Hacker News: axiom92

New comment by axiom92 in "Program-of-Thought Prompting Outperforms Chain-of-Thought by 15% (2022)"

axiom92 — Sun, 30 Nov 2025 22:01:04 +0000

https://www.reddit.com/r/ChatGPT/comments/14sqcg8/anyone_els...

New comment by axiom92 in "Program-of-Thought Prompting Outperforms Chain-of-Thought by 15% (2022)"

axiom92 — Sun, 30 Nov 2025 22:00:38 +0000

This was integrated in gpt4 2 years ago:

https://www.reddit.com/r/ChatGPT/comments/14sqcg8/anyone_els...

New comment by axiom92 in "Program-of-Thought Prompting Outperforms Chain-of-Thought by 15% (2022)"

axiom92 — Sun, 30 Nov 2025 20:25:49 +0000

And even before this work, there was "PAL: Program-aided Language Models" (https://arxiv.org/abs/2211.10435, https://reasonwithpal.com/).

Afaik PaLM (Google's OG big models) tried this trick, but it didn't work for them. I think it's because PaL used descriptive inline comments + meaningful variable names. Compare the following:

```python

# calculate the remaining apples

apples_left = apples_bought - apples_eaten

```

vs.

```python

x = y - z

```

We have ablations in https://arxiv.org/abs/2211.10435 showing that both are indeed useful (see "Crafting prompts for PAL").

New comment by axiom92 in "ChatGPT Atlas"

axiom92 — Tue, 21 Oct 2025 23:55:07 +0000

You can do this at grok.com.

There is a "start thread" option below every conversation. You can also read the responses aloud (helpful if you want to do something async).

New comment by axiom92 in "BERT is just a single text diffusion step"

axiom92 — Mon, 20 Oct 2025 18:29:40 +0000

Yeah, that's the first formal reference I remember as well (although, BERT is probably the first thing NLP folks will think of after reading about diffusion).

I collected a few other text-diffusion early references here about 3 years ago: https://github.com/madaan/minimal-text-diffusion?tab=readme-....

New comment by axiom92 in "Adaptive LLM routing under budget constraints"

axiom92 — Mon, 01 Sep 2025 21:51:28 +0000

From last neurips https://automix-llm.github.io/automix/

New comment by axiom92 in "Grok 4 Launch [video]"

axiom92 — Tue, 15 Jul 2025 08:53:14 +0000

The demo was done live (as was everything else).

New comment by axiom92 in "The Dark Forest hypothesis is absurd"

axiom92 — Tue, 23 Jan 2024 20:55:38 +0000

Looks pretty cool https://www.youtube.com/watch?v=mogSbMD6EcY

Although, it seems it's only going to cover the first book (which makes sense, given how difficult the other two would be to film). The real magic for me was in book 3. It was inspiring to see someone think so far out, so boldly.

New comment by axiom92 in "Turing Complete Transformers: Two Transformers Are More Powerful Than One"

axiom92 — Tue, 09 Jan 2024 05:13:05 +0000

Welcome to one of the most hated parts of the academia.

New comment by axiom92 in "Sam Altman: if I start going off, board should go after me for my shares"

axiom92 — Sat, 18 Nov 2023 05:44:48 +0000

The joke is that he doesn't own any OpenAI shares.

[1] https://www.cnbc.com/2023/03/24/openai-ceo-sam-altman-didnt-...

New comment by axiom92 in "Fuyu-8B: A multimodal architecture for AI agents"

axiom92 — Wed, 18 Oct 2023 21:31:20 +0000

Right, but no separate image encoder + half the size could be very helpful for many applications.

New comment by axiom92 in "Against LLM Maximalism"

axiom92 — Wed, 13 Sep 2023 15:29:02 +0000

> tasks that need deterministic outputs and the thing you need to create is already known statically

Wow, interesting. Do you have any example for this?

I've realized that LLMs are fairly good at string processing tasks that a really complex regex might also do, so I can see the point in those.

New comment by axiom92 in "How Is LLaMa.cpp Possible?"

axiom92 — Wed, 16 Aug 2023 00:47:11 +0000

> And basically all servers will have 8xA100

for those wondering: no this is not the norm. My lab at CMU doesn't own any A100s (we have A6000s).

New comment by axiom92 in "Never waste a midlife crisis"

axiom92 — Tue, 18 Jul 2023 01:12:25 +0000

> The options seemed to be: If I went for it, I’d be penniless, and if I didn’t go for it, I’d be bitter. I’d be bitter going forward. Penniless certainly beats bitter. So I made the decision.

Kind of like industry -> PhD decision.

New comment by axiom92 in "AI agents that “self-reflect” perform better in changing environments"

axiom92 — Fri, 07 Jul 2023 06:07:37 +0000

Some of our recent/relevant work: https://selfrefine.info/

New comment by axiom92 in "Alien Supercivilizations Absent from 100k Nearby Galaxies (2015)"

axiom92 — Sun, 30 Apr 2023 19:05:00 +0000

https://en.wikipedia.org/wiki/Dark_forest_hypothesis

Also a major theme of https://www.amazon.com/Dark-Forest-Remembrance-Earths-Past/d...

New comment by axiom92 in "Large Language Models Are Human-Level Prompt Engineers"

axiom92 — Sun, 09 Apr 2023 23:20:53 +0000

For those curious about self-refining systems: https://selfrefine.info/ (our recent work).

Request a ride by dialing 1-833-USE-Uber

axiom92 — Wed, 22 Mar 2023 23:47:29 +0000

Article URL: https://www.uber.com/blog/california/1-833-use-uber/

Comments URL: https://news.ycombinator.com/item?id=35268667

Points: 3

# Comments: 3

New comment by axiom92 in "OpenAI to discontinue support for the Codex API"

axiom92 — Tue, 21 Mar 2023 14:40:44 +0000

cushman and code-davinci are similar for sure (same architecture). Perhaps that's what they meant.

New comment by axiom92 in "OpenAI to discontinue support for the Codex API"

axiom92 — Tue, 21 Mar 2023 04:27:46 +0000

Codex (code-davinci-002) is free (limited beta).