Hacker News: woodson

New comment by woodson in "Apple Silicon and Virtual Machines: Beating the 2 VM Limit (2023)"

woodson — Sat, 11 Apr 2026 22:55:33 +0000

Well, but their customers are those that buy Apple hardware.

New comment by woodson in "Hybrid Attention"

woodson — Tue, 07 Apr 2026 14:51:13 +0000

Look into RWKV.

New comment by woodson in "Claude Code Unpacked : A visual guide"

woodson — Wed, 01 Apr 2026 15:29:31 +0000

It’s a distinction that IMHO likely doesn’t make much difference, at least for the mostly automated/non-interactive coding agent use case. What matters more is how well the post-training on synthetic harness traces works.

New comment by woodson in "An unsolicited guide to being a researcher [pdf]"

woodson — Mon, 23 Mar 2026 15:45:26 +0000

There are still opportunities, but they aren’t paid nearly as well as less researchy positions in industry. US post-doc salaries at state universities aren’t that high.

New comment by woodson in "Show HN: I built a sub-500ms latency voice agent from scratch"

woodson — Mon, 02 Mar 2026 23:17:27 +0000

You mean Moshi (https://github.com/kyutai-labs/moshi)? Since Personaplex is just a finetuned Moshi model.

New comment by woodson in "Euro firms must ditch Uncle Sam's clouds and go EU-native"

woodson — Sat, 31 Jan 2026 19:34:58 +0000

…who lives in Oregon, in the US.

New comment by woodson in "Qwen3-TTS family is now open sourced: Voice design, clone, and generation"

woodson — Thu, 22 Jan 2026 22:25:06 +0000

This is about speech to text, not speech recognition.

New comment by woodson in "EU–INC – A new pan-European legal entity"

woodson — Wed, 21 Jan 2026 17:07:26 +0000

You still need a tax accountant in France to register the FTE and file paperwork with the tax office and social insurance.

New comment by woodson in "EU–INC – A new pan-European legal entity"

woodson — Wed, 21 Jan 2026 16:27:05 +0000

The bank issues/refusals may have something to do with FATCA. If you have anything to do with the US in terms of taxes, many EU banks don’t want you as their customer. If it’s a subsidiary of a foreign company, then a lot of paperwork is required to prove that the foreign owners actually exist.

New comment by woodson in "Show HN: Yolobox – Run AI coding agents with full sudo without nuking home dir"

woodson — Mon, 12 Jan 2026 20:13:12 +0000

This is basically a devcontainer, right?

New comment by woodson in "Sopro TTS: A 169M model with zero-shot voice cloning that runs on the CPU"

woodson — Fri, 09 Jan 2026 00:08:43 +0000

Jeez, OP asked what it means in this context (zero-shot voice cloning), where you quoted a generic definition copied from Wikipedia. I defined it concretely for this context. Don't take it as a slight, there is no need to get all argumentative.

New comment by woodson in "Sopro TTS: A 169M model with zero-shot voice cloning that runs on the CPU"

woodson — Thu, 08 Jan 2026 23:36:31 +0000

From your definition:

> a learner observes samples from classes which were not observed during training, and needs to predict the class that they belong to.

That's not what happens in zero-shot voice cloning, which is why I dismissed your definition copied from Wikipedia.

New comment by woodson in "Sopro TTS: A 169M model with zero-shot voice cloning that runs on the CPU"

woodson — Thu, 08 Jan 2026 22:39:37 +0000

I don't disagree, but that's what people started calling it. Zero-shot doesn't make sense anyway, as how would the model know what voice it should sound like (unless it's a celebrity voice or similar included in the training data where it's enough to specify a name).

New comment by woodson in "Sopro TTS: A 169M model with zero-shot voice cloning that runs on the CPU"

woodson — Thu, 08 Jan 2026 22:36:51 +0000

Does the 169M include the ~90M params for the Mimi codec? Interesting approach using FiLM for speaker conditioning.

New comment by woodson in "Sopro TTS: A 169M model with zero-shot voice cloning that runs on the CPU"

woodson — Thu, 08 Jan 2026 22:19:26 +0000

This generic answer from Wikipedia is not very helpful in this context. Zero-shot voice cloning in TTS usually means that data of the target speaker you want the generated speech to sound like does not need to be included in the training data used to train the TTS models. In other words, you can provide an audio sample of the target speaker together with the text to be spoken to generate the audio that sounds like it was spoken by that speaker.

New comment by woodson in "Google Opal"

woodson — Wed, 31 Dec 2025 07:25:43 +0000

The local Gemma models are pretty good for tasks involving multilingual inputs (translation, summarization, etc.). They have their niche.

New comment by woodson in "Texas is suing all of the big TV makers for spying on what you watch"

woodson — Thu, 18 Dec 2025 23:13:57 +0000

There are perceptual hashing algorithms for images/video/audio (dsp and ML based) that could work for that.

New comment by woodson in "Nvidia Nemotron 3 Family of Models"

woodson — Wed, 17 Dec 2025 00:46:26 +0000

Not OP, but perhaps they mean not putting too much faith in common benchmarks (thanks to benchmaxxing).

New comment by woodson in "Generative Optogenetics"

woodson — Mon, 15 Dec 2025 05:05:00 +0000

> Sometimes DARPA is real slow on identifying good ideas. (more than 8 years later today...) better late than never.

Perhaps they noticed a development (breakthrough?) in an adjacent field that makes this now more interesting and/or practical/feasible.

New comment by woodson in "Open (Apache 2.0) TTS model for streaming conversational audio in realtime"

woodson — Fri, 28 Nov 2025 12:53:12 +0000

Looks very similar to Kyutai’s models, given that it uses the same neural audio codec (Mimi) and Depformer module etc.