potatoman22

Reminds me of https://github.com/lemonade-sdk/infinity-arcade

I really liked this, thank you

I think a good "traditional" approach would look like a BM25 algorithm over an index of trigger phrases for each category, sitting behind a majority-vote classifier. The "fine tuning" would be done by reindexing the data, generating different/new phrases, and tuning the classification threshold.

I think that's to demonstrate its speed

Adding this to my homelab, thanks for the web browser option. How does your son use the backing tracks? Does he mute bass/drum, or play along with them?

I have about 60 hours of my own playing in midi data. I wonder if I could fine-tune this model on that?

It's not specific to SQLlite per se, but that's what most dashboard builders are

I was trying to say there's precedent for using ML in this field. I don't think they even said this product is an image generation model. It's probably some ML version of this https://en.wikipedia.org/wiki/Ultrasound_computer_tomography

I also found this researcher on their staff who studies tomography https://scholar.google.com/citations?hl=en&user=idvD2yYAAAAJ

Generative models have been used in healthcare for a while for things like drug design and data generation. Not to mention all the algorithms (and probably ML) used in generating results for MRI and CT scans. I don't think this is that crazy provided they can prove it's effective.

I'd wager that for 99.9% of "Apple Intelligence" tasks, Google's models perform just as well as other frontier labs. Google also has done more work on getting LLMs running on edge devices compared to anthropic and openAI.

The source also says > The new architecture centers on Apple Foundation Models co-developed with Google, which Apple says are adapted to run both on-device and on servers through its existing Private Cloud Compute infrastructure

Which could mean Google and Apple have trained some custom models, probably the on-device ones, specifically tailored towards Apple's hardware.

Picking an equal weight fund is closer to picking specific stocks than investing in the S&P500 imo

I wonder why they didn't use a Bonsai model as the text encoder

It reminds me of the Nord theme https://github.com/nordtheme/nord

Google should know why a human accepted the automated suggestion, or if and why there wasn't any human oversight in the first place.

This reminds me of control vectors, especially this line in the linked DwarfStar repo:

> y = y - scale * direction[layer] * dot(direction[layer], y)

From https://vgel.me/posts/representation-engineering/

> A control vector is a vector (technically a list of vectors, one per layer) that you can apply to model activations during inference to control the model's behavior without additional prompting

Neat! It seems like Qwen 9b took the same amount of time as gemma4-e4b too, which is interesting. I haven't been able to get Qwen to stop thinking so much

Not to be nitpicky, but many of the 4-12b models are somewhere between GPT-3.5 and GPT-4o-mini. It's hard to find a good comparison though, because the benchmarks people score models against change so often. For reference, Sonnet 3.6 came out about a year after GPT 3.5

The website has a "chat" link near the top

There's a point where they meet, but "faking it until you make it" doesn't work for productivity in the same way it doesn't work for getting rich.

But there's a secret: just buy my $399 masterclass and I'll teach you 17 simple productivity hacks to 100x your income.

Bloomberg made a good video about this https://www.youtube.com/watch?v=-gjnrMg9iSo

How did you install parakeet? It was a nightmare to install on windows

This project is called Levin, so Anna Kareninina. However, I learned Anna (as in the archive) is a pseudonym, so this is probably not the case.

Great name haha. Is Anna a reference to who I think it is?

I'm a software engineer, but I use it to write down hypotheses: the cause of a bug, how I'm guessing a system works, potential fixes for said bugs, what I think this piece of documentation means, etc

What theme did you use? I really like the "garden" theme

Someone should analyze this and share results. The data should be there

I've definitely noticed an association between how much I vibe code something and how good my internal model of the system is. That bit about LLM users not being able to quote their essay resonates too: "oh we have that unit test?"

I wouldn't say so. Audible gives you 1 book a month for $15. Most e-books I see are around $10.

You need to buy an apple device