FeedCity

Autoresearching Apple's "LLM in a Flash" to run Qwen 397B locally Here's a fascinating piece of research by Dan Woods, who managed to get a custom version of Qwen3.5-397B-A17B running at 5.5+ tokens/second on a 48GB MacBook Pro M3 Max despite that model taking up 209GB (120G...

I have an early adulthood trauma from struggling to understand consensus amidst a myriad of poor explanations. I am overcompensating for that by adding my own attempts to the fray. Today, I want to draw a series of pictures which could be helpful. You can see this post as a set of missing illustrations for Notes on Paxos, or, alternatively, you can view that post as a more formal narrative counter-part for the present one.

I know, I know, this sounds controversial. But hear me out. A few years ago a new trend in UI design emerged where related elements would appear more and more detached and unrelated to the things they are meant to point to. Here's a screenshot of Google Chrome circa 2010: ...

Release: datasette 1.0a26

Datasette now has a mechanism for assigning semantic column types. Built-in column types include url, email, and json, and plugins can register additional types using the new register_column_types() plugin hook.

Snowflake Cortex AI Escapes Sandbox and Executes Malware PromptArmor report on a prompt injection attack chain in Snowflake's Cortex Agent, now fixed. The attack started when a Cortex user asked the agent to review a GitHub repository that had a prompt injection attack hidde...

Why not benchmark with physical experiments?

Turns out everything DID NOT go offline somehow. Yay!

Great news—we’ve hit our (very modest) performance goals for the CPython JIT over a year early for macOS AArch64, and a few months early for x86_64 Linux. The 3.15 alpha JIT is about 11-12% faster on macOS AArch64 than the tail calling interpreter, and 5-6%faster than the standard interpreter on x86_64 Linux.

— Ken Jin, Python 3.15’s JIT is now back on track

Tags: python

OpenAI today: Introducing GPT‑5.4 mini and nano. These models join GPT-5.4 which was released two weeks ago. OpenAI's self-reported benchmarks show the new 5.4-nano out-performing their previous GPT-5 mini model when run at maximum reasoning effort. The new mini is also 2x f...

Release: llm 0.29

Adds support for OpenAI's new models gpt-5.4, gpt-5.4-mini, and gpt-5.4-nano.

If you do not understand the ticket, if you do not understand the solution, or if you do not understand the feedback on your PR, then your use of LLM is hurting Django as a whole. [...]

For a reviewer, it’s demoralizing to communicate with a facade of a human.

This is because contributing to open source, especially Django, is a communal endeavor. Removing your humanity from that experience makes that endeavor more difficult. If you use an LLM to contribute to Django, it needs to be as a complementary tool, not as your vehicle.

— Tim Schilling, Give Django your time and money, not your tokens

Tags: ai-ethics, open-source, generative-ai, ai, django, llms

Imagine I’m the design leader at your org and I present the following guidelines I want us to adopt as a team for doing design work: Typography: Use expressive, purposeful fonts and avoid default stacks (Inter, Roboto, Arial, system). Motion: Use a few meaningful animations...

Dang, guess I have to suffer.

Today's links William Gibson vs Margaret Thatcher: The Street Finds Its Own Alternatives For Things. Hey look at this: Delights to delectate. Object permanence: Prison for spamming; Dotcom layoffs; Ethernet action-figures; UK libel reform; "Poe's Detective"; God's customer se...

OR: rubberneckers anonymous

Your Startup Is Probably Dead On Arrival If you started a company more than two years ago, it’s likely that many of your assumptions are no longer true. You need to stop coding, building, recruiting, fund raising, etc., and take stock of what changed around you. Or your company will die. I just had coffee […]

Agentic Engineering Patterns > LLMs are restricted by their context limit - how many tokens they can fit in their working memory at any given time. These values have not increased much over the past two years even as the LLMs themselves have seen dramatic improvements...

A more obscure Commodore employee (where he apparently met his wife), but still quite influential during the Amiga era. Along with more pedestrian units such as the Amiga 3000 and the monochrome Amiga 2024 monitor, my favourite device he worked on was the 1987 prototype SX-5...

In the beginning, it was simple. A website, a database and 150M+ email addresses to search. Time has added serverless functions (which run on servers 🤷‍♂️), code on the edge, new data storage constructs and a completely different mechanism for even just querying a simple email address.

Introducing Mistral Small 4 Big new release from Mistral today (despite the name) - a new Apache 2 licensed 119B parameter (Mixture-of-Experts, 6B active) model which they describe like this: Mistral Small 4 is the first Mistral model to unify the capabilities of our flagsh...