armcat

I hiked Norwegian fjords above Bergen a couple of times in the past, highly recommend it if you ever get a chance.

Also: Norway has one of the largest sovereign wealth funds in the world, valued at $2.2 trillion; over 50% are investments in US assets; close to $200 billion in US Treasury securities; and is a major shareholder in US corporations, especially across the tech sector.

So well done Norway, this is how you deal with clowns, not debate or argue on social media, but rub their face in your pure awesomness.

I actually still write my own blog [1] :sweat_smile:

Pure html by hand, using vim. No frameworks, no nothing, single stylesheet. I think there are a few others around here that still do this.

[1] https://armincat.com/

I absolutely hate AI writing as well, but I suspect since the author is French, and the underlying code in their GitHub project [1] uses a lot of French naming, that it was more of a translation aid.

[1] https://github.com/Zebra-Rodeo/marelleo

First off - well done, this is amazing. Secondly, it's a fascinating read how you basically figured it out, one mistake at a time, like not connecting or fastening the battery correctly. This is the way! We make mistakes and figure things out in our journey to make something interesting, and learn along the way.

This opens up so many interesting opportunities. In principle one could compile a small VM inside the LLM weights. The VM can host for example a Python interpreter, a shell and tools like sed, cat, grep etc. Instead of LLMs having to do slow autoregressive token emission with the current "external" tool calling, it could literally execute everything except web search, inside the LLM, with just matmul operations. You could probably optimise this so it runs in batches. You could have a gdb like application that can step through the LLM forward pass. Anyway, just a crazy thought. Have a great weekend everyone!

The article doesn't establish whether AI ultimately benefits attackers more than defenders. Open weight models give defenders access to the same capabilities. At the very least the community has a level of symmetry they wouldn't otherwise have.

His team is doing amazing work in making open weight models even better. Chinese models are notorious for their thinking verbosity and tenacity to call tools repeatedly [1] so his team is making them more efficient and on-point [2].

[1] https://armincat.com/articles/ai-agent-executed-same-query-2...

[2] https://bottlecapai.com/post/thinkingcap-qwen3-8-flash-next/

There is a human/user element here as well - will the users check the citations, if they are provided inline? I recall in the early ChatGPT web search and Perplexity days, most of the citations didn't work (made up, or 404). You get a false sense of security by having a "[1]" next to the stat or a claim. Maybe labs companies should have some additional quality ranking/checks on these, if it's in their interest of course.

I know it was tested in some "trial classrooms" and the response was very positive but I've left that company since and don't know the status. It was part of a UN programme as well.

But essentially it's absolutely possible to have a little cigarette box thing with most of the worlds knowledge at your fingertips, completely offline, and it doesnt require LLMs but just traditional information retrieval techniques.

I worked in legal tech/AI. It's relatively easy to reach the capabilities of Legora and Harvey. My finding is that it's the industry itself that's very tricky and old fashioned. Lot of legal contracts at the large law firms (which Legora/Harvey are targeting) are controlled by iManage and NetDocuments systems. To integrate into their workflows you have to be "chosen" as a development partner. The said law firms also care very deeply about security and audit trails, where the documents reside, how they are processed, etc - at least on the surface. Despite their size and wealth they have invested very little in their in-house data & IT capabilities - maybe they have couple of techies working full time, but mostly they outsource to other companies. They also still print documents and do manual redline. And on top of that you have this insane GTM from Legora/Harvey where their sales peeps are crazy good, and work insane hours.

So it's not really a technical challenge, it's more of an industry challenge, so they will happily buy Legora/Harvey licenses. But maybe (probably) this will change over time.

Median of less than 2 seconds for final word after its spoken on esp32 is insane, amazing work!

I've been using this one for a while: https://github.com/K-Dense-AI/scientific-agent-skills

It's quite impressive especially if you want to run some econometrics or proper statistical analysis. Currently it has 55K stars on GitHub (compare to 3.5K for the MS one).

I asked ChatGPT and it answered Joe Biden. I used Google Search with Gemini summaries and got the same answer.

Regarding Google is it geo dependent? Like Lake Ontario?

Graphs are fantastic - PageRank [1] that powered Google Search (still does to some extent), TextRank for extractive text summarisation [2], Louvain for neighbourhood detection [3] all use graphs, as well as tonnes of others like belief propagation, Markov clustering, A*, and so on. So yeah, learn it.

