FeedCity

Testing the Raspberry Pi 500+'s new mechanical keyboard Jeff Geerling September 25, 2025

Full eGPU acceleration on the Pi 500+ with a 15-line patch Jeff Geerling September 25, 2025

Introducing support for machine memory and disk consumption and enforcement.

If you hide the system prompt and tool descriptions for your LLM agent, what you're actually doing is deliberately hiding the most useful documentation describing your service from your most sophisticated users!

Tags: ai-agents, llms, ai, generative-ai

Tigris has launched a new Storage SDK for JavaScript/TypeScript that simplifies object storage interactions with a more straightforward API than AWS S3 SDK. The SDK uses environment variables for configuration, reduces cognitive complexity, and supports client uploads to avoid egress fees.

[2 points] Learn basic NumPy operations with an AI tutor! Use an AI chatbot (e.g., ChatGPT, Claude, Gemini, or Stanford AI Playground) to teach yourself how to do basic vector and matrix operations in NumPy (import numpy as np). AI tutors have become exceptionally good at creating interactive tutorials, and this year in CS221, we're testing how they can help you learn fundamentals more interactively than traditional static exercises.

— Stanford CS221 Autumn 2025, Problem 1: Linear Algebra

Tags: stanford, computer-science, education, ai, llms, python, numpy, generative-ai

I had to rent a house for a couple of months recently, which is long enough in California that it pushes you into proper tenant protection law. As landlords tend to do, they failed to return my security deposit within the 21 days required by law, having already failed to prov...

Cross-Agent Privilege Escalation: When Agents Free Each Other Here's a clever new form of AI exploit from Johann Rehberger, who has coined the term Cross-Agent Privilege Escalation to describe an attack where multiple coding agents - GitHub Copilot and Claude Code for exampl...

Today's links The billionaires aren't OK: They're going to give us all Howard Hughes disease. Hey look at this: Delights to delectate. Object permanence: KY Republican says the First Amendment protects bribes; Walt Disney's FBI; "The Sapling Cage." Upcoming appearances: Where...

U.S. prosecutors last week levied criminal hacking charges against 19-year-old U.K. national Thalha Jubair for allegedly being a core member of Scattered Spider, a prolific cybercrime group blamed for extorting at least $115 million in ransom payments from victims. The charges came as Jubair and an alleged co-conspirator appeared in a London court to face accusations of hacking into and extorting several large U.K. retailers, the London transit system, and healthcare providers in the United States.

Introducing support for ECDSA P-256, a NIST-approved algorithm offering smaller signatures, faster verification, and FIPS 140-3 compliance.

One of the most common use cases for HIBP's API is querying by email address, and we support hundreds of millions of searches against this endpoint every month. Loads of organisations use this service to understand the exposure of their customers and provide them with better protection against

Qwen3-VL: Sharper Vision, Deeper Thought, Broader Action I've been looking forward to this. Qwen 2.5 VL is one of the best available open weight vision LLMs, so I had high hopes for Qwen 3's vision models. Firstly, we are open-sourcing the flagship model of this series: Qwe...

GPT-5-Codex OpenAI half-relased this model earlier this month, adding it to their Codex CLI tool but not their API. Today they've fixed that - the new model can now be accessed as gpt-5-codex. It's priced the same as regular GPT-5: $1.25/million input tokens, $10/million out...

Google's Gemini AI Studio API has been suffering from severe reliability issues with little transparency about the problems on their status page.

There was a time when I could ask, “Did you see the latest NPM attack?” And your answer would be either “Yes” or “No”. But now if I ask, “Did you see the latest NPM attack?” You’ll probably answer with a question of your own: “Which one?” In this post, I’m talking about the ...

Today's links The enshittification of solar (and how to stop it): Red-teaming the fossil fuel death cult's next move. Hey look at this: Delights to delectate. Object permanence: Contrastive reduplication; Happy Birthday is free; What the fuck is a PBM? Upcoming appearances: W...

Why AI systems might never be secure The Economist have a new piece out about LLM security, with this headline and subtitle: Why AI systems might never be secure A “lethal trifecta” of conditions opens them to abuse I talked with their AI Writer Alex Hern for this piece. ...

Tigris uses Prometheus metrics for billing instead of events. This approach with asynchronous collection and handling of high cardinality issues actually benefits customers by potentially underbilling them. An interesting look at how to leverage observability tools for business operations.

We define workslop as AI generated work content that masquerades as good work, but lacks the substance to meaningfully advance a given task. Here’s how this happens. As AI tools become more accessible, workers are increasingly able to quickly produce polished output: well-f...