The underlying classifier, BirdNET, is a traditional neural network and not an LLM:

https://doi.org/10.1016/j.ecoinf.2021.101236

Why would anyone assume this uses an LLM? It classifies bird sounds, not human language. I don't mean this as an attack, I'm genuinely curious! This seemed obvious to me, and I want to know what line of thinking might lead one to believe that an LLM is the better (or more likely) tool for this job over a purpose-built classifier.

Are there any LLMs being widely used for audio classification? I know VLMs are being used a lot in image stuff.

It always seems kind of silly to me to throw everything at an LLM. I know they’re huge and can automatically handle a huge number of tasks but something in me finds it wasteful when we could be creating easily trainable, cheap to run bespoke models for a lot of stuff

There are LLMs that support audio input, similar to those with vision support.

From my testing of open weights LLMs with audio support, they basically are only trained to recognize audio as an alternative to text input, they treat audio as basically equivalent to a transcript, and can't recognize or distinguish things like music, accents, background sounds, etc.

So they're only really good for transcribing or summarizing or using audio input in place of text input for prompts, but not anything that requires distinguishing any information about the audio that would not be present in a transcript.

It can be tempting to try to use an LLM for a variety of tasks; kind of the whole thing about an LLM is that you don't have to do a separate complex training run for every task, but can just provide instructions in natural language. But it only works as far as what the training data covers, if the training basically always treated audio and a text transcript as equivalent, the model has nothing causing it to learn other relevant features of the audio. If there's enough bird call identification in the training data of an LLM, it might be able to do that, but I think multimodal training data tends to be much more limited than the text training corpus

Have you had any luck fine tuning one with musical data for classification or music-aware QA? I've been hacking on https://trebel.la/ which I would like to be a music practice companion, and the biggest missing feature is actually useful audio-based feedback pipeline.

My current approach, not yet validated, is trying to generate training data from masterclass recordings on Youtube, and then fine tuning MOSS-audio on a bunch of those. But I'm interested if there are better models, or large training sets I don't know about.

You're selling a music feedback app and it doesn't do the music feedback part yet?

The harness that connects to a chatbot, API or voice interaction is the place to route requests to different systems. If you remember the early days of ChatGPT it explicitly said it was routing image generation to Dall-E after embellishing your request itself first.

Determining which tool to use should be a lightweight operation but I’m not expert enough to understand exactly how much lighter than a full LLM call just to recognize it needs a different tool or model.

For sure, I just know it’s tempting given the power of large transformers to throw things at an existing model.

For instance, OCR is something that can be done locally with no access to a GPU but people (including me) still often use cloud hosted multi-modal large language models for it.

wav2vec or similar approaches are used a lot these days, which is basically BERT with audio inputs. Whether that counts as an LLM or not, I am not sure. It's a transformer architecture in any case.

People will often reach for "easily trainable, cheap" solutions when they can; the reason people reach for Transformers and LLMs is because when you throw more data at them, they get better.

More info please. Can I add this into birdnet go for more species detection?

Note that while the underlying birdnet-go project started as BirdNET only, it can now use Google Perch v2, BattyBirdNET (for bats!) and other models in the future. It's a really cool project!

Curious if there is a similar project for dog barks.

I wanted to say that seems like a stretch but then I often find myself visualizing my best guess at the appearance, including species, of an unknown dog. It seems like size of dog and pitch of bark are negatively correlated

fm = Foundation Models

https://developer.apple.com/videos/play/wwdc2026/334/

Quite an interesting selection committee:

  Neal Stephenson
  Gwern Branwen
  Matt Huang

Right? I'd submit something just for the chance of lodging an idea in the head of Stephenson or Gwern.

This geographic puzzle feels like the modern day version of the Seven Bridges of Königsberg:

https://en.wikipedia.org/wiki/Seven_Bridges_of_Königsberg

It was a $1 million donation to Vanderbilt in April 2020 to support three COVID research projects, including one related to the Moderna vaccine:

https://www.npr.org/sections/coronavirus-live-updates/2021/0...

Thank you

Dolly Parton’s Imagination Library has been gifting free books to kids from birth to age 5 for over three decades: https://imaginationlibrary.com/

It’s a matter of preference: Everything OP says about GUI development getting easier is also true about TUI development getting easier.

> What makes Github special is the ecosystem and integrations. Sentry, Linear, and so many other tools have special UX around Github. You need centralization (or at the very least, a standard) for that.

I think “a standard” is the key here. There is no reason why integrations must work through a proprietary centralized service.

You can set latitude and longitude after opening settings with the gear icon.