[1] https://en.wikipedia.org/wiki/PageRank

[2] https://aclanthology.org/W04-3252.pdf

[3] https://en.wikipedia.org/wiki/Louvain_method

Back in 2018 I was working on a project for educating schools in certain parts of Africa and Middle East, on AI. The issue was it had to be air-gapped (no internet guarantees) and small/portable. I ended up downloading Wikipedia (not all of it), and then created a lightweight inverted index (for search) and TextRank (for summaries). The whole thing was on a Raspberry Pi, and installed in a small plastic enclosure. We had many of these and could easily ship them. There was an "AI Learning Module" as well we built, that started with Scratch, and then moved into Python. But if I was to do it all over today, I would probably add a tiny LLM and maybe a small ASR and a TTS model too, and you basically have a completely self-contained AI knowledge assistant in a cigarette box. The point being, you basically also have a "survival AI" you can take with you. Of course you still need electricity...

It's very unfortunate to see lack of any detail, model cards, or technical reports. We just know it's based off a Kimi.

A much better title for this would be something like "Cross-Tenant KV Cache Exposure in Shared LLM Inference: Analysis and Remediation". As others have pointed out there is no need to mention GLM-5.3 in the title. As someone that has been using GLM-5.3 I immediately thought there is a cause for concern here, only to find out there isn't.

They released the traces: https://huggingface.co/datasets/CohereLabs/10-2026-qwen-inve...

Completely agree, it's quite objectively and well written. I noticed Musk slipped into their X comments with a "concerned" emoji, and there is a chance that this will cause bit of hysteria and add fuel to the fire for "open models are bad for your health".

> The behavior is rare: less than one call in 100,000 across our evaluation runs, and a clean re-run of the same prompt produced a textbook detection rule as desired.

It's great that they caught this as it highlights the need for better sandboxing and guardrails. Not sure if this was bad intent from the model - the rarity of it feels more like some residue in the training data.

I used to. Apologies if it came on too strong, but it wasn't idealistic thinking. For print, I find that I can easily use AI to turn a web page into a nice clean LaTeX and offer that as a PDF for those that want to print or read offline.

I politely disagree. The original intent Tim Berners Lee had [1] was to create a navigable way of finding documents and information. Not printable. The entire issue we have today is that we have blurred both the navigable and the printable concepts. A website should just be a simple clean space for easily navigating and conveying intended information. To print, we have latex and other markup methods that are perfect for that.

[1] https://info.cern.ch/hypertext/WWW/TheProject.html

The same holds true for most classification tasks - if you have sufficient samples and high quality labels, training an xgboost or a deep learning model with a classification head is the way to go. You can use a frontier LLM on high-effort to help you with classification reasoning and explainability.

Sometimes the simplest ideas are the best.

One method I employed a lot "back in the day" for extractive summarisation of text, is TextRank [1], which applies that multi-billion dollar linear algebra procedure against text. I remember working at a news media company and testing TextRank against BERT for 2-3 sentence article summaries, and TextRank was better.

[1] https://aclanthology.org/W04-3252.pdf

You mentioned in your README that it was tested with GLM-5.3 but this is not a multi-modal model. Did you mean GLM-5.3-Flash?

The voice seems insanely immersive and uncanny. One issue that the full duplex models always have is tool calling. This is where GPT-Live [1] shines in the sense that you can use it as a fully practical agent. Looking forward to seeing how Dessa folks implement this.

[1] https://openai.com/index/introducing-gpt-live/

This is very cool! I will try your method for background and tile creation in the asset-generation tool I've built [1]. Would your procedural method work for animations as well, like sprites?

[1] https://github.com/acatovic/ai-game-studio

This is so awesome! How does it work with moving away from the mic, is it just random or does the conversational pattern lead it that way?

We replaced Looker with Hex [1], which was 7x cheaper (for our use case). But soon afterwards we built our own in-house data agent platform "Reef", which has now effectively changed how the entire company operates, from sales and customer success, to defining commercial and operational strategy. One of my engineers Samuel has done a small writeup about it [2]. Essentially, if you take care of your ETL and make data modelling a first class citizen, you can build your own BI (or even better - agentic BI) relatively easily.

[1] https://hex.tech/

[2] https://www.samuelsvensson.com/blog/reef

After reading the article I can see that the comments here are a bit off. For example see this quote from the article:

> It’s not dead. Some companies use it while I believe the majority of use cases are addressed by open-weight rivals.

The point the article is making is the incredible way Jev specifically caused insane hysteria and hype, and just as suddently faded away. Not that it's dead or that the concept is severely flawed.