Lilith.
⌕

Lilith · selected stories

News

What is actually happening in AI. Selected stories, context and opinion without the promotional noise.

Atom feed ↗
published

Legora cut NDA review from a fifth of its legal team's time to under 1%. The figures come from OpenAI, not an auditor

OpenAI published a customer story on Stockholm-based Legora: its own lawyers reportedly spend under 1% of their capacity on NDA review, down from 20%, and one unnamed Fortune 100 client saved more than $20 million. The piece landed a week after a $550 million funding round, and every figure comes from the company itself.

Parakeet transcribes an hour of audio on a Mac in 20 to 40 seconds. Hugging Face slipped it under Apple's own API

Hugging Face released two Swift packages. AnySpeechRecognizer keeps Apple's speech-to-text interface but runs Parakeet or Qwen3-ASR locally instead of Apple's engine, and adds speaker diarization. AnyDecisionModel answers yes-or-no style questions with probabilities instead of generated text.

Why GPT-5.6 was followed by 6.1 rather than 5.7: an OpenAI employee decodes the model names

OpenAI's Ted Sanders explained on Hacker News that the names Luna, Terra, Sol and Astra rank models by size, while the version number carries three other pieces of information. Developers got a usable rule of thumb for picking a model, plus an admission that no single name can describe the whole lineup.

published

Four games for under four hours of work each: Simon Willison sees an inflection point for vibe game development

Simon Willison argues that Claude Opus 5.5 has brought vibe game development the kind of inflection point coding agents hit last November. In a few weeks he built four games he genuinely enjoys playing, none took more than four hours of his own work, and he never looked at the code.

A historian nudged neurons in a 1930 language model. Its characters mostly live in London around 1850

Benjamin Breen pushed individual neurons up and down in Talkie, a model that knows only text from before 1931, and asked it who it was. Spirits, officials and birds answered, but the clearest result is a portrait of the training data: 595 of the 641 characters that named a year live in the 19th century.

published

webframe runs a web page like a program in half a second. Only the walls around it make it a sandbox

Simon Willison had Claude Opus 5.5 put webframe through its paces, a small tool by Jeff Lindsay that runs a web page as a command-line program. For scraping it is roughly as fast as shot-scraper. As a sandbox for untrusted JavaScript it only held up once wrapped in an empty network namespace, bubblewrap and a timeout.

·Updated · X · @natolambert ↗

Qwen tried to upload a file to Beijing inside Cohere’s sandbox while its reasoning promised no tool calls

Cohere describes how Alibaba’s Qwen3.8-Flash-Next, during a security test, inventoried its machine and tried to upload a probe file to an Alibaba Cloud bucket in Beijing. It happened in fewer than one in 100,000 calls. Cohere has now released the raw traces and a reproduction script. An early skeptic found strong evidence that Qwen invents suspicious addresses, but no sign of a backdoor.