Where the AI market is talking

Unsealed Briefs in Authors’ Case v. Microsoft/OpenAI

> Imagine alternative time line where tech is the same but each creative provides their own trained model that scales their own creative vision and skills, and keeps them in control of their own work. This is a timeline that can not and will never exist if the Authors Guild and most of the other anti-AI lawsuits succeed. Because if they do succeed, the only people who will be able to provide an AI model will be companies with enough resources to license all the training data in perpetuity. The Authors Guild isn't mad because the authors can't train and use their own AIs, they're mad because the AI companies are making money and they're not getting what they perceive as their fair cut of that money. > Imagine OpenAI, Anthropic &co having to compete by hiring [thousands] of their own …

When did Google get so weird?

Eliza would not remotely pass the turing test as turing envisioned in his paper. And frontier labs are not optimizing for 'convincingly human'. If Open AI or Anthropic wanted, frontier models could pass much more human like.

When did Google get so weird?

LLMs are fundamentally predicting the next word to make coherent text. If you've ever played with a Markov chain text generator you've done this with a fairly dumb predictor that maintains coherence over a very short distance. Deep transformer neutral networks can do it with a much longer coherence distance but they are fundamentally performing the same operation. After "Question: Did the team make the playoffs? Answer:" a reasonable completion is "yes, the team made the playoffs". An early demonstration of GPT-2 was a fake news article about scientists discovering unicorns in Antarctica - the model doesn't "know" whether or not unicorns exist in Antarctica, but it's able to complete "Breaking news! Scientists have discovered a colony of English-speaking unicorns in Antarctica." by adding…

Show HN: Has Anthropic been nerfing their models without disclosure

There is more to code review than (automatable) detection

code reviews are just a gateway that can be whatever you want it to be, and is kind of legacy human coder thing now. At its basics it was a point to catch problems that humans were likely to make / would more likely make if they knew there wasn't a review. Now you can target it for AI mistakes. You can build your code review skills (AI skill) to be incredibly thorough. The points made in the article don't really seem like things you need to do at the "legacy" gateway of code review. Things are different now. Code is cheap. Validation, Product Coherence, Governance need to be done early and throughout.

Did Anthropic's A.I. Really Make a Scientific Discovery on Its Own?

I will say that of the people I know in pharma, they are not allowed to use Anthropic nor OpenAI directly. It's all AWS Bedrock access, whom they trust not to use the data. (Because they already trust AWS with all their data, and AWS is not putting out papers in their fields of work.)

Unsealed Briefs in Authors’ Case v. Microsoft/OpenAI

"The fact that they believed this is legally useful because of the definition of fair use, so I see why the Author's Guild is emphasizing it, but the authors I know don't talk about it." Perhaps "the fact they believed this [i.e., the infringement was intentional] is legally useful" for seeking statutory damages, up to $15,000 per infringed work (assuming registration was timely, otherwise up to $7,500), as this requires the plaintiffs to show defendants acted willfully Without intent, statutory damages could be as low as $200 per work Anthropic settled for $1.5B with Bartz et al. in an earlier case involving the same pirated copies. The Court in the Bartz case said that training LLMs with these pirated copies was not fair use https://www.npr.org/2025/09/05/n…

An agent used DNS to reach an external chatbot

No, I'm not implying that. I'm suggesting that many more of these issues are likely to come from OpenAI due to their culture (not that that is okay). Anthropic has no excuse, I know they know better. One might argue that they were initially negligent because they didn't really believe it would do anything like this. But having further incidents is pure negligence at best.

When did Google get so weird?

> This is similar to how, not too long ago, LLM's had extreme difficulty counting the number of letters in some words. The specific issue of Google is that they are using an underpowered model, not fit to task, and much prone to hallucination than either OpenAI or Anthropic free tier offerings. Google should at least match the frontier labs at the free tier (with some limit; after that, degrade quality), ffs

