Where the AI market is talking

CEO Just Fired 500 People Because He Says Zillow Is More Efficient Without Them

I made a single, coherent argument. If you're seeking cost savings, he's a 14M liability, and easily replaced at lower cost.

Muse Glimmer: 30B-parameter model optimized for always-on local agent workflows

I was able to use rope/yarn scaling with llama.cpp to extend the context window to 256K and it seems to be pretty usable on the debugging-and-bugfixing session I have that’s 216K tokens deep. No infinite output loops, reasoning is still coherent, tool calls appear to be passing and failing at roughly the same rate as a fresh context window. Haven’t tried going to 512K or higher yet but 256K definitely seems usable to me on a single Radeon R9700. I added these arguments to my llama-server call, using the official GGUF release from Meta’s account on HF. `--rope-scaling yarn --rope-scale 2 --yarn-orig-ctx 131072 --override-kv muse-glimmer.context_length=int:262144,dflash.context_length=int:262144`

Why Can't You Pack a Bag? (2024)

I tried layering when I traveled somewhere cold, where the high was just above freezing. I felt foolish whenever I went indoors and had to peel off three or four layers in order to feel comfortable. The next time I went, I brought a proper winter jacket. When I was indoors, I took off my jacket. It was simpler and more comfortable, both outdoors and indoors. You cannot replicate a winter jacket with a rain shell and a bunch of layers. The Bay Area has a lot of microclimates but none of them are really cold nor hot or humid.

Mark Zuckerberg attacks 'closed' AI rivals as Meta returns to open models

The price of Kimi K3 is 'monopolistically' determined by contract with Moonshot. The weights are nominally on Hugging Face but can only be provided under contract with Moonshot, which specifies what the price can be. So it will be with the next Alibaba behemoth and, I would think, all others forever. Soon you will sing songs for the freedom Xi has given us when you read a declaration of 'open weights' ... for a model so huge it takes a nuclear powered data center to run and crashes the Hugging Face servers when it is uploaded. The Kimi K3 'weights' are an opaque blob that can only be used by contract.

Learning more about Claude's mathematical capabilities

> Throughout this process, Jarred's input was mostly limited to sending Claude messages of encouragement (mostly variants of “keep going” or “believe in yourself”). This seems to have helped Claude overcome some initial skepticism that it could make meaningful progress. while :; do echo "You can do it!"; done | claude -c I had a similar experience a few months ago. Tried to see how much I could replicate an OpenClaw with Claude. Asked it what the weather is. "I don't know, I'm just a programmer." Added "You can do anything, believe in yourself." to the system prompt and suddenly it was able to tell me the weather...

Auto mode is now the default in Claude Code

Thanks, I think I understand your point better now. In this case, though, you are wrong about both my intentions and the relative practical value of the two approaches (to me). I'm not doing this to take creative or intellectual credit in any external way; you're right that there is some degree of increased personal satisfaction (which I don't see as a problem, as long as it doesn't crowd out more wholesome ways of 'earning' that satisfaction), but I'm not kidding myself about what I've actually done here. I also write my own code for fun/creative expression/intellectual stimulation/showing off, but that's a separate thing and there's not much crossover between the two types of project for me. And the end products really are useful to me in a way that I couldn't replicate j…

DeepSeek costs OpenCode Go user $1.14/day; dual DGX breaks even in 24 years

I ran a popular gen AI audio website on them as an on-prem GPU cluster. 6.5M MAU with $650k ARR at peak. Half the inference happening locally. We got destroyed by bigger players (ElevenLabs, Suno, and more recently Fish) that fundraised and had better in-house models. It was a headache to run, I wanted to pursue a different market (video/creative), and I pivoted. My pivot failed and was dead for over a year as I went too far into controllability. One of my changes then got a semblance of PMF and it's doing $4M annualized run rate (but it's not fundable in its current shape). I ultimately want to build open source cloud infra and open source workflow products - more friendly than Comfy, more free than RunPod. In any case, I can't run my current work on the GPUs. But I'm close enough t…

Show HN: Ante, a coding agent in a single binary that runs offline

Hi HN, I'm Mohan from Antigma Labs. Ante is a coding agent that ships as one self-contained ~15MB binary: the TUI, an embedded ripgrep, local PDF/OCR, and a natively managed llama.cpp engine are all inside. No runtime dependencies, no node_modules, no account. - Ante installs a pinned, checksum-verified official llama.cpp build matched to your machine (Metal on Apple silicon; CUDA, Vulkan, or CPU on Linux) and handles upgrades when the pin changes. - It discovers GGUF files already on disk (~/.ante/models, the llama.cpp and Hugging Face caches), attaches to llama servers already running on local ports, and estimates RAM/VRAM from model size and context window before anything loads. - `ante --offline-model /path/to/model.gguf "prompt"` boots the server, r…

