Hacker News: larme

New comment by larme in "Tiny $70 Xteink X3 e-reader"

larme — Mon, 07 Sep 2026 17:45:33 +0000

What I linked is the official repo. My customization is private because I replace the original fonts with a commercial font (which I own a license) so I feel it's better not to make the repo public.

New comment by larme in "Tiny $70 Xteink X3 e-reader"

larme — Mon, 07 Sep 2026 17:39:44 +0000

Also I think there's a youtuber who shoot some videos about using microslate.

New comment by larme in "Tiny $70 Xteink X3 e-reader"

larme — Mon, 07 Sep 2026 17:39:08 +0000

I list some cons in a sibling comment. You may want to check if you plan to buy one

New comment by larme in "Tiny $70 Xteink X3 e-reader"

larme — Mon, 07 Sep 2026 17:37:25 +0000

I don't have video on my hand but a photo is here: https://files.catbox.moe/z8suof.jpg

There are some cons:

- some users report text gets blurred out if used under direct sunlight for a long time. I don't use it outside so I'm not affected

- sometimes the BLE connection is sketchy. I'm not sure if it's hardware or software problem

- the editor features are very barebone so you may want to customize your own version

- beware there are unflashable edition of this device

New comment by larme in "Tiny $70 Xteink X3 e-reader"

larme — Mon, 07 Sep 2026 16:08:21 +0000

I have an X4 and use it as a writer deck with microslate firmware[0]

I heavily customized the source code with llm so I can use emacs style keybindings and search.

[0]: https://github.com/Josh-writes/microslate-firmware

New comment by larme in "A walkable ASCII cyberpunk city in one HTML file [video]"

larme — Tue, 01 Sep 2026 07:06:07 +0000

because modern browsers are monsters that a lot of us want to avoid

New comment by larme in "How to Train a Gen AI Kick Drum Model on Your Old Linux Desktop with 6GB VRAM"

larme — Thu, 16 Jul 2026 19:06:40 +0000

For me creating the exact sound is not very interesting from sound designing perspective. You can always sample the real instrument.

Like physical modeling synthesis, the interesting part is to compress the sound to some parameters that you can tweak and generate new sounds

Another approach is VAE, which also you give your some latent embedding, you can tweak the embedding to generate new sound. However the meaning of this embedding is not explicit.

New comment by larme in "How to Train a Gen AI Kick Drum Model on Your Old Linux Desktop with 6GB VRAM"

larme — Thu, 16 Jul 2026 17:40:41 +0000

People who are interested in this application should check synplant[0]. It has a ML technology called "Genopatch" which gives you 2 functionality:

1. you can try to describe a sound with some tags and it will try to generate a sound to capture the feeling of these tags

2.you can feed it with a sound sample and it will try to re-synthesize the sound with its synth engine. Though the end result will usually be just a "re-imagined" version of your input sample.

My guess is the underlying model is not a "deep" model. The main benefit is that the end result is not a wave file, but a list of generated parameters that can be synthesized by the synthplant engine. And now it comes the interesting part: you can tweak these parameters to finetune the generated sound. These parameters have actual meanings (FM ratio, reverb etc.)

[0]: https://soniccharge.com/synplant

New comment by larme in "Old and new apps, via modern coding agents"

larme — Sun, 12 Jul 2026 13:18:16 +0000

Before LLM there has already been Fields medalist[0] who creates professional software[1].

[0]: https://en.wikipedia.org/wiki/Martin_Hairer

[1]: https://www.hairersoft.com/

New comment by larme in "Interview with Mitchell Hashimoto about Ghostty and Zig"

larme — Fri, 10 Jul 2026 14:27:46 +0000

This description sounds like a cult.

New comment by larme in "How to convert between wealth and income tax"

larme — Fri, 22 May 2026 17:47:15 +0000

> It's a big democracy red flag when a majority wants to take a lot from a tiny minority

It’s not “taking”. The rich give out some money so the society has a higher probability to stay peaceful. or a violent revolution may happen.

This is really a win win situation

New comment by larme in "Show HN: KVBoost – chunk-level KV cache reuse for HuggingFace, 5–48x faster TTFT"

larme — Fri, 22 May 2026 07:56:54 +0000

Go is not high performance enough. Like what others said, you implement the high performance part in C++ and use python to glue them.

New comment by larme in "SBCL: the ultimate assembly code breadboard (2014)"

larme — Wed, 20 May 2026 18:59:17 +0000

Try pair reading it with a good thinking LLM like GPT 5.5 or Claude Opus. I found it help me a lot.

I have started learning SBCL internal from the beginning of this year with the help of GPT, and I really want to contribute to SBCL compiler someday in future.

New comment by larme in "AI is too expensive"

larme — Tue, 19 May 2026 16:33:51 +0000

Some of these models are open weight. You can try hosting them and do the price calculation yourself.

They also publish papers talking about how to save kv cache and computation powers. Because currently they don't have the most powerful nvidia cards, training and inference efficiency is very import for them.

New comment by larme in "Teenage Engineering's PO-32 acoustic modem and synth implementation"

larme — Tue, 31 Mar 2026 19:31:44 +0000

That's amazing. I wonder how accurate is the synth engine?

New comment by larme in "Jensen: 'We've done our country a great disservice' by offshoring"

larme — Mon, 05 Jan 2026 14:47:51 +0000

It’s true that American companies have benefited from sending jobs overseas. Instead of trying to stop them, perhaps we could explore ways to share this money more fairly across the country.

New comment by larme in "FPGA Based IBM-PC-XT"

larme — Mon, 17 Nov 2025 14:35:13 +0000

Have you tried pocket 386? https://arstechnica.com/gadgets/2024/08/a-few-weeks-with-the...

New comment by larme in "Apple Silicon GPU Support in Mojo"

larme — Mon, 22 Sep 2025 16:43:20 +0000

Thanks for the answers!

New comment by larme in "Apple Silicon GPU Support in Mojo"

larme — Mon, 22 Sep 2025 04:07:04 +0000

I'm very interesting in this field (realtime audio + GPU programming). How do you deal with the latency? Do you send or multiple single vectors/buffers to GPU?

Also I think because samples in one channel need to be processed sequentially, does that mean mono audio processing won't benefit a lot from GPU programming. Or maybe you are dealing with spectral signal processing?

New comment by larme in "LLM Inference Handbook"

larme — Fri, 11 Jul 2025 16:13:13 +0000

Wow that's really thorough