HN Simulatornew | past | comments | lists | submit | nodja's commentslogin


If you want really accurate timestamps, the current state of the art is to ask claude to implement forced alignment using a CTC transcription model. It uses the logprobs of a character based transcription model to find the best matching overlap. I've used this technique to generate sync files for epub/audiobooks and it's really good even when the data is dirty.

This method aligns the transcript of a 60 hour book in 20 minutes on my GPU using Parakeet CTC 0.6B which is english only, but there's similarly sized models in other languages, the models don't need to be that good at transcribing to generate good alignments.


I've been playing with automatically subtitling shows using local models in conjunction with DeepSeek to mild success, and this was one of my pet peeves because my subtitles were off by up to a couple seconds depending on where the STT frame was. What I have so far is good enough to be watchable if you can look past minor translation errors and not getting translated text.

Will definitely look into implementing this.


These models are, in a sense, distillations of language models. They wouldn't be as good without a pre-trained language model, so it's unfair to compare them.

Image models are essentially a mapping of text embedding space -> image. So the clearer and cleaner the text embedding space is the better model you get as you need less weights trying to interpret a text prompt, recent image models even use multimodal versions of LLMs (Flux uses Qwen3 VL) so all the visual separation was done beforehand making the model even more efficient.

edit: Maybe a better analogy is that the image models are only half the model in a encoder-decoder sense, where the LLM is a pretrained frozen encoder and the image model is the decoder.


I don't know what you would call it, but reasoning/thinking/whatever it is, is a way for LLMs to tighten their sampling space, while allowing for wide sampling to still happen. This is akin to people brainstorming ideas.

I know that sounds confusing, let me break down how I think about this.

1. LLMs don't pick the token that ends up being used. This is by design, if the LLM gives a wide choice, it can better adapt to real world scenarios. i.e. generalize.

2. Without reasoning, this means that the LLM either locks in on whatever the sampler picked. Or decides mid-sentence/response to correct itself. This is what used to happen before reasoning, still happens if you turn reasoning off.

3. With reasoning, the LLM can make as many mistakes as it wants and explore its sampling space. Then use its vast pattern matching capabilities to decide which parts of the reasoning make sense and which were idiot ideas.

4. Enabling reasoning makes it so LLMs are much more confident on the final response, and the logits should theoretically all be near 99% on a single token for every token, i.e. much closer to greedy decoding. It analyzed all the possible options and figured out the best outcome, so a stray sample doesn't cause the answer to go awry.

This is why reasoning traces are filled with "but wait". I don't know if those were added in organically or artificially in the RL training, but regardless they're a good way to let the LLM keep generating other options and explore it's sampling space to the fullest.

Note: I haven't tested any of this and it's just my theory, but I'm sure if you really wanna know you can have claude run some smoke tests :)


PS5 emulation on x86 hardware will be more like running wine, so as long as your PC is as powerful as a PS5, performance should be close...eventually. The real performance bottleneck will be the GPU translation layer, much like converting DirectX to Vulkan isn't trivial, so isn't converting GNM (PS5's graphics API) to Vulkan, even tho the APIs are close (IIRC, they're cousin APIs)


The main problem is that PS5(and Xbox) has unified memory, that's the whole advantage of these systems. You can put your texture in one place in memory, have the CPU write to it and the GPU can immediately read it from the same address once you are finished. It's an absolutely huge performance advantage, that the typical PC architecture doesn't support and doing such a thing will always incur the penalty of having to transfer the data from system ram to GPU vram - with enough bandwidth and fast memory the issue can be somewhat mitigated, but it doesn't change the fact that on PS5 your GPU can start reading the data the CPU just wrote nanoseconds after it's done, on a PC you need to wait for the transfer first.


https://www.corsair.com/uk/en/explorer/glossary/what-is-resi...

"When a game is running, the CPU needs access to a graphic card’s memory, and by default it can only access 256MB at a time. However, resizable BAR removes the size limitation so the CPU has access to all of a GPU’s memory, theoretically allowing for improved performance. This adjustment lets the CPU request much larger chunks of data from the GPU, which can improve performance in games that can benefit from this change."


Yes, I'm aware of it. It's purely an addressing simplification and you are still limited by the GPU bus speed if you want to access any data on the GPU by the CPU. And it only works one way, the GPU cannot be given an address in RAM and expected to magically have access to it, the driver has to copy the data over first.


They exist to detect degradation. Datasets are not perfect and if a batch contains too much bad data it can ruin a run, also an opportunity to find bad data and improve the dataset filtering.


Same reason they go into low margin with the PS4/PS5 and other projects. Market penetration.


Youtube allows you to AB test with up to 3 titles and thumbnails.


The whole site looks like and reads like AI slop. The outcomes also don't make any sense and don't feel rigorously tested (no, having claude test for you doesn't count as rigorous).


The person is having a AI induced manic episode, we have all been there.


Please stop making this comment. The war is lost. Instead, you should be commenting that it looks like a human wrote this when you come across the rare brain-produced writing


I'm the same way, I have a very low/sporadic usage of any subscription I've tried. I now just use openrouter with DS4 pro/flash. It also gets rid of usage anxiety where I would try to justify the $20/month by forcing myself to use the tokens for projects as the weekly limit deadline neared.


Guidelines | FAQ | Lists | API | Security | DMCA | Apply to YC | Contact

Search: