Well yeah, so are neural nets. I just meant that these are engineering accomplishments, not scientific per se. Of course experimental science will often take advantage of cutting edge technology, including from computer science.
NNs have absolutely revolutionized systems biology (itself a John Hopfield joint, and the AlphaFold team are reasonably likely to get a Nobel for medicine and physiology, possibly as soon as 'this year') and are becoming relevant in all kinds of weird parts of solid-state physics (trained functionals for DFT, eg https://www.nature.com/articles/s41598-020-64619-8).
The idea that academic disciplines are in any way isolated from each other is nonsense. Machine learning is computer science; it's also information theory; that means it's thermodynamics, which means it's physics. (Or, rather, it can be understood properly through all of these lenses).
John Hopfield himself has written about this; he views his work as physics because _it is performed from the viewpoint of a physicist_. Disciplines are subjective, not objective, phenomena.
My personal theory is that Demis and John will win the Chemistry prize for AlphaFold this year and that they decided to also award this one to help bolster the idea that ML is making fundamental improvements in academic science.
I would prefer if there was an actual Nobel Prize for Mathematics (not sure if the Fields would become that, or a new prize created).
I did have a similar thought, but awarding a prize to AF this year would be a very bold move, given how things went today. Right folks, we've demoralized all the physicists today, tomorrow we can do the same for the chemists!
To be perfectly honest, I'm not really sure the physics community being demoralized by this prize award is a really negative thing. I think many aspects of HEP and other areas have stagnated, requiring exponentially more money for marginal gains (I don't have a problem with LIGO or other large projects, but "we found another high energy particle consistent with the standard models" not so much). Perhaps this prize will give the community a shot in the arm to move away from "safe but boring".
latexrun does a pretty reasonable job with LaTeX files, and only runs when needed, etc. Would be nice to have this integrated into a build system for plots, data generation, etc.
I don't think that 4o will actually be available for free. It seemed like they were quite careful in choosing their words. My guess is 3.5 is free without an account, and accessing 4o requires linking your OpenAI account.
They only mentioned 4o, but they mentioned it explicitly at the start, well before they mentioned one can also tie in to their openAI account, if you have one, at the end of the presentation.
To me that implies 4o by default, but I guess we'll find out.
Frequencies beyond the Nyquist limit alias back to resolved modes. So a combination of the camera plus a human eye to check that the flashing is “fast” (i.e. above the Nyquist limit) should be enough to verify the frequency.
A minor correction: anti-matter is actually regular matter. We understand it quite well, and are even able to create anti-atoms in the lab. On the other hand, dark matter is much more poorly understood: essentially the only evidence we have for its existence are observations of "weird" gravitational effects in the universe.
This has been my conclusion. I've spent the past few weeks polishing my yt-dlp scripts. Honestly, the experience of being able to watch any video I want offline and buffer-free is way better than the default YouTube experience. Sometimes I miss the comments section, but it is probably better for me that I don't have access to that.
I wonder if there is a market for a youtube downlod only mirror that creators can opt into, for this use case. Concurrent bandwidth requirements could be lower and no need for streaming tech. The service probably couldn't legally download from youtube and re-host, but if you made a desktop app that uploads to YT and the mirror service at the same time and mirrors the YT video ID on the mirror, that could be attractive to users.
All that said, how would such a service sustain itself? Probably ads. So really, same boat, different captain. I don't know.
Would you pay a subscription to be able to download your YT video queue?
...which just got a 40+% price bump. And has zero guarantee that it won't have ads, doesn't remove sponsored segments, and sends a pittance to the actual creators.
It already has ads, promoted videos are still in your feed and still promoted over your own preferences, whether you pay or not. And of course it doesn't stop creators from having sponsored content.
So it already doesn't stop the ads. It just stops pre-roll ads.
So cancel your premium subscription said some years later? At this hypothetical future time you’d be a lot more justified in piracy IMO, but today you don’t have much of a leg to stand on.
Grayjay[0] for android does this with their 'Polycentric Id'.
All videos from multiple platforms supported by the app have a special comment section only available through grayjay. Functionality is still pretty limited, but it's refreshing to write a comment and know that it won't be picked up by a bot and shadow-banned/deleted without reason.
Yep, a while back I grabbed yt-dlp and I've been going through my bookmarks and locally archiving useful (or nostalgic) YouTube videos. My 8TB hard drive beats Google's enshittification.
