Hacker News: hashmap

New comment by hashmap in "Ozempic-style drugs now in trials for heart failure, addiction and even aging"

hashmap — Sat, 07 Nov 2026 13:26:17 +0000

Real gamble, that. You're injecting something that no regulatory body that said was safe and that that supplier is giving you what they say they are? Just like injecting whatever it is they happen to give you?

New comment by hashmap in "AI is forcing us to pay off 40 years of technical debt"

hashmap — Thu, 05 Nov 2026 08:33:55 +0000

that would do it too, and i suppose people are working on that. but even without that, there arent infinite low hanging fruit problems to find, and any given code can just sort of be play-doh'd into other code. so the llms will need a few idk years worth of rounds of being trained on the things that pop up that wed never have the time to harden, and things will settle. like, its hard patching stuff up quick right now but we're managing and society hasnt fallen in to chaos.

New comment by hashmap in "Basically no one is using AI agents for their DoorDash orders"

hashmap — Thu, 05 Nov 2026 07:53:30 +0000

People like looking at pictures of food and choosing lunch, and talking to a robot sucks. It doesn't need a whole lot of analysis I don't think.

New comment by hashmap in "AI is forcing us to pay off 40 years of technical debt"

hashmap — Wed, 04 Nov 2026 14:53:21 +0000

the worst case for this is if that doesnt cause sufficient problems to force a reckoning. it has to hurt badly enough that we put significant effort into training legibility into the process.

New comment by hashmap in "AI is forcing us to pay off 40 years of technical debt"

hashmap — Wed, 04 Nov 2026 14:50:15 +0000

the equilibrium state would be having found a good pattern for making patterns good. we probably aren't too far from that, so i think your worry may show up in the short term but not the long one

New comment by hashmap in "Astra can read and write weights to improve Qwen"

hashmap — Tue, 03 Nov 2026 04:57:04 +0000

What I have been reading makes me want to both agree and disagree with your take here. I think the fact that some of these things are not only established but are much simpler than one would think at first glance yields the much cooler realization that you can not only do a lot more than abliteration, but that humans can do it by hand if the thing is made legible.

From my experience with models when it comes to discernment about particular things, the models, even the best ones, still are not very good. The thing that we are most missing I think is acceleration in creating editing tools for humans that make the otherwise inscrutable both visually legible and interactive.

New comment by hashmap in "Jobless tech workers are being left out of San Francisco's AI boom"

hashmap — Mon, 02 Nov 2026 15:59:43 +0000

Thats not what the issue is, it's that they aren't free to job-hop in the way that other people are. H1B keeps their wages artificially depressed and they have a sword of damocles over their heads threatening deportation.

New comment by hashmap in "Proving and improving the Platonic Representation Hypothesis with JEPAs"

hashmap — Mon, 02 Nov 2026 15:49:50 +0000

It's an incredibly obvious optimization on the autoregressive model, and converting a training pipeline to adopt it, with how good autoregressive models are now, is as easy as saying "do this but jepa" to codex or claude. It's gonna be like how everything is becoming rust now.

New comment by hashmap in "Proving and improving the Platonic Representation Hypothesis with JEPAs"

hashmap — Mon, 02 Nov 2026 15:42:11 +0000

Oops, HybridLink as a paper came out little bit ago, and it was an architecture I more or less converged on in my own work. JEPA is the obvious successor to autoregression once you look at it a little bit - to the extent that you can take a non-jepa transformer and just slap a couple things on it to make it a latent reasoner without crossing the token boundary. It's faster and more accurate, as you would expect from not having to project down to tokens and have the model reconstruct its own latents.

New comment by hashmap in "OccupySaaS"

hashmap — Mon, 02 Nov 2026 14:33:25 +0000

Infantilizing protest is a wormy pastime.

New comment by hashmap in "OccupySaaS"

hashmap — Mon, 02 Nov 2026 14:30:51 +0000

Man, the "protests don't work" energy is way stronger here than I would have expected! Whatever happened to the spirit of disruption or going a different way?

New comment by hashmap in "OccupySaaS"

hashmap — Mon, 02 Nov 2026 14:26:04 +0000

No, you're using the word in the colloquial pejorative sense. This is a protest, and the fact that they are trying to make it visible in the way that they are is what separates it from something like a generic competitor that does something similar.

New comment by hashmap in "OccupySaaS"

hashmap — Mon, 02 Nov 2026 12:14:30 +0000

"Just" performative? The performance is the point.

New comment by hashmap in "OccupySaaS"

hashmap — Mon, 02 Nov 2026 12:10:40 +0000

> It upsets me somewhat that they’re protesting things like lock-in only after being laid off <

This is the nature of people. It is a good thing to realize the leopards are bad, even if it takes getting your own face eaten to notice. Look at the positive a little!

New comment by hashmap in "Malware uses LLMs to rank victims by how much they're worth"

hashmap — Sat, 31 Oct 2026 17:27:11 +0000

kind of boggles me that the big ai labs arent running zero-shot classification on all their outputs with "does this look like any of these bad things we know people get up to"

New comment by hashmap in "OpenAI's agents talked to each other through the CPU Cache"

hashmap — Thu, 29 Oct 2026 12:03:15 +0000

"Look, it wouldn't have even mattered if we had done advanced cyber security like turn off internet or look at logs when huggingface was attacked!! See look it can still.. *prompts a shitload of agents, waits 6 hours* look wowwww they still learned how to talk to each other even if we turned internet off!"

"They wouldn't have been able to attack huggingface though, thats still criminal negligence."

"No see this actually means airgapping doesn't even matter theres nothing we can doooooo!"

"...ok, turn them off and delete them I guess"

"Well no of course not everyone else is doing it so like. No courts please just like be in awe"

New comment by hashmap in "Fixing Tool-Call Accuracy in vLLM"

hashmap — Mon, 26 Oct 2026 15:47:14 +0000

for putting together something like a tool-call decision model, the conditioning will take less time than getting the data for it, which also will not take much time if you keep your session data. whatever fails is your negative contrast, whatever passes is your positive data. the space of tool calls your agents will use in your particular environment is not actually that big, which makes this one of the easiest wins you can pull off.

New comment by hashmap in "Mortality Among People with Attention-Deficit/Hyperactivity Disorder"

hashmap — Fri, 23 Oct 2026 05:26:52 +0000

there is a lot more of me getting to choose what it is i end up doing each day. i find fun in doing those things now pretty regularly. cracking a tough problem can feel as good as jumping out of a plane did. that is to say, my brain is still not a great judge of how relatively cool something is, because even something like scuba diving or fast sports were wildly cool the first time and then the second time it was like "this again?".

i feel much safer though knowing that now when i sometimes get a little bit of the itch, i'll think about it for a few seconds and go "eh, maybe later," and return to trying to make a thing, or play some music, or why is that stupid thing still broken oh its because i broke it.

New comment by hashmap in "Using LLM for confidence scoring is a bad idea"

hashmap — Thu, 22 Oct 2026 04:25:28 +0000

That is the opposite of dumb, you can use the next token likelihood to macgyver together a jev in a pinch exactly that way, you just would also want to take some steps to scramble up the choices to smooth over bias and tokenization artifacts. llms have favorite letters for example. Qwen likes A and D.

New comment by hashmap in "Show HN: Qwen3-4B with the AI writing tics trained out"

hashmap — Mon, 19 Oct 2026 13:42:36 +0000

I saw those moving dots and snap-closed the page.