Hacker News: nmca

New comment by nmca in "There is no comfortable reading position"

nmca — Mon, 19 Jan 2026 19:56:54 +0000

I put a 5 dollar phone-holder-ring on the back of my kindle, problem solved.

New comment by nmca in "Autism's confusing cousins"

nmca — Sat, 06 Dec 2025 22:44:19 +0000

In a maximally earnest way — to what degree should we be sure these language harms are real on net? Are there data as opposed to anecdata? (Of course many phenomena are real without data; we can just be more confident in cases with data)

New comment by nmca in "Weighting an average to minimize variance"

nmca — Sun, 16 Nov 2025 06:03:30 +0000

This is also a nice way to combine the ratings of a number of noisy annotators with variable annotations noise.

New comment by nmca in "Meta Superintelligence Labs' first paper is about RAG"

nmca — Sun, 12 Oct 2025 02:12:17 +0000

This is not work by any of the high profile new hires, in case folks are confused.

New comment by nmca in "Why We Think"

nmca — Sat, 27 Sep 2025 21:49:47 +0000

Lillian’s blog is extremely good in general & if it’s new to you I suggest checking out the other posts also. I particularly enjoyed the one on human data.

New comment by nmca in "Code formatting comes to uv experimentally"

nmca — Fri, 22 Aug 2025 00:18:44 +0000

right, but you definitely shouldn’t be using any other formatter than ruff and this helps with that

New comment by nmca in "Ask HN: How can ChatGPT serve 700M users when I can't run one GPT-4 locally?"

nmca — Fri, 08 Aug 2025 21:07:22 +0000

It’s the same principle as:

https://www.tripadvisor.com/Restaurant_Review-g60763-d477541...

New comment by nmca in "Ask HN: How can ChatGPT serve 700M users when I can't run one GPT-4 locally?"

nmca — Fri, 08 Aug 2025 21:05:33 +0000

Isn’t the answer to the question just classic economies of scale?

You can’t run GPT4 for yourself because the fixed costs are high. But the variable costs are low, so OAI can serve a shit ton.

Or equivalently the smallest available unit of “serving a gpt4” is more gpt4 than one person needs.

I think all the inference optimisation answers are plain wrong for the actual question asked?

New comment by nmca in "200k Flemish drivers can turn traffic lights green"

nmca — Mon, 28 Jul 2025 17:13:52 +0000

You can dream of better yet! If the spec was required to be open source for the government project, then you could have commercial choices and some less feature rich open source version.

New comment by nmca in "Tom Lehrer has died"

nmca — Sun, 27 Jul 2025 18:01:29 +0000

I feel the AI safety community has not made enough of Lehrer’s masterpiece on the topic:

https://youtu.be/frAEmhqdLFs?si=DYsY5Juco-kJ5eWD

New comment by nmca in "Quantitative AI progress needs accurate and transparent evaluation"

nmca — Fri, 25 Jul 2025 13:07:50 +0000

Indeed, it’s like saying a jet plane can fly!

New comment by nmca in "OpenAI claims gold-medal performance at IMO 2025"

nmca — Sat, 19 Jul 2025 21:06:09 +0000

It takes Tim Gowers more than hour and a half to go through q4! (Sure, he could go faster without video. But Tim Gowers! An hour and a half!!)

New comment by nmca in "OpenAI claims gold-medal performance at IMO 2025"

nmca — Sat, 19 Jul 2025 17:46:24 +0000

It’s interesting that this is a competition elite enough that several posters on a programming website don’t seem to understand what it is.

My very rough napkin math suggests that against the US reference class, imo gold is literally a one in a million talent (very roughly 20 people who make camp could get gold out of very roughly twenty million relevant high schoolers).

New comment by nmca in "OpenAI claims gold-medal performance at IMO 2025"

nmca — Sat, 19 Jul 2025 17:36:10 +0000

There wasn’t in the CoT for these problems.

New comment by nmca in "Emergent Misalignment: Narrow finetuning can produce broadly misaligned LLMs"

nmca — Mon, 14 Jul 2025 10:39:03 +0000

Great follow-up work from OpenAI on this:

https://openai.com/index/emergent-misalignment/

New comment by nmca in "Async Queue – One of my favorite programming interview questions"

nmca — Mon, 07 Jul 2025 01:40:54 +0000

Why is the method called sendOnce? It’s send with a capacity limiter / semaphore right, so what about it is Once?

New comment by nmca in "Typr – TUI typing test with a word selection algorithm inspired by keybr"

nmca — Fri, 27 Jun 2025 02:31:19 +0000

This is great! I have daydreamed about how to do a more complex algo: instead of character speed and frequency you could do bi-character speed/frequency/error rate and probably improve over keybr further.

New comment by nmca in "The Hollow Men of Hims"

nmca — Thu, 26 Jun 2025 18:54:54 +0000

“bad” seems fine?

New comment by nmca in "The Hollow Men of Hims"

nmca — Thu, 26 Jun 2025 04:27:13 +0000

This article is quite bad: haughty in tone and confused in content.

If you are interested in learning something about a key part subject matter (compounding loopholes and their impact on drug prices), this article is much better:

https://www.astralcodexten.com/p/the-compounding-loophole

New comment by nmca in "U.S. bombs Iranian nuclear sites"

nmca — Sun, 22 Jun 2025 05:46:47 +0000

How do you purport to know this?