Ryan Tremblay

Training music models at (hiring researchers!) - Opinions *directly* from orgs I'm associated with.

[Ryan Tremblay @zaptrem]

@griffinli_ @jjuulliiaallii @ericgkrehbiel @hackneyapp Do the tokens you're collecting allow you to theoretically book rides as me?

[Ryan Tremblay @zaptrem]

@jjuulliiaallii @ericgkrehbiel @hackneyapp I thought they all had account-dependent pricing?

[Ryan Tremblay @zaptrem]

@difficultyang I assumed it was at least partially deployed because the docs reference it now: https://t.co/GNkReVWtZK

[Ryan Tremblay @zaptrem]

@MatternJustus How many requests were rejected or swapped to Opus 4.8?

[Ryan Tremblay @zaptrem]

@ZDi____ Awesome, you should do a lot more patchification of the mel spectrogram then use a diffusion head like VAR and you'll have a great long-duration tts model.

[Ryan Tremblay @zaptrem]

@difficultyang Haven’t tested yet but with opus 4.8 in the web ui the answer was definitely no.

[Ryan Tremblay @zaptrem]

woooo spelling

[Treblo @TrebloApp]

Sonauto is now Treblo!

https://t.co/MF6pc9fx7W

Your account, your songs, and your shared links all still work. We're the same four people working towards the same goal.

Check out our blog post for more details: https://t.co/9Y0mPF0SU0

[Ryan Tremblay @zaptrem]

@epstein_stylist @difficultyang @Smit_Chaudhary3 The problem is the ML research classifier is invisible. I have no way of knowing if the classifier has improved because they don't tell me when they turn on Loboto-Mythos mode, it just happens silently.

[Ryan Tremblay @zaptrem]

@difficultyang @Smit_Chaudhary3 And what about those of us working on powerful models for other domains -- those can't and never will turn anyone into paperclips -- but get caught in the "frontier" filter all the same? (or maybe we don't, but we're stuck constantly looking over our shoulders now)

[Ryan Tremblay @zaptrem]

Unironically evil behavior. How can we even tell whether mythos has decided our LLM music model counts as "frontier LLM research"? We know models can be overzealous about declaring requests are unsafe/bio-uplift/etc, but now it includes ML research and won't even tell us?

[elie @eliebakouch]

mythos will be bad ON PURPOSE on ai "frontier llm research" tasks, this is very very sad for the research community

also the fact that this is un purpose not visible to the user is crazy

[Ryan Tremblay @zaptrem]

The composition here and the kid’s facial expression gives Norman Rockwell vibes

[jonathan liu @jonathanzliu]

holy fucking shit

[Ryan Tremblay @zaptrem]

@___Harald___ "little practical use" is this ragebait

[Ryan Tremblay @zaptrem]

@ZDi____ Just give Claude Code/Codex your node’s IP/SSH key and you can have this now? I do it every day

[Ryan Tremblay @zaptrem]

@rosmine Where did this blogpost go? Seems to be gone

[Ryan Tremblay @zaptrem]

@oops4041555 What is a fallback paywall?

[Ryan Tremblay @zaptrem]

An AR model takes 2N flops per token/timestep. A diffusion model is the above *times the number of steps* (or even more due to bidirectional attention). Unfortunately AR mogs diffusion in training *and* inference compute efficiency.

[David @DavidSHolz]

Most researchers agree that autoregression is best when memory bandwidth is cheap and diffusion is best when FLOPS are cheap. They also admit the future of compute is all FLOPS because memory scaling is hard and scaling FLOPS is easy. So why not go all in on diffusion????

[Ryan Tremblay @zaptrem]

@greentoepalacio Seems like he was kinda effed over by lane position and the pool lacking buffer lanes

[Ryan Tremblay @zaptrem]

@ZDi____ Have you tried just doing off the shelf spectrostream? Seems to work fine for lyria….

[Ryan Tremblay @zaptrem]

@jchencxh The problem with pixel prediction (noise) hits hard in my domain (audio is dominated by stuff we don’t care about). Re: “if there’s more signal, keep learning from it” you can do that with latent prediction too with stuff like hierarchical prediction (newest vjepa?)