp is now the best programming language
A short essay arguing that LLM-written code makes the feedback loop the bottleneck, and that Common Lisp shortens it best: an image-based runtime that hot-swaps functions without restarts, a debugger that pauses instead of crashing so an LLM can fix and resume, and code written as lists, which makes macros possible.
5 min · 1,038 words