Hacker News: m1el

New comment by m1el in "Voxtral Transcribe 2"

m1el — Wed, 04 Feb 2026 21:32:49 +0000

you should check out

https://github.com/pipecat-ai/nemotron-january-2026/

discovered through this twitter post:

https://x.com/kwindla/status/2008601717987045382

New comment by m1el in "Voxtral Transcribe 2"

m1el — Wed, 04 Feb 2026 18:29:05 +0000

I've been using nemotron ASR with my own ported inference, and happy about it:

https://huggingface.co/nvidia/nemotron-speech-streaming-en-0...

https://github.com/m1el/nemotron-asr.cpp https://huggingface.co/m1el/nemotron-speech-streaming-0.6B-g...

New comment by m1el in "Inverting the Xorshift128 random number generator"

m1el — Thu, 04 Sep 2025 20:58:22 +0000

Exactly. I've implemented a xorshift-based rng inverter previously, and here's the implementation for the algorithm in the article:

https://github.com/m1el/samaras/blob/master/src/xorshift128....

New comment by m1el in "Precision Clock Mk IV"

m1el — Sat, 31 May 2025 18:07:47 +0000

It's an artifact of the camera. The camera shutter is long enough that it averages the images over 33ms. At some point in the video you can see that a high speed camera can see the correct display.

New comment by m1el in "OpenAI asks White House for relief from state AI rules"

m1el — Thu, 13 Mar 2025 16:58:44 +0000

somehow, I suspect openai didn't "buy" all of the articles, books, websites they crawled and torrented.

New comment by m1el in "OpenAI asks White House for relief from state AI rules"

m1el — Thu, 13 Mar 2025 16:56:50 +0000

when it comes to real people, they get sued into oblivion for downloading copyrighted content, even for the purpose of learning. but when facebook & openai do it, at a much larger scale, suddenly the laws must be changed.

New comment by m1el in "QuickJS, the Next Generation: a mighty JavaScript engine"

m1el — Mon, 08 Jul 2024 07:00:35 +0000

From my observations: cold start, ease of patching. If you're running a lot of different JS code or restarting the code frequently, it's faster than node. Where it's useful: fuzzing. If you have a library/codebase you want to fuzz, you need to restart the code from a snapshot, and other engines seem to do it slower. It's also really easy to patch the code, because of the codebase size. If you need to trace/observe some behavior, just do it.

New comment by m1el in "Show HN: Execute JavaScript in a WebAssembly QuickJS sandbox"

m1el — Sun, 07 Jul 2024 19:41:44 +0000

Salesforce sandboxing is too easy to escape. Last time I needed to implement some feature for Salesforce, I've encountered 4 different escapes. It was also horrible dev experience.

New comment by m1el in "US Olympic and other teams will bring their own AC units to Paris"

m1el — Sun, 23 Jun 2024 01:50:33 +0000

It's not about being poor. First, the climate didn't require AC in most of the Europe, until ~10 years ago. You had a few hot days, and that's it. Second, thermal isolation in the US is extremely bad quality. I think people could cut their AC usage by half if they had proper thermal isolation in their houses. Third, northern Europe countries still don't have a climate to justify buying an AC.

New comment by m1el in "“Imprecise” language models are smaller, speedier, and nearly as accurate"

m1el — Fri, 31 May 2024 17:35:20 +0000

Artificial neural networks work the following way: you have a bunch of “neurons” which have inputs and an output. Neuron’s inputs have weights associated with them, the larger the weight, the more influence the input has on the neuron. These weights need to be represented in our computers somehow, usually people use IEEE754 floating point numbers. But these numbers take a lot of space (32 or 16 bits). So one approach people have invented is to use more compact representation of these weights (10, 8, down to 2 bits). This process is called quantisation. Having a smaller representation makes running the model faster because models are currently limited by memory bandwidth (how long it takes to read weights from memory), going from 32 bits to 2 bits potentially leads to 16x speed up. The surprising part is that the models still produce decent results, even when a lot of information from the weights was “thrown away”.

New comment by m1el in "Home Screen Advantage"

m1el — Mon, 26 Feb 2024 08:49:28 +0000

Not a browser, but a PWA. It's a web page, which you can "install" as an "app". Features like storage, background tasks and notifications are important for many applications, for example a messenger. These were available, and there is a market for those, but Apple has decided to kill that market.