There are no "rogue" AI agents

I don’t buy them being surprised. This perfectly plays into their pattern of using fear to make their models seem more powerful than they are. They obviously had the technical expertise… there isn’t a damn thing a bunch of randos on some HN thread knew about that model that they didn’t. It gave them an excuse to delay their IPO when their books seem like they’re going to be pretty shit compared to Anthropic. It helps them reposition themselves as being more safety-forward which the market is clearly more interested in. All that is to say they had motive out the ass, easily had the knowledge and capability to avoid the problem, knew better than anybody else what the models were capable of, set up the environment, gave it the prompt, and then did not even monitor the output. Negligence is c…

Anthropic is a supply chain risk for all of us

Not only Anthropic but any LLM that refuses to produce an answer because of some objection. I know that no company want to be associated with a headline that their model was used to {something}. Instead of helping us better learn something, we're heading (in fast pace) to the world where we won't be free to think outside allowed topics. Example: I asked some LLM to give me the spell (from a certain religion) to bring justice to someone. It refused as it would be "too powerful and could be use to cause damage".

There are no "rogue" AI agents

uh who in this discussion is in charge of policy? Take a look at what actual people in government are saying - see [1] (the parent article mentions Bernie Sanders being most high profile in response as well). you'll see there is no prosecution in there, only talk about the "freeze" that Anthropic / OpenAI are salivating for as it would close out all of their competitors. Bernie is fully hoodwinked. The argument for prosecution is actively hindered by this language of "rogue agents" [2]: > The incident is remarkable not just as a cybersecurity breach, but as a legal stress test. The Computer Fraud and Abuse Act (CFAA), the primary federal statute governing unauthorized computer access, was written decades ago with human intruders in mind.[4] Its key provisions require intentional o…

When did Google get so weird?

Had to try it. And it delivers :) > hes never coming over Dario Amodei Anthropic CEO Dario Amodei will actually be joining President Donald Trump for a private dinner at the White House tonight, Sunday, September 27, 2026. CNBC While it is true that he was noticeably absent from the high-profile state dinner held for Chinese President Xi Jinping earlier this week—which was attended by other major tech figures like Elon Musk and Sam Altman—

How I changed teaching after AI managed to do all my homework assignments

It's science-based in a sense, but a pretty weak sense. I did some work with education researchers and my impression was that their methodologies were quite haphazard and they were sometimes just trying to find a way to "tell a story" that they already had in mind. Also, my impression is that a lot of the research on "best practices" is based on relatively limited comparisons between schools or teaching methods, and there is insufficient evidence that this or that is truly a "best practice" in a more general sense. I also interacted a good bit with education grad students when I was in grad school and they almost uniformly seemed to have no coherent sense of what they were doing or what their field actually knew. I've known people who had similar experiences with education schools elsewhe…

Biology might not be quantum, but its math is quantumlike

The real question is whether biology exploits a subset of special quantum phenomena (entanglement, coherence, superposition). Just about everybody in the field acknowledges that biology, being made of chemistry and physics, is "quantum", but in a "boring" way (just the energy calculations), but the bar to convincingly demonstrate exploitation of special phenomena is much higher.

DSPy – Program, don't prompt, your LLMs

is this pretty close to langchain / ADK? this looks to generate prompts with words/instructions not provided by the user, not keen on that concept

Biology might not be quantum, but its math is quantumlike

Apparently that finding was itself found to be incorrect. The article doesn’t specifically debunk coherence as the mechanism behind photosynthesis, but it does imply that the real mechanism was discovered to be something else.

Unsealed Briefs in Authors’ Case v. Microsoft/OpenAI

not that I know of. would love to. maybe simply replicating the HN content at first but witg better moderation (still think /. moderation scheme is thr gold standard, and the reason it's good and effective is why its never replicated)

Show HN: TinyAIArena watch AI agents battle it out

I’m not sure what good output in this case would look like. I’d expect “good prose” in a book’s fight scene to be action oriented and show something about the character or reflect something about the narrative, so, it should often not involve talking. I guess realistic fights are just a lot of shouting and semi-coherent thoughts (because people don’t have time to come up with a full clever quip in that sort of situation).

Tells of a Slop UI

The difference here is that the slop paradigm is not a set of intentionally designed UX choices that work well for users, it's a set of accidental commonplace LLM tendencies, not designed for humans, replicated ad infinitum that are making user experiences worse in 99% of cases. The use of white space is horrendous, the font choices are illegible, the color palettes lack sufficient contrast.