Hacker News: error9348

New comment by error9348 in "Introducing command And commandfor In HTML"

error9348 — Fri, 07 Mar 2025 18:02:37 +0000

Would make interaction easier for ai agents on the web.

New comment by error9348 in "Ilya Sutskever NeurIPS talk [video]"

error9348 — Sat, 14 Dec 2024 21:09:16 +0000

Wow, thanks for the correction. Didn't know this existed - to be fair, I only found a preview last year I tried, and paid up.

New comment by error9348 in "Ilya Sutskever NeurIPS talk [video]"

error9348 — Sat, 14 Dec 2024 20:58:18 +0000

It would be great if all NeurIPS talks were accessible for free like this one. I understand they generate some revenue from online ticket sales, but it would be a great resource. Maybe some big org could sponsor it.

New comment by error9348 in "OpenAI and Other AI Companies Pay $2 an Hour for Kenyan Jobs"

error9348 — Tue, 26 Nov 2024 02:15:56 +0000

> Documents reviewed by 60 Minutes show OpenAI agreed to pay SAMA $12.50 an hour per worker, much more than the $2 the workers actually got, though SAMA says what it paid is a fair wage for the region.

Let me guess. SAMA had the classic sku/$/hr saas pricing.

New comment by error9348 in "Trump wins presidency for second time"

error9348 — Wed, 06 Nov 2024 17:55:00 +0000

Post hoc ergo propter hoc

New comment by error9348 in "A 9.9 CVE has been announced for Linux"

error9348 — Thu, 26 Sep 2024 14:16:51 +0000

Original source seems to be https://archive.is/wwoQZ

New comment by error9348 in "OpenAI threatens to revoke o1 access for asking it about its chain of thought"

error9348 — Fri, 13 Sep 2024 23:20:51 +0000

will these (or have these) notifications been paused while the decision is being reconsidered?

New comment by error9348 in "CSS Can Get You in Jail – Browser renderers, now deemed criminals"

error9348 — Wed, 03 Jul 2024 06:55:54 +0000

> by using CSS for displaying the numbers, you put the validity of the documents in the hands of the rendering engine.

Any automation tool could result in a regression which may generate invalid numbering. At any rate, I doubt a court would ignore mens rea.

New comment by error9348 in "CoreNet: A library for training deep neural networks"

error9348 — Wed, 24 Apr 2024 04:26:57 +0000

The interface looks very Apple as well. Looks like you create a config file, and you already have a model in mind with the hyperparameters and it provides a simple interface. How useful is this to researchers trying to hack the model architecture?

One example: https://github.com/apple/corenet/tree/main/projects/clip#tra...

New comment by error9348 in "Penzai: JAX research toolkit for building, editing, and visualizing neural nets"

error9348 — Sun, 21 Apr 2024 18:05:02 +0000

Jax trends on papers with code:

https://paperswithcode.com/trends

New comment by error9348 in "OpenAI's comment to the NTIA on open model weights"

error9348 — Tue, 02 Apr 2024 01:16:33 +0000

Q3-7 & Q3-5d get to the workability. I don't think OpenAI responds to that part of the RFC. Meta's comment on that issue seems to be fairly clear, they oppose the proposed rules on KYC for IaaS and are "not aware of technical capabilities that could not be overcome by determined, well-resourced, and capable actors".

https://www.ntia.gov/sites/default/files/publications/open_m...

https://about.fb.com/wp-content/uploads/2024/03/NTIA-RFC-Met...

New comment by error9348 in "MIT Unveils Gen AI Tool That Generates High Res Images 30 Times Faster"

error9348 — Wed, 27 Mar 2024 03:52:38 +0000

Could be. Could also be that LLM is to OpenAI what information retrieval is to Google. A lot is publicly known in the information retrieval space, but google still dominates.

New comment by error9348 in "OpenAI scrapped a promise to disclose key documents to the public"

error9348 — Wed, 24 Jan 2024 20:46:24 +0000

Looks like they draw a line at generative AI. CLIP / Whisper / Gym are open; Jukebox / GPT / DallE are not.

Masked Audio Generation Using a Single Non-Autoregressive Transformer

error9348 — Mon, 15 Jan 2024 23:47:39 +0000

Article URL: https://pages.cs.huji.ac.il/adiyoss-lab/MAGNeT/

Comments URL: https://news.ycombinator.com/item?id=39007629

Points: 1

# Comments: 0

New comment by error9348 in "How OpenAI is approaching 2024 worldwide elections"

error9348 — Mon, 15 Jan 2024 22:28:59 +0000

The question the blog asks on election procedure is easy to get right, why does ChatGPT refuse to answer? I'd bet the answer it would otherwise produce is reasonable. Perplexity gives a fine answer. The collateral damage from fixing hallucinations seems to hurt accessibility.

New comment by error9348 in "Show HN: #!/usr/bin/env docker run"

error9348 — Sun, 14 Jan 2024 07:39:10 +0000

Should be fine, you can even compile and run a C file using a shebang

New comment by error9348 in "US regulator considers stripping Boeing's right to self-inspect planes"

error9348 — Fri, 12 Jan 2024 21:25:28 +0000

There are all kinds of externalities, including safety, which are ignored. US tax policy which allows write offs only for large vehicles which are more likely to cause deaths. Tire emissions. Congestion pricing -- common in Europe -- needs years or reviews and lawsuits. NHTSA doesn't require bicycle test devices.

New comment by error9348 in "CLI user experience case study"

error9348 — Fri, 12 Jan 2024 15:31:23 +0000

Terseness is the point. Man pages can be discoverable too if your shell supports autocomplete

New comment by error9348 in "U.S. moves closer to filing antitrust case against Apple"

error9348 — Fri, 05 Jan 2024 23:23:42 +0000

Thanks. That explains why I never notice the difference.

New comment by error9348 in "U.S. moves closer to filing antitrust case against Apple"

error9348 — Fri, 05 Jan 2024 21:44:52 +0000

This may be ignorant, but why does green/blue bubble matter for US folks? Almost all cell providers have free SMS/MMS. Would people even notice it if the color was always blue?