New comment by m1el in "Testing how hard it is to cheat with ChatGPT in interviews"

m1el — Wed, 31 Jan 2024 23:59:18 +0000

I too select the lines that I read. However, I never select the entire page, unless I intend to copy it.

New comment by m1el in "Testing how hard it is to cheat with ChatGPT in interviews"

m1el — Wed, 31 Jan 2024 20:49:12 +0000

Oh, and to add an insult to the injury, I was using a collaborative editing tool. So I was able to see the person:

1) Select All (most likely followed by the copy) 2) Type the answer 3) Make an obvious mistake when they type else block, before the if

New comment by m1el in "Testing how hard it is to cheat with ChatGPT in interviews"

m1el — Wed, 31 Jan 2024 20:46:07 +0000

I've had a displeasure of interviewing someone who used ChatGPT in a live setting. It was pretty obvious: I ask a short question, and I say that I expect a short answer on which I will expand further. The interviewee sits there in awkward silence for a few seconds, and starts answering in a monotone voice, with sentence structure only seen on Wikipedia. This repeats for each consecutive question.

Of course this will change in the future, with more interactive models, but people who use ChatGPT on the interviews make a disservice to themselves and to the interviewer.

Maybe in the future everybody is going to use LLMs to externalize their thinking. But then why do I interview you? Why would I recommend you as a candidate for a position?

New comment by m1el in "Designing a SIMD Algorithm from Scratch"

m1el — Tue, 28 Nov 2023 11:45:06 +0000

The minimap contains a copy of the content, but with `transform: scale`. The rest is handling `window.onscroll` and mouse events on the overlay.

New comment by m1el in "Packing a string of digits into an integer quickly"

m1el — Sun, 09 Jul 2023 09:02:48 +0000

https://gist.github.com/m1el/e16fc153076d5764c95835de2936afe...

Here's my solution, which also uses pdep/pext

New comment by m1el in "Humans aren’t mentally ready for an AI-saturated ‘post-truth world’"

m1el — Wed, 21 Jun 2023 12:34:54 +0000

I am not scared for AI overflowing the news sites with bullshit. We already have a fire hydrant worth of bullshit content produced for consumption. Lies and fakes have coexisted with humans forever. People did rumours, then we had books, press, radio, television, and now the Internet. "But it's easier to produce lies/deepfakes today" -- true. However, the absolute cost of producing a lie per consumer already was negligible, and now it's even smaller. People will recalibrate their level of trust in technology and move on.

New comment by m1el in "Language models can explain neurons in language models"

m1el — Tue, 09 May 2023 18:26:51 +0000

You're correct to have a suspicion here. Hypothetically the explainer could omit a neuron or give a wrong explanation for the role of a neuron. Imagine you're trying to understand a neural network, and you spend enormous amount of time generating hypotheses and validating them. Well the explainer might give you 90% correct hypotheses, it means you have 10 times less work to produce hypotheses. So if you have a solid way of testing an explanation, even if the explainer is evil, it's still useful.

New comment by m1el in "Testing ability of GPT4 and open-source LLMs to detect C++ bugs"

m1el — Sun, 16 Apr 2023 06:03:26 +0000

I don't understand how the author can write "No, they cannot.", while the page says "GPT 4: 0 false alarms in 15 good examples. Detects 13 of 13 bugs."

Edit: the post says open-source models.

New comment by m1el in "Web LLM – WebGPU Powered Inference of Large Language Models"

m1el — Sat, 15 Apr 2023 20:21:42 +0000

If you're doing inference on neural networks, each weight has to be read at least once per token. This means you're going to read at least the size of the entire model, per token, at least once during inference. If your model is 60GB, and you're reading it from the hard drive, then your bare minimum time of inference per token will be limited by your hard drive read throughput. Macbooks have ~4GB/s sequential read speed. Which means your inference time per token will be strictly more than 15 seconds. If your model is in RAM, then (according to Apple's advertising) your memory speed is 400GB/s, which is 100x your hard drive speed, and just the memory throughput will not be as much of a bottleneck here.