Hmm, not so sure. He produced a digital signal, who’s spectrogram happened to be an image, and then played that digital signal to a bird. Dunno if a analogue spectrogram really even makes sense as a concept. The only analogue part of the chain would be the birds vocalisations, right?

That doesn’t sound great. What benefits do you see in mirroring this behavior

Well, guess I can’t deny such compelling evidence

As much of a prick as this guy is, I don’t think that’s true. The behind the bastards episode on him couldn’t substantiate it at least

Reverse proxy with mTLS in front might be a simple solution depending on your setup

https://www.youtube.com/watch?v=YhuWay9XJyw

Yup, that’s what I was alluding to, while it may not still be the case for transistors, they did manage to take 50 odd years to get there, push that trend line from the figure 50 years heh (not saying you should, 5 seems much more conservative)

Take a look at Nvidias pace wrt Moore’s law (of FLOPS) https://netrouting.com/nvidia-surpassing-moores-law-gpu-innovation/

Or like looking at the early days of semiconductors and extrapolating that CPU speed will double every 18 months …smh these people

Can you go into a bit more details on why you think these papers are such a home run for your point?

1. Where do you get 95% from, these papers don’t really go into much detail on human performance and 95% isn’t mentioned in either of them

2. These papers are for transformer architectures using next token loss. There are other architectures (spiking, tsetlin, graph etc) and other losses (contrastive, RL, flow matching) to which these particular curves do not apply

3. These papers assume early stopping, have you heard of the grokking phenomenon? (Not to be confused with the Twitter bot)

4. These papers only consider finite size datasets, and relatively small ones at that. I.e. How many “tokens” would a 4 year old have processed? I imagine that question should be somewhat quantifiable

5. These papers do not consider multimodal systems.

6. You talked about permeance, does a RAG solution not overcome this problem?

I think there is a lot more we don’t know about these things than what we do know. To say we solved it all 2-5 years ago is, perhaps, optimistic

Unfortunately not, here is a little kitchen sink type demo though https://myst-nb.readthedocs.io/en/latest/authoring/jupyter-notebooks.html

Myst-nb is probably the place to start looking btw - forgot to mention it in previous post

I use sphinx with Myst markdown for this, and usually plotly express to generate the js visuals. Jupyterbook looks pretty good as well

Devils advocate: Splatting, dlss, neural codecs to name a few things that will change the way we make games

Looks like the Barbican in London to me, it’s apartments and a public bar/drinking/working area, nice spot to hang out!

Something makes me uneasy about this being a Google sheet, you need to use credentials to view it and someone has a log of who has accessed it…you can probably even see who’s viewing it in realtime

Use an anonymous account! Or someone should host this on a website or something with higher privacy

I’m with you, I drew the line at calculators though. I can do the damn sums by myself!

Relative point to point

• which Blockchain are we talking here? How does it compare to the current banking infrastructure?
• again, which one? How does it compare to the current pricing?
• escrow is a thing, someone can build up a PayPal equivalent on top of a Blockchain, the list goes on
• the current system doesn’t do great here, some Blockchains makes it way more traceable, in fact
• skill issue, but also solvable with a PayPal equivalent
• not a fact, what does this even mean?
• does it?

You could say the Linux kernel is an astronomically terrible idea because it doesn’t do anything…but it is just the platform, the good comes from what people build on top of it that add all these quality of life features you miss

Buy ydy

I don’t really follow your logic, how else would you propose to shape the audio that is not “just an effect”.

Your analogy to real life does not take into account that the audio source itself is moving, so their is an extra variable outside of just stereo signal -which is what spatial audio is modelling

And your muffling example sounds a bit over simplified maybe? My understanding is that the spatial stuff is produced by phase shifting the LR signals slightly

Finally why not go further? “I don’t listen to speaker audio because it’s all just effects and mirages to sound like a real sound, what only 2^16 discrete positions the diaphragm can be in” :p

This sounds like a fun project “samples of chemistry”

There’s is a huge difference though.

That being one is making hardware and the other is copying books into your training pipeline though

The copy occurs in the dataset preparation.

Privacy preserving federated learning is a thing - essentially you train a local model and send the weight updates back to Google rather than the data itself…but also it’s early days so who knows what vulnerabilities may exist

You could try dexed, it’s a YamahaDX7 clone https://github.com/asb2m10/dexed/releases