News for hackers. Stuff that matters.

binarymax

Re:Ask HN: What hardware are you using to run local LLMs? on Monday November 02, 2026 @05:18PM (#50344730)
OK, that gpu-unlock is something I've never seen before and...wow. I did some research and NVidia really did kneecap their RTX for market segmentation. Grrrrr
Re:Ask HN: What hardware are you using to run local LLMs? on Monday November 02, 2026 @09:52AM (#50337609)
I have a 3090 w/ 24G VRAM, it supports flash attention and bf16, but is a bit old. I dont run generative models on it as they would be slow and inaccurate, but I do run embedding models, OCR, and system one (decision) models on it.
Well, technically yes, ethically no.
Doesn't matter. The entire thing is vibe coded, so you can just point Claude at it and say "Copy this".
Whether or not you agree or disagree with the merits of USAID is beside the point. A single unelected and unqualified person must not be given the authority to dismantle a massive US government program.
Re:Anthropic S-1 on Monday October 26, 2026 @09:34AM (#50243428)
In the simplest calculation, if you get 200M people on a $200/mo pro max plan, that's $500B annual revenue.

This is a "utility" mindset. If you bet that Anthropic is just like a monthly electric bill, then it's within the realm of possibility.

For the record - I don't think that's gonna happen.

lol that was a fun typo. Fixed :)
12% of annual revenue (~200B for 2025). Not a small amount.
Same! Here's the short film of Paddle to the Sea produced by NFB of Canada - We used to watch this in grade school in the 80s. https://www.youtube.com/watch?v=uhjb1IG1pnQ

I loved it. Still watch it every couple years.

Re:Why do LLM pelicans look like this? on Wednesday October 14, 2026 @01:09PM (#50091088)
I got downvoted[0] heavily for even mentioning the possibility. But I'll say that even if it isn't, there are plenty of pelicans in the training set for new models. People saying "well then why does it also work for flamingos???" don't understand how LLMs work.

[0] https://news.ycombinator.com/item?id=49526966

Re:Anthropic is out with new nonsense about Claude on Sunday October 11, 2026 @03:04PM (#50046666)
It’s wild that this got flagged. There are some really important points here that deserves a discussion.
Re:Orkut.com on Thursday October 08, 2026 @08:20PM (#50014319)
I agree. I see this announcement and think, “why?”

The social media experiment is over. Time to go back to meeting with real people.

I don’t need to connect with someone 5000km away with the same hobbies as me. I can do that in my town and have an actual, real connection. I don’t need to wonder if someone is a bot, and I don’t need to hand over an ID to prove I’m not.

Let’s go back outside.

Re:Theranos.World on Thursday October 08, 2026 @04:48PM (#50012030)
Burying the api key in notes on the laptop is a nice touch.
Re:Anti-Patterns in Software Blogging on Wednesday October 07, 2026 @03:35PM (#49997742)
People are still reading. At least I am :)
If I'm getting citations, and the distillation has strong alignment with citations, then it's fine. In other words: the if the summary is a highlight and not a rewrite then I'm happy. I've studied this extensively and have even created metrics around this problem: https://maxirwin.com/articles/llm-rag/
It’s not a blanket ‘Yeah’. It’s a ‘sometimes’ and that sometimes depends on use. Distilling down a whole days worth of research is one of the best uses of AI.
Re:Web Search API on Monday October 05, 2026 @07:38AM (#49963549)
Not sure where the trust claim lands, but the first two are now exceedingly trivial with code agents. A little more work perhaps, but not hard at all. I’ve done this myself (not with those providers) with several search platforms.
Re:Web Search API on Monday October 05, 2026 @07:32AM (#49963500)
Why not use those providers directly? Does Cloudflare need to be in the middle of everything?
Are search summaries slop?
There is the mysterious Ballmer peak after all.
Re:There are no "rogue" AI agents on Sunday September 27, 2026 @12:55PM (#49868444)
Exactly this. At worst, OpenAI knew about these behaviors and should be prosecuted under CFAA. At best, OpenAI is negligent and should be prosecuted for negligence.

Luckily there are states and legal departments pursuing such action. So while OpenAI can deflect as much as it wants, that doesn't mean there aren't people who know better and will still do what is necessary to set precedent.

If you read the first couple paragraphs it’s very self explanatory
Re:OpenAI agents carried out an undisclosed attack on RubyGems on Friday September 11, 2026 @10:20PM (#49668055)
3. They're just doing this shit on purpose.
Re:So you want to use OpenRouter? on Friday September 11, 2026 @10:50AM (#49659456)
Pretty damning. We're using OpenRouter for some research tasks and it's making me question everything. I suspect lots of the model providers are running into issues like these: https://forum.level1techs.com/t/why-your-local-llm-feels-dum... (HN discussion: https://news.ycombinator.com/item?id=49402232)
Re:Stockfish 19 on Thursday September 10, 2026 @08:39PM (#49652076)
Is doing something you love, and practicing to the point of expertise, wasting your life? Replace chess with playing an instrument, or basketball. This quote is just exclusionary.
Re:This Month in Ladybird – August 2026 on Monday September 07, 2026 @04:18PM (#49602557)
Really excited for alpha. When it hits I plan on using it for my daily driver to help them with usage stats and crash reports.
I can’t tel if you’re being sarcastic or not.

I mean it’s obviously not true - since if it were then why would you waste your time with a masterclass?

So the sarcasm bit is I don’t know if you’re parroting all the other masterclass scammers out there for lols or if you actually are a masterclass scammer.

Re:Claude Fable 5.1 and Claude Mythos 5.1 on Tuesday September 01, 2026 @03:36PM (#49526966)
Do you think model trainers are pelicanmaxxing now?
Re:Restroom Archive on Tuesday September 01, 2026 @10:56AM (#49522832)
Wow this is amazing! I had a habit of visiting the nicest bathroom I can find (typically 5 star hotel lobbies) while traveling. I took some photos of a couple a long way back, but never even thought to use something like LIDAR. I was thinking of using them as a coffee table book with the title "The Best Places To Go".

So from a fellow public bathroom enthusiast, I wholeheartedly thank you for this project. I travel much less these days, but if I'm ever in a nice bathroom I'll see if I can get a scan to work and upload it!

Older »