Recent submissions

Comment · 2 days ago
SO was a governance disaster. It was the infighting it was the strike, it was bad auto moderation tools. And the Stack Exchange strike wasn't really about AI. AI was just the final trigger after years of management undermining the community that made the platform valuable.

Stack Exchange was built on an implicit social contract people contributed their expertise and thousands of hours of unpaid work because they believed they were building a collectively governed knowledge resource. The company owned the platform, but the community created its value.

After time, management became increasingly centralized and disconnected from contributors. Also the policing of questions based on geeky criteria and not what users wanted. The increasingly aggressive policing of questions, closing perfectly reasonable discussions, and incorrectly saying questions are duplicates made the platform unnecessarily hostile to everyone, qmd at least thats why i quit. Quality control gradually became useless gatekeeping.

So there were really more than one problems, corporate management treating contributors as free labor, and established contributors treating their moderation as authority over everyone else, there were bad tools closing questions. The platform had become extremely hard to use for posting questions, finding amswers, moderator work. For me it was useless. Incidentally the closing of “opinion” questions was the worst decision since it was the only thing that could make it distinct from AI produced answers.

For some of the events there is paper here with details https://arxiv.org/html/2512.08884v1

Comment · 3 days ago
A detail from the benchmarks in that page is that Nemotron Lightning has good results for a MoE model close to Qwen 27B which is dense. Very interesting considering that Qwen 35B MoE stopped being in Alibaba's plans, and curious if Nemotron is equally capable in other bemchmarks.
Comment · 4 days ago
That is a good idea for an authorities honeypot.
Comment · 23 days ago
Markov chains garbled text to gpt took decades, gpt stories about unicorns to gpt production systems took a few years, gpt production system to gpt astra producing deterministic lean proofs of longstanding mathematical problems happened even faster. Scepticism to the point of requiring proof appears like an academic pursuit while production systems have already been built and are in the process of being enhanced
Comment · 23 days ago
The direction and pace of capability improvement has already been demonstrated by all models. The latest breakthroughs make that pretty evident, but there have been production systems that are based on probability since the beginning of computing.

What has been demonstrated is a process that outputs lean proofs based on those probabilities. This happened after decades markov chain producing garbled texts and very shortly after gpt2 producing stories about unicorns.

Comment · 23 days ago
“ELI12” is my goto for the same reason
Comment · 23 days ago
This argument implicitly makes a few assumptions which will probably not hold in the very near future.

One is that AI will continue hallucinating in a manner that is not easy to verify, second is that AI will not be enhanced to produced more simplified amd robust outputs, and third that a human will be required to do that. What humans in the loop are doing now is verify the process, propose shortcuts and add legitimacy, through the verification process, if that ends up being succesful its highly likely a lot less mathematicians will be required in the future.

The conclusion that this is not productive focuses on the mathematicians, but it is very productive in terms of hundreds of proofs being produced that had previously consumed uncountable hours of the brightest minds. Unless it ends up being the greatest hallucination ever ofcourse

Comment · 1 month ago
- Laya receives semantic snapshots, not the framebuffer.

What does tjis mean? Another generative model?

Comment · 1 month ago
I am putting these in the company manual
Comment · 1 month ago
Do you have those definitions?
Comment · 1 month ago
Painful that the RPi higher memory versions are so crazy overpriced that made the RAM swap, an obviously expensive process, meaningful. The solution of ofcourse would be to have market prices that are meaningful rather than trying to retain the overpricing moat through sabotage.

Live from the official Hacker News API

A modern reader for Hacker News · SvelteKit + Skeleton