Plankaluel

I don't think this is comparable. One is generally taking long-time exposures for astronomy pictures (hence the streaks).

So if the satellite is simply blocking light, in the worst case, it will block the light of a star, maybe for a few milliseconds(?) dimming the star in the final picture by a minuscule amount.

This might still be problematic (influencing spectroscopy? dimming the star slightly, influencing some measurements there?) but probably still better than potentially mixing in light from a difference source (?)

Regarding wind and mountains. Some perspective from someone from neighboring Tyrol:

The reason there is so little wind power: Probably the same reason the western, alpine parts of Austria have basically zero wind power - and why neighbouring Carinthia recently voted in a referendum to ban it completely.

People who live in the Alps generally don't like seeing the mountains altered. It is treated almost as sacrilege. And since these areas are heavily dependent on tourism, where the appeal rests on a romantic, Disney-fied fantasy of wild, untamed nature, locals worry that turbines would make the region less attractive to tourists. Of course, this "untouched" landscape is largely a fiction in the first place: most of it looks the way it does precisely because people have lived in it and shaped it for centuries.

An RTX 5090 for 20-30fps for the small model: That is not as unreasonable as I had feared :D

Yeah, that was my first reaction as well

I think this is just another case of "over-optimization to make shareholders happy in the end ruins everything". I.e., the normal enshittification problem.

Pretty sure all of that does make financial sense: - Being able to write 4k will bring people in to re-watching/watching the show for the first time. - Redoing the CGI, etc., would have cost a lot of money. - Very few people will cancel their subscription or stop watching because of stuff like that - So in the end, no one cares

I.e., it makes financial sense to do the minimum possible. Sure, if this were a project you care about, if it were your company that you are also emotionally invested in and maybe proud of, etc., things might look different. But your actual customers are shareholders, which in the end are predominantly giant ETF brokers and pension funds, that don't care about anything else but what your stock price looks like and whether you are in the S&P500. They probably don't even know what your company is doing.

Sorry, rant over ;P

See, that's why you should read the article, I guess :D So the influence is even worse than I thought ...

It's shocking how much pictures influence judgment: Without reading much, at first, I thought: Poor guy, maybe he got pulled into something, ...

Then I saw the pictures of him in a leopard fur pajama and indoor sunglasses, and with his (an assumption on my side) trophy wife, and thought: "Naah, he probably deserves it"

Just because something has been done for hundreds of years does not mean it has been scientifically shown to actually do more than placebo. A lot of people do a lot of stuff that has been proven to do nothing detectable, and they still swear by it.

"It has been done for hundreds of years" isn't a good argument. There is a reason "Appeal to tradition" is one of the more famous logical fallacies.

Also: The radiation doses used in this trial are very likely much, much higher than what you would get from such a bath in radioactive water (otherwise the water would be so radioactive that staying in there or even drinking it would kill you very quickly), so this doesn't really tell us anything about whether the traditional modalities do anything or not. And yes, stuff like that also exists in Austria with Radon caves, and many other places.

I have to say:

After all the talk in the article how the Romanian system makes part of the population perform higher by throwing a lot of the resources at them, lowering the performance of the people that are already below average. Which is a problem for Romania because a lot of the highly educated people leave the country afterwards ...

I was not really prepared for the final sentence where the author recommends as a solution that more countries should do it like that.

The thing about Arsenal is they always try to walk it in.

Yeah, it's a typical "startup research post", mainly there to have stuff to show to potential investors and customers.

Thank you for the tip! I just bought the (german) original which is fortunately still in print and very affordable as a hard cover.

You are running Qwen2.5 32b that has been fine tuned on data that was generated by R1

Not sure a TV documentary from 22 years ago is the source to get scientific explanations from ...

Once installed, the bridge can be driven along the road when a 100m segment is finished

GPT-4 is not a single model. The GPT-4 that was released initially a year ago is way worse in benchmarks than the newest versions of it and the original version has been beat by quite a lot of other models by this point.

The newest version of GPT-4 is probably still overall the best model currently, but it is only a few months old, and the picture depends a lot on what benchmarks you are looking at.

E.g. for what we are doing at our company (document processing, etc.) Claude-3 Opus and Gemini-1.5 Pro are currently the better models. The newest GPT-4 even performed worse than a previous version.

So to me it def. seems like the gap is getting smaller. Of course, OpenAI could be coming out with GPT-5 next week and it could be vastly better than all other current models.

This was wonderful. Thank you!

VoiceCraft, which has been released open source a week or so ago, can do the same with 5-10sec audio and is pretty convincing. It's pretty fun to play around with: https://github.com/jasonppy/VoiceCraft

Thanks! I was not aware of ellama. Maybe the problem is more one of discoverability :D

Super interesting and I will try it out for sure!

But: The mode of operation is quite different from how GitHub CoPilot works, so maybe the name is not very well chosen.

It's somewhat surprising that there isn't more development happening in integrating Large Language Models with Emacs. Given its architecture etc., Emacs appears to be an ideal platform for such integration. But most projects haven't been worked on for months etc. But maybe the crowd that uses Emacs is mostly also the crowd that would be against utilizing LLMs ?

Sure, but it seems like the US is doing the same for all other drugs except for aspirin so I assume it is not a huge problem?

This seems to be a US (UK?) thing? At least here in the German speaking countries in Europe aspirin also has nice round numbers. A dosage for heart attack prevention is usually 100mg. Normal dose is 500mg

Yeah, I think unless the authors are deliberatly misleading the public (which is of course a possibility, though not very likely I think) it is hard to imagine a scenario where all of this won't turn out to be highly interesting in one way or another (even if it might turn out to not be a superconductor)