Hacker News: wills_forward

New comment by wills_forward in "Mistral OCR 3"

wills_forward — Sat, 20 Dec 2025 00:00:55 +0000

Why not use both? I just built a pipeline for document data extraction that uses PaddleOCR, then Gemini 3 to check + fix errors. It gets close to 99.9% on extraction from financial statements finally on par with humans.

New comment by wills_forward in "Qwen3-Omni: Native Omni AI model for text, image and video"

wills_forward — Mon, 22 Sep 2025 19:32:08 +0000

https://x.com/whowillrickwill/status/1920723985311903767

New comment by wills_forward in "A computer upgrade shut down BART"

wills_forward — Fri, 05 Sep 2025 16:04:42 +0000

The cheap easy take: it's tragically ironic that the software running the infrastructure in Silicon Valley is such a problem

New comment by wills_forward in "Lossless LLM compression for efficient GPU inference via dynamic-length float"

wills_forward — Fri, 25 Apr 2025 19:01:34 +0000

So this could universally decrease the memory requirements by un-quantitized LLMs by 30%? Seems big if true.

New comment by wills_forward in "Emergence of a second law of thermodynamics in isolated quantum systems"

wills_forward — Sun, 02 Feb 2025 23:11:42 +0000

This paper is basically statistical mechanics with a quantum veneer. Two major issues:

1. Scale: They're simulating just 13 qubits with QuTiP and making grand claims about quantum thermodynamics. The computational complexity they're glossing over here is astronomical. Anyone who's actually worked with quantum systems knows you can't just handwave away the scaling problems.

2. Measurement Problem: Their whole argument about instantaneous vs time-averaged measurements is just repackaging the quantum measurement problem without actually solving anything. They're doing the same philosophical shell game that every "breakthrough" quantum paper does by moving around where they put the observer and pretending they've discovered something profound.

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

wills_forward — Sun, 15 Dec 2024 00:51:13 +0000

It was funny to hear the same guy warning LMMs were getting too powerful now talking about the limits of available original training data.

Transformer by Hand in Excel

wills_forward — Sat, 23 Nov 2024 14:58:53 +0000

Article URL: https://twitter.com/proftomyeh/status/1859282491955130452

Comments URL: https://news.ycombinator.com/item?id=42221241

Points: 2

# Comments: 0

Tool to transcode videos in nothing but a browser

wills_forward — Thu, 03 Oct 2024 02:53:35 +0000

Article URL: https://rgbcu.be/yacine

Comments URL: https://news.ycombinator.com/item?id=41726906

Points: 2

# Comments: 0

Satellite internet has quickly changed life in Alaska

wills_forward — Sun, 18 Aug 2024 22:31:26 +0000

Article URL: https://www.adn.com/alaska-news/2024/08/18/satellite-internet-has-disrupted-the-market-in-alaska-and-transformed-everyday-life-for-many/

Comments URL: https://news.ycombinator.com/item?id=41286078

Points: 5

# Comments: 2

Three week satellite time-lapse video of earth

wills_forward — Mon, 05 Aug 2024 22:24:18 +0000

Article URL: https://www.youtube.com/watch?v=7nQvMa6QfMM

Comments URL: https://news.ycombinator.com/item?id=41166174

Points: 3

# Comments: 0

New comment by wills_forward in "Remembering Bell Labs as legendary idea factory prepares to leave N.J. home"

wills_forward — Sun, 21 Jan 2024 18:57:03 +0000

Citadel or Jump?

Helluva pivot

wills_forward — Sat, 13 Jan 2024 03:47:46 +0000

Article URL: https://twitter.com/whowillrickwill/status/1746015172928454725

Comments URL: https://news.ycombinator.com/item?id=38977177

Points: 3

# Comments: 0

New comment by wills_forward in "Show HN: GPT-V and OCR for Screen Control"

wills_forward — Wed, 13 Dec 2023 14:03:43 +0000

I really like the elegant simplicity of tagging the screen elements like that and not obfuscating it away.

Nice work too!

New comment by wills_forward in "Show HN: Beeper Mini – iMessage client for Android"

wills_forward — Wed, 06 Dec 2023 00:03:56 +0000

Is this part of the reason Apple decided to support RCS? They knew the iMessage system would get opened up eventually anyway...

New comment by wills_forward in "LLM Visualization"

wills_forward — Tue, 05 Dec 2023 03:21:49 +0000

Aw thanks for such encouragement all

New comment by wills_forward in "LLM Visualization"

wills_forward — Sun, 03 Dec 2023 18:22:52 +0000

My jaw drop to see algorhythmic complexity laid out so clearly in a 3d space like that. I wish I was smart enough to know if it's accurate or not.

New comment by wills_forward in "Fast Llama 2 on CPUs with Sparse Fine-Tuning and DeepSparse"

wills_forward — Thu, 23 Nov 2023 15:46:00 +0000

This seems like a big step forward in terms of running specific use-case trained inference locally, right? At least given current hardware generally deployed in business.

New comment by wills_forward in "Theory of interstellar trade (1978) [pdf]"

wills_forward — Wed, 16 Aug 2023 17:55:47 +0000

Great podcast

New comment by wills_forward in "Retentive Network: A Successor to Transformer for Large Language Models"

wills_forward — Sun, 23 Jul 2023 04:05:41 +0000

MIT and Microsoft Researchers Introduce "RetNet" - An 8X Faster Transformer Alternative for AI

New comment by wills_forward in "PredictIt is shutting down after CFTC withdraws no-action letter"

wills_forward — Fri, 05 Aug 2022 04:17:34 +0000

This is such a bummer, but not shocked they're out-of-bounds given the bullets in the letter and how big they've become: https://www.cftc.gov/csl/22-08/download