FeedCity

The bot-fighting is a non-stop battle. In this week's video, I discuss how we're tweaking Cloudflare Turnstile and combining more attributes around how bot-like requests are, and... it almost worked. Just as I was preparing to write this intro, I found a small spike of anomalous

From rate limits to no limits: How IPv6's massive address space and a crafty botguard bypass left every Google user's phone number vulnerable

I don't think I've ever felt more tension in the room during a live performance than when I was standing five feet in front of Sly Stone, in a room that felt like a powder keg ready to blow. And I don't think I've ever seen a room feel greater relief than watching him tear i...

I don't think I've ever felt more tension in the room during a live performance than when I was standing five feet in front of Sly Stone, in a room that felt like a powder keg ready to blow. And I don't think I've ever seen a room feel greater relief than watching him tear i...

I don't think I've ever felt more tension in the room during a live performance than when I was standing five feet in front of Sly Stone, in a room that felt like a powder keg ready to blow. And I don't think I've ever seen a room feel greater relief than watching him tear i...

EDIT(2025-06-09 20:51 UTC): The containerization stuff they're using is open source on GitHub. Digging into it. Will post something else when I have something to say. This year's WWDC keynote was cool. They announced a redesign of the OSes, unified the versi...

Qwen3 Embedding New family of embedding models from Qwen, in three sizes: 0.6B, 4B, 8B - and two categories: Text Embedding and Text Reranking. The full collection can be browsed on Hugging Face. The smallest available model is the 0.6B Q8 one, which is available as a 639MB ...

It's been a long time coming, but we finally have some promising LLMs to try out which are trained entirely on openly licensed text! EleutherAI released the Pile four and a half years ago: "an 800GB dataset of diverse text for language modeling". It's been used as the basis ...

For [Natasha] Lyonne, the draw of AI isn’t speed or scale — it’s independence. “I’m not trying to run a tech company,” she told me. “It’s more that I’m a filmmaker who doesn’t want the tech people deciding the future of the medium.” She imagines a future in which indie film...

De hele korte versie: De kern van de Nederlandse overheid draait nu op eigen emailservers, in eigen beheer. Op 14 oktober 2025 verloopt de ondersteuning op de gebruikte software. Microsoft stimuleert sterk om dan alle email over te dragen aan de Amerikaanse cloud, met alle gevolgen van dien. Maar het is ook mogelijk (met wat moeite) om een nieuwe versie van de email software aan te schaffen, zoals 9 ministeries van plan zijn.

I presented an invited keynote at the AI Engineer World's Fair in San Francisco this week. This is my third time speaking at the event - here's my talks from October 2023 and June 2024. My topic this time was "The last six months in LLMs" - originally planned as the last yea...

Radek Sienkiewicz in a funny-because-its-true piece titled “Why do AI company logos look like buttholes?“: We made a circular shape [logo] with some angles because it looked nice, then wrote flowery language to justify why our…design is actually profound. As someone who ha...

Sipeed NanoCluster fits 7-node Pi cluster in 6cm

Sipeed's NanoCluster is a tiny compute module clusterboard with room for up to 7 tiny computers.

Each slot has two inline M.2 (NGFF) slots which accept either a custom-designed SoM (System on Module) or an adapter board to adapt a standard Compute Module form-factor board into the slot (as pictured above).

One end has a large fan for cooling, and the other end has power and IO. You can power the board via either PoE++ (60W) or through USB-C PD (65W). Or both, with redundant power fail-over.

Jeff Geerling June 6, 2025

Self-hosting your own media considered harmful

I just received my second community guidelines violation for my video demonstrating the use of LibreELEC on a Raspberry Pi 5, for 4K video playback.

I purposefully avoid demonstrating any of the tools (with a suffix that rhymes with "car") that are popularly used to circumvent purchasing movie, TV, and other media content, or any tools that automatically slurp up YouTube content.

Jeff Geerling June 5, 2025

Ukraine has seen nearly one-fifth of its Internet space come under Russian control or sold to Internet address brokers since February 2022, a new study finds. The analysis indicates large chunks of Ukrainian Internet address space are now in the hands of proxy and anonymity services nested at some of America's largest Internet service providers (ISPs).

gemini-2.5-pro-preview-06-05: Try the latest Gemini 2.5 Pro before general availability Announced on stage today by Logan Kilpatrick at the AI Engineer World’s Fair, who indicated that this will likely be the last in the Gemini 2.5 Pro series. The previous model ID was gemin...

Solomon Hykes just presented the best definition of an AI agent I've seen yet, on stage at the AI Engineer World's Fair:

An AI agent is an LLM wrecking its environment in a loop.

I collect AI agent definitions and I really like this how this one combines the currently popular "tools in a loop" one (see Anthropic) with the classic academic definition that I think dates back to at least the 90s:

An agent is something that acts in an environment; it does something. Agents include worms, dogs, thermostats, airplanes, robots, humans, companies, and countries.

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

what brewing temperature is best?

OpenAI slams court order to save all ChatGPT logs, including deleted chats This is very worrying. The New York Times v OpenAI lawsuit, now in its 17th month, includes accusations that OpenAI's models can output verbatim copies of New York Times content - both from training d...

As I wrote in my last post, Twitter's new encrypted DM infrastructure is pretty awful. But the amount of work required to make it somewhat better isn't large.When Juicebox is used with HSMs, it supports encrypting the communication between the client and the backend. This is ...