Hacker News: chompychop

New comment by chompychop in "Show HN: Shoehorn – Quantize any model down to run on your machine"

chompychop — Fri, 21 Aug 2026 20:04:58 +0000

Well, then don't get all snarky and dismissive of things you might not be knowledgeable about ("you" here referring to OP).

New comment by chompychop in "Show HN: Shoehorn – Quantize any model down to run on your machine"

chompychop — Fri, 21 Aug 2026 19:28:15 +0000

I gotta laugh at your thought process: knowing Fable 5 is a large frontier model, you're telling me that the first thing that came to your mind on seeing that model name is that it's a quantized version of Fable? As opposed to a distillation/fine-tuning on Fable responses?

New comment by chompychop in "Phosh 0.56.0"

chompychop — Sun, 05 Jul 2026 15:42:51 +0000

With tap-to-pay, you can store multiple cards in your digital wallet, and you don't have to remember any of their PINs. You can use your fingerprint to sign transactions. I believe this makes it faster.

New comment by chompychop in "Jamesob's guide to running SOTA LLMs locally"

chompychop — Fri, 03 Jul 2026 17:29:48 +0000

Is Whisper still considered SOTA for STT? Since it came out years ago, I'd have assumed there are better models by now.

New comment by chompychop in "ETH-Zurich: Digital Design and Computer Architecture; 227-0003-10L, Spring, 2025"

chompychop — Sun, 30 Nov 2025 19:45:41 +0000

Does anyone know how this course compares to the NAND2Tetris course?

New comment by chompychop in "Mixpanel Security Breach"

chompychop — Thu, 27 Nov 2025 08:52:24 +0000

Sam Altman.

New comment by chompychop in "Await Is Not a Context Switch: Understanding Python's Coroutines vs. Tasks"

chompychop — Wed, 26 Nov 2025 12:31:20 +0000

Asynchronous and parallel programming are concepts I've never really learned, and admittedly scared to use, because I never really understand what the code is doing or even the right way of writing such code. Does anyone have any good resources that helped you learn this and have a good mental model of the concepts involved?

New comment by chompychop in "Meta Segment Anything Model 3"

chompychop — Tue, 25 Nov 2025 13:51:18 +0000

Thanks! TIL there's a class of segmentation models with the YOLO naming scheme.

New comment by chompychop in "Meta Segment Anything Model 3"

chompychop — Tue, 25 Nov 2025 13:01:43 +0000

YOLO is not a segmentation model.

New comment by chompychop in "Cormac McCarthy's personal library"

chompychop — Thu, 02 Oct 2025 16:53:18 +0000

Where can I read more about your research? Knowledge graphs interest me.

New comment by chompychop in "FFmpeg merges WebRTC support"

chompychop — Thu, 05 Jun 2025 09:11:31 +0000

I have a beginner question - Can WebRTC be used as an alternative to sending base64-encoded images to a backend server for image processing? Is this approach recommended?

New comment by chompychop in "Ask HN: What projects do you donate to?"

chompychop — Tue, 27 May 2025 10:12:12 +0000

But modern browsers do have this feature, don't they? I can do this on Chrome and Firefox, at least.

New comment by chompychop in "Harnessing the Universal Geometry of Embeddings"

chompychop — Wed, 21 May 2025 21:12:26 +0000

Don't you mean "John" instead of "Jack"? :)

New comment by chompychop in "Show HN: Factorio Learning Environment – Agents Build Factories"

chompychop — Tue, 11 Mar 2025 17:18:03 +0000

It's edging pretty close to this - multimodal LLMs can already beat early bosses in soulslike games: https://arxiv.org/html/2409.12889v2

New comment by chompychop in "Show HN: Open-Source Colab Notebooks to Implement Advanced RAG Techniques"

chompychop — Wed, 04 Dec 2024 11:41:20 +0000

Does it still have the "abstraction hell" issue when trying to work with it for custom, non out-of-the-box use cases?

New comment by chompychop in "Show HN: Open-Source Colab Notebooks to Implement Advanced RAG Techniques"

chompychop — Wed, 04 Dec 2024 06:48:10 +0000

Huh? All of their notebooks use LangChain.

New comment by chompychop in "AlphaProof's Greatest Hits"

chompychop — Mon, 18 Nov 2024 09:31:08 +0000

Is it currently possible to reliably limit the cut-off knowledge of an LLM (either during training or inference)? An interesting experiment would be to feed an LLM mathematical knowledge only up to the year of proving a theorem, and then see if it can actually come up with the novel techniques used in the proof. For example, having only access to papers prior to 1993, can an LLM come up with Wiles' proof of FLT?

New comment by chompychop in "LoRA vs. Full Fine-Tuning: An Illusion of Equivalence"

chompychop — Fri, 08 Nov 2024 22:13:53 +0000

I don't see how the authorship by Christopher Manning shifts favour towards the other paper; this paper has Antonio Torralba as a co-author, who's also one of the big shots in AI.

New comment by chompychop in "Math is still catching up to the genius of Ramanujan"

chompychop — Tue, 22 Oct 2024 07:24:56 +0000

Manjul Bhargava is not Indian.

New comment by chompychop in "Show HN: Fast and Exact Algorithm for Image Merging"

chompychop — Thu, 26 Sep 2024 12:15:18 +0000

I believe the term in research literature for this is "Dynamic Video Synopsis". Check this CVPR 2006 paper for instance: https://www.cs.huji.ac.il/~peleg/papers/cvpr06-synopsis.pdf