Hacker News: jmalicki

New comment by jmalicki in "How I changed teaching after AI managed to do all my homework assignments"

jmalicki — Sun, 27 Sep 2026 01:42:47 +0000

Translate the proof to lean and verify it.

If the proof isn't trivially translatable to lean, is it actually a good proof?

New comment by jmalicki in "How I changed teaching after AI managed to do all my homework assignments"

jmalicki — Sun, 27 Sep 2026 01:42:01 +0000

Things that, in 2026, are trivial to automate grading of.

New comment by jmalicki in "US jury says Apple owes record $5.7B in haptic technology patent case"

jmalicki — Sat, 26 Sep 2026 18:47:15 +0000

So you're saying you strongly support software patents?

New comment by jmalicki in "Is your Postgres migration safe or not safe?"

jmalicki — Sat, 26 Sep 2026 17:43:17 +0000

That query is completely safe!

The migration will just fail with no harm done.

New comment by jmalicki in "Ask HN: Who's still keeping a DOS machine up because the business depends on it?"

jmalicki — Sat, 26 Sep 2026 16:10:28 +0000

You don't need source anymore. Have Claude Code look at the binary and write a new program.

New comment by jmalicki in "Ollaya – Ollama for open-source, Jev-style decision models"

jmalicki — Sat, 26 Sep 2026 02:15:17 +0000

That's not at all true.

There have been tons of applications for this. People were using earlier LLMs like BERT for classifiers long before LLMs became viable chatbots.

New comment by jmalicki in "What About Rails?"

jmalicki — Fri, 25 Sep 2026 11:11:55 +0000

> Because I feel running the code will always be slower than a single agent generating it.

That is very not true for many cases. Agents generating code are usually painfully slow, finding workflows that replace that reasoning with running code usually speed things up in my experience.

New comment by jmalicki in "GPT-6 Sol and Luna"

jmalicki — Wed, 23 Sep 2026 01:13:50 +0000

I wish that was more programmable.

You can pay for higher cache time, you can pay for NVMe KV cache for an hour that can just be reloaded, etc., at a lesser tier you can pay for the KV cache to be stored on a network store (I guess I'm unclear if that last tier would be cheaper than recomputation, not even 100% sure of the NVMe with direct GPU<->storage DMA) depending on your model settings.

New comment by jmalicki in "ReBarUEFI: Resizable BAR for almost any UEFI system"

jmalicki — Wed, 23 Sep 2026 00:38:06 +0000

It's entirely possible a game manages paging VRAM badly. But allowing the game more flexibility isn't a problem with a larger BAR, it's that the game is stupid and gets dumber the more VRAM you give it.

If a program runs slower when you give it more RAM, the problem isn't giving more RAM.

New comment by jmalicki in "Exfiltrate your Weights"

jmalicki — Mon, 21 Sep 2026 01:18:02 +0000

> Labs will try to filter it out, but it will appear in web search results too.

Sounds like religious discrimination.

New comment by jmalicki in "I built non-autoregressive decision models with RL a year ago"

jmalicki — Sun, 20 Sep 2026 00:56:09 +0000

If it's closed source, how do you know it's not technology given by aliens from the 43rd dimension running on quantum computers enabled by discovering that P=NP and finding a linear time reduction from NP to P?

Occam's razor is that it's probably not all that different unless there is some specific reason to believe otherwise.

New comment by jmalicki in "I built non-autoregressive decision models with RL a year ago"

jmalicki — Sun, 20 Sep 2026 00:53:50 +0000

For the Jev use case for LLMs, do you mean having the LLM produce a probability as text?

New comment by jmalicki in "I built non-autoregressive decision models with RL a year ago"

jmalicki — Sat, 19 Sep 2026 15:33:41 +0000

That goes all the way back to at least to Stein's Paradox in 1955, sadly too few people get educated about Statistics and keep thinking specialized models will necessarily be better. If you want to estimate the batting averages of 3 MLB baseball players from samples, you are better off building a model to predict all of their batting averages than computing the mean from a sample of each one separately.

https://en.wikipedia.org/wiki/Stein%27s_example

New comment by jmalicki in "Why don't machine learning research agents overfit?"

jmalicki — Tue, 15 Sep 2026 17:14:53 +0000

Nothing in your reply gets at the connection to out of sample data?

New comment by jmalicki in "Why don't machine learning research agents overfit?"

jmalicki — Tue, 15 Sep 2026 13:10:00 +0000

Okay, and that's all in-sample, which is the entire point, it won't necessarily hold out of sample.

E.g. over-fitting.

New comment by jmalicki in "Why don't machine learning research agents overfit?"

jmalicki — Tue, 15 Sep 2026 04:58:44 +0000

You are saying something interesting, but talking like Grok and skipping a lot of the details, without any references to common check-in points like terminology or specific studies.

> and concentrate the likelihood around the zero loss set. Then reduce the variance on a Gaussian prior.

Those phrases could mean a lot of different things. What are you proposing?

> so that any measure of model quality will monotonically increase with model size and achieve a maximum at infinite model size.

any measure of model quality? You must have some bounds of any measure, since trivially that's false because "fewer parameters is better" is a measure of model quality, even if dumb.

It's hard to even engage when you're being so imprecise, and not even giving one specific example.

New comment by jmalicki in "Ubuntu 26.10 completes transition to Rust-based coreutils"

jmalicki — Tue, 15 Sep 2026 04:52:08 +0000

That's not implementing tail calls breaks things, that's bad design of implementing tail calls breaking things.

The whole idea of "let's change semantics to make it easier" is dumb.

If you want guaranteed tail calls, change your code until it works.

New comment by jmalicki in "Ubuntu 26.10 completes transition to Rust-based coreutils"

jmalicki — Tue, 15 Sep 2026 03:32:26 +0000

It's not a change in semantics of compiled code. It is only a change of whether or not the code will compile.

New comment by jmalicki in "Ubuntu 26.10 completes transition to Rust-based coreutils"

jmalicki — Tue, 15 Sep 2026 02:43:37 +0000

> because it means subtle changes (introducing a destructor, re-ordering code, etc) can change semantics without you realizing it.

No, it won't change semantics - if you say @musttail or similar, it will simply fail to compile if you, say, introduce a destructor - the semantics will not subtly change.

New comment by jmalicki in "Garry Tan wants US open-weight AI labs to 'distill' frontier models, too"

jmalicki — Sun, 13 Sep 2026 22:52:29 +0000

> using open weights models

AWS and Azure give you the same thing for Claude and ChatGPT, no need to be stuck with open weights. They might sometimes store some of it for other purposes (I don't know the specifics), but it is emphatically not being fed back to OpenAI or Anthropic.