Where the AI market is talking

Data centers raise nearby temperatures by up to 4 degrees in Phoenix

Even at today's demand Anthropic and OpenAI often return 'model overloaded' errors. Amodei, who used to research cancer biomarkers at Stanford, said this: >I think it will actually be possible to cure most human disease in ~5-10 years Pessimists are going to say he is lying or crazy. Given the trajectory of progress over the last year, I do not see where they get the confidence to say with certainty that AI will not change the world. If there is even a small chance for the biomedical breakthroughs mentioned, this to me justifies any data center buildout, a hundred times over, and I still would not need to think twice. The benefit to effort ratio is great already today, and the upside basically has no known limit.

Pacing model development in an era of cyber-critical capabilities

What a breath of fresh air. If 2026's Anthropic did an announcement like that, it'd be so many words it'd crash the browser.

Show HN: ChatOSS – A Codex alternative for Open Source AI built on Ollama

If you sign up for the $8/m Starter plan, you can use any OpenAI compatible keys. So your OpenRouter keys would work. The business model choice is what made sense to try out right now. It also makes it easy to get up and running with open source cloud models. If you build an application that runs on ChatOSS, you own the app, it doesn't belong to ChatOSS. You can do whatever you want. You could create an app for the platform that you charge a subscription fee for or offer for free or whatever you want.

Field measurements of neighborhood-scale air temperature impacts of data centers

I want to understand how his mind works. Not sure what value this is, now I think about it: I'm unlikely to time a bet against him well enough to win it. So, for this point, I don't care that the words and actions he emits have no apparent rational long-term coherency, but I do care about the coherency of the system behind them which generates them. Think of it like me wanting to learn how an LLM works so I don't get surprised by it doing more than a Markov chain.

Turbovec – Google's TurboQuant for vector search in Rust

Why not just use Qdrant? They've been integrating TurboQuant for months, works well.

Pacing model development in an era of cyber-critical capabilities

I suspect the labs are relying on frictions such as the models being extremely large (e.g. 2TB for a 2T parameter model, making exfiltration more difficult) and also not yet displaying any desire to survive or self-replicate beyond their immediate task (that we know of).

Launch HN: Speko (YC S26) – OpenRouter for Voice AI

not really counter-positioned, they are one of the providers we route to. elevenlabs sells models - voices, stt, and their agent platform on top of them. we don't sell any models: we measure all of them (elevenlabs included, their scribe is near the top of our stt board) and route each request to whatever wins for your language and constraints. so for them the best outcome is that you use their models; for us the best outcome is that you use the right one. sometimes that is the same thing.

Claude Code May–August 2026 weekly limits promotion

On the one hand it feels true. On the other I ask - what good software have anthropic, or anyone else, produced that was fully vibed? As someone who does near 100% of my coding via LLM these days, i still find that for anything complex i am still looking at and thinking in terms of code. Im still quality checking and steering at some interval via code. And im still not sure how or whether i can replicate that level of thought without still dealing in code at times.

Ask HN: I've quit six systems for tracking my illness. What works?

Yeah, maybe. The issue I foresee is more that the moat is quite small if you can just build it, so the moat would need to exist in stuff that can't just be built and replicated within a few weeks or something. User base maybe, connections to real resources, some other form of deeper value you can't just copy. Not sure though; I'm not a business person

Ask HN: I've quit six systems for tracking my illness. What works?

What do you need to track? I have a personal app I call Genomicly where I am, quite amateurishly, trying to replicate MaxGen panel and similar precision medicine with whoop, and LabCorp integration. My main goal is privacy which is tricky without the right hardware. The whoop is really easy to use from someone with ADHD and 20 years on a fire engine. I need simple.

Apple announces changes for apps in the European Union

I asked a question: "How do you not use their SDK?" Your answer seemed nonsensical to me. What I meant, though, is that I'm not going to argue about a specific point, the viability of "some FOSS SDK that perfectly replicates the original's functionality." If you don't even have any experience with iOS development, then you have no basis for making any claims about that subject, and thus I'm not going to bother explaining "What would be so outlandish about that?" In any case, most iOS devs don't want to opt out of using the SDK. That would not be a particularly attractive route for making iOS apps. What I want as an iOS dev is the same freedom of software distribution that we have on macOS.

Cursor launches Origin, GitHub alternative

git is such a good platform for decentralized tech. Forfeiting that advantage is just criminal. What I really want (and almost have) is a local GitHub for me and agents. With 4+ orchestrators working with a ton of subagents, things get difficult to coordinate. Several days ago, I went back to plain git and it was real slow and abrasive to work with. With some wizardry applied, I effectively have an "Excel for worktrees", a very productive and low-friction environment. https://replicated.live/blog/crdt

Apple announces changes for apps in the European Union

> I didn't mean to imply that this would be a vibe-coded slop project. That's not really the issue. LLM participation is a red herring. The issue is "replicates the original's functionality".

Apple announces changes for apps in the European Union

> Nowadays with LLMs people could probably patch up some FOSS SDK that perfectly replicates the original's functionality without infringing Apple's copyrights. I don't intend to insult you, but as an iOS developer myself I have serious doubts about whether you have any expertise in the subject you're talking about here.

Apple announces changes for apps in the European Union

Nowadays with LLMs people could probably patch up some FOSS SDK that perfectly replicates the original's functionality without infringing Apple's copyrights. If the DMA was enforced correctly this would have already happened.

Splitting a Git Commit

The bit that made this click for me: it works on any commit in history, not just HEAD. Mark the commit as edit in git rebase -i, then git reset HEAD^ to uncommit it while keeping the changes in the working tree, and rebuild the pieces from there. git add -p is the real workhorse for the rebuild: s splits a hunk into smaller ones, and e lets you hand-edit the hunk when the boundary doesn't fall on clean line breaks. Stage a coherent slice, git commit, repeat until the tree is empty, then git rebase --continue. The rebase's only job is to drop you at the right spot; add -p does the actual splitting.

GPT-4chan mirror hosted on Hugging Face behind gated access

Israel creates fake think tank in likely attempt to dupe AI chatbots

I am trying to understand your points. But claiming there is no direct nor potential relation here is a new way of thinking for me, and I am struggling to compartmentalize. If we have: 1. https://www.calcalistech.com/ctechnews/article/hya411vqezg 2. https://www.wsj.com/politics/israels-50-million-experiment-t... 3. https://www.wired.com/story/super-pac-backed-by-openai-and-p... We can see that at least some sources are covering the topic of using AI to influence the public on political topics, which regardless of any particularly nation, suggests there is an active interest in manipulating the epistemic integrity of AI systems for purposes of biasing outputs. And please allow me to make something very clear: I do not care wh…

Fairphone is now officially available in the United States

Is it pettiness or just sticking to their morals? Murena, as they mentioned, forwards data to OpenAI which is against pretty much everything Graphene stands for.

What We Learned Moving Our Agent Loops from Anthropic to GLM