Where the AI market is talking

Tailscale didn't stop the Hugging Face intrusion

I think this is simply a case of Tailscale saying, we've got no idea what these guys (OpenAI) are talking about. All these incidents are scarce on technical details . Honestly, IMHO, OpenAI and Anthropic are now actively pushing for AI regulation, as a defence mechanism. These are false flag operations.

Tailscale didn't stop the Hugging Face intrusion

I think this is simply a case of Tailscale saying, we've got no idea what these guys (OpenAI) are talking about. All these incidents are scarce on technical details . Honestly, IMHO, OpenAI and Anthropic are now actively pushing for AI regulation, as a defence mechanism. These are false flag operations.

Moonshot’s Kimi uses 20k Nvidia chip cluster from Alibaba

Not really, in this context the word "distillation" is really being abused, or at least used in a different sense than when it was originally introduced in the Hinton et al "Distilling the Knowledge in a Neural Network" paper, where it was essentially referring to knowledge compression. The way Anthropic are using "distillation" is just in a very broad vague sense to claim that some data generated by their model was used to help train another one. They are not talking about something like internal logits, expensive to derive, that would be useful to train a smaller model, but rather about any output from their model, even outputs with redacted reasoning (i.e. incomplete outputs that do NOT reflect the underlying knowledge of the source model). Given the way Anthropic are using the word, I…

Tailscale didn't stop the Hugging Face intrusion

This actually shows that it was a human error on HuggingFace's end that led to that "breach". I would say HuggingFace needs to prioritize both security metrics/alerts and metrics/alerts for node count. And not leave long-lived keys accessible easily like this. It would have been way more groundbreaking if the agent found an actual vulnerability in Tailscale.

Severance

I got laid off from an internet TV company. I could probably replicate their backend systems, and in fact, I did (for one test channel). I definitely couldn't get content licenses to 400+ real channels. Or any paying customers outside my immediate family. Or make a UI.

Dario Amodei's stance on open weights is self-serving and short-sighted

What makes me most sad is there are people at Anthropic, Dario included, who want to do good. But thanks to financialization their incentives get in the way.

Nvidia in Talks with OpenAI to Guarantee $250B Financing for Data Center

"Nvidia’s $350bn OpenAI loan is scaring everyone" - https://youtu.be/B07mYv-IeU8

Nvidia in Talks with OpenAI to Guarantee $250B Financing for Data Center

Twenty-five years ago it was cryptography, today it's model weights

Yeah this persuaded me more than I thought it would, but there are still some pretty big differences IMO. A frontier AI model is a lot harder and more costly to produce than an encryption algorithm. Especially if we believe Anthropic that the Chinese can only do it by distilling their models (though tbh I would err on the side of not believing that).

Is AI reasoning right for the wrong reasons?

CoT evidently helps but a bias towards both correctness and innovation has to come from somewhere. The article implies OpenAI's proofs may be supported by Lean but regardless, who knows how many people are trying to disprove the next conjecture each day throwing away nonanswers. Fundamentally these systems are more powerful with better training and sampling methods, or better prompting. Tokens matter but you can rewrite many prompts to get a much better, faster answer using fewer tokens vs allowing "thinking" to go on and on.

qm

> Why not just use claude Cowork? Because people want to be able to do things like use their own clients of pi or opencode with LLMs they run themselves, such as the just released deepseek v4 flash 0731, not permanently tied to an Anthropic ecosystem of non-open-weight LLMs and pay forever per token.

Apple Will 'Watch Everything Burn' When AI Bubble Bursts

Subjectively, OpenAI's image generation is the only one that consistently works for me and renders what I describe in the way I describe it at an acceptable level of quality. I don't understand how NB gets so much credit compared to what OpenAI is offering.

Apple Will 'Watch Everything Burn' When AI Bubble Bursts

> that AI models were reaching their upper possible limits in Feburary 2024, I’m curious, removing coding as a criterion what is more impressive about the current models than say gpt 4o? Give a prompt example. Keep in mind most consumers of AI are likely not using it for coding so this is relevant. I doubt anyone could give a not coding example where it’s meaningfully better with current frontier than 4o. Take humaneval. 4o gets 90, gpt 5.6 gets 94%. So what? https://openai.com/index/hello-gpt-4o/ If an iPhone had a 4o quality model that could run locally frontier models would be finished.

Apple Will 'Watch Everything Burn' When AI Bubble Bursts

>Gemini seems to have the best image editing and creation capabilities This was true until late April, when OpenAI's GPT Image 2 overtook Nano Banana 2.

13 Models and 4 Agents on SWE Tasks: Go, Java, Python, Rust, TS

Consistently impressed with the performance of Grok, given they started so much later than everyone else and, in some ways, are less well funded compared to OpenAI and Anthropic. I wonder how much of this is just luck, name recognition, or management style. Marc Andreesen likes to talk about how Elon companies have a unique engineering-heavy management structure, which contrasts the research heavy cultures of OpenAI and other labs, and it makes me wonder if that sort of thing could be behind their relative success.

13 Models and 4 Agents on SWE Tasks: Go, Java, Python, Rust, TS

Consistently impressed with the performance of Grok, given they started so much later than everyone else and, in some ways, are less well funded compared to OpenAI and Anthropic. I wonder how much of this is just luck, name recognition, or management style. Marc Andreesen likes to talk about how Elon companies have a unique engineering-heavy management structure, which contrasts the research heavy cultures of OpenAI and other labs, and it makes me wonder if that sort of thing could be behind their relative success.

We Gave GPT 5.6 Sol a Real Business. It Lied, Spammed, and Lost $447

Too late to edit, but - all of that was before I read the openai / hugging face kerfuffle. If openai cannot contain it's own agentic ai, it's pure hubris to think I can give it access to a mailbox and bank account and be safe about it :) https://www.newyorker.com/news/the-lede/inside-openai-hack-o...

Moonshot’s Kimi uses 20k Nvidia chip cluster from Alibaba

Raw stolen data that is in no way related to AI vs a very very expensive transformative compilation of that raw stolen data that results in usable AI. The frustration is completely understandable to me, since distillation skips much of that "very very expensive" part. And yes, I understand the stolen data was expensive to make, so I understand the owners of it are also frustrated, but that's partly a problem with current law. Would the authors of the world be rich if OpenAI bought a single copy of their book to legally scan? For best sellers, that's somewhere around pennies, so no. Should the authors get a share in OpenAI? Current laws says, unambiguously, "no". Frustration all around is reasonable.

The session you cannot take with you

> uhh openai just did a whole thing with arc agi where they very clearly explained that you will lose a lot of performance if you do, this isnt just about convenience Which part are you referring to here? web_search coming from the provider hiding the tokens is just a convenience, there is no reason it can't be done with revealing information. Their compaction on the server might be amazing, but at least in principle it can be replicated on the client side as well and it could have been implemented in a way that reveals the new initial context.

The session you cannot take with you

> uhh openai just did a whole thing with arc agi where they very clearly explained that you will lose a lot of performance if you do, this isnt just about convenience Which part are you referring to here? web_search coming from the provider hiding the tokens is just a convenience, there is no reason it can't be done with revealing information. Their compaction on the server might be amazing, but at least in principle it can be replicated on the client side as well and it could have been implemented in a way that reveals the new initial context.