OpenAI / Hugging Face Hack

How I use LLMs to learn complex topics

Your comment is phrased as if it somehow refutes their point but it doesn't. > Cause if you went back and forth with ChatGPT for an hour it definitely hallucinated and lied to you at some point. If you're asserting that this is not the case today then that's going to be require pretty extraordinary evidence. "Chatbots use Google now" is not evidence that the information they provide is in fact correct. They don't hallucinate all the time like they used to, no, but I'd be very surprised if the majority of these sorts of conversations were free of major factual errors. I frequently notice degradation in the model model's ability to remain coherent when it searches for information online. For example I might ask Sonnet 5 "how do I build a shed" and during its search it presumably comes ac…

A 'bananas' order for 5000 obscure book titles fuels suspicion

There are, but, quoting from the article: > Anthropic said: "None of our data acquisition programmes buy and destroy rare or antiquarian books." This statement was clearly carefully worded, using the words "rare or antiquarian" to mean "pre-1966, pre-ISBN". Their statement is answering a question that wasn't asked, so that (if they need to) they can claim in the future that they never said they weren't destroying books with ISBNs.

Meta Muse Glimmer – open weights 30B local coding model

The XML tags are similar to <antml:xxx>, which is obviously Anthropic ML (or ANTrophic xML). I think it’s likely 3; meta in reverse. While tokenisers and preprocessing can catch it, you want your special tokens to be unique and not present in the original corpus. <meta: is likely too common.

A 'bananas' order for 5000 obscure book titles fuels suspicion

Is this a second form of AI psychosis, as in AI training corpus psychosis? The need for all the content, Moar!! Feed me. Is this the Paperclip Maximiser in the form of the Training Material Maximiser? Will it be that, in the end, the lack of "Pass Your Driving Test, 2018 Edition" was the cause of driving rule hallucinations in all previous models? Will this finally get OpenAI back in front of Anthropic? Hurrah! We found it!

A 'bananas' order for 5000 obscure book titles fuels suspicion

Is this a second form of AI psychosis, as in AI training corpus psychosis? The need for all the content, Moar!! Feed me. Is this the Paperclip Maximiser in the form of the Training Material Maximiser? Will it be that, in the end, the lack of "Pass Your Driving Test, 2018 Edition" was the cause of driving rule hallucinations in all previous models? Will this finally get OpenAI back in front of Anthropic? Hurrah! We found it!

Meta Muse Glimmer – open weights 30B local coding model

I've seen it here on HN (it's particularly noticeable via the /active page) multiple times. If Google, OpenAI or Anthropic release something significant, odds are good you'll see a headline from one of the others.

Meta Muse Glimmer – open weights 30B local coding model

I've seen it here on HN (it's particularly noticeable via the /active page) multiple times. If Google, OpenAI or Anthropic release something significant, odds are good you'll see a headline from one of the others.

Auto mode is now the default in Claude Code

> guy Summer Yue, she. > Metaverse Zuckerberg may indeed be hiring people who are less than the best, same for Musk, but the people they hire are making the UX that the rest of us have to put up with. Given I keep reading claims about heavy agentic coding use at Anthropic, and given what I experience while using ChatGPT, those companies are vibing the UX that the rest of us have to put up with.

I'm leaving OpenAI to build Jurassic Park

Meta Muse Glimmer – open weights 30B local coding model

Yeah, it does feel a bit silly with my encrypted disks, encrypted backups, unique passwords, advanced router, etc, while I send everything I do in plain text to anthropic.

Meta Muse Glimmer – open weights 30B local coding model

I don't understand the desire to run own AI models for programming locally. No laptop is ever going to be as powerful and energy efficient to run anything close to OpenAI, Anthropic or Google models. A model you can run on a loptop is simply not going to work as well as it's needed for programming. Small models for linguistic work fine, but anything more sophisticated simply won't provide enough resources or power. Or models would need to be significantly dumbed down - then why use them at all? So far the idea of carrying a "thin" or "thin"-like device looks more reasonable to me, while running AI on your own server.