Carriers are irrelevant now. It's the OEMs (especially Samsung) who have the power. Samsung would kick Google to the curb if Google tried to take over.

How many companies used llama instead of paying OpenAI for API access?

If Meta saw OpenAI as a market competitor then denying them revenue is a cunning move.

Especially if llama isn't Meta's core competency but ChatGPT is OpenAI's.

Maybe even legal in that aspect, though they both seem super guilty of not paying for copyrighted material.

To the extent that the situation you refer to is a problem, it's a problem of market share and the resulting funding for ongoing development of an open source OS. Google's ability to enforce requirements on Android OEMs is limited because the big players or any significant consortium of the smaller players can simply choose to cut ties with Google if Google is too pushy.

Yes, Chrome established a different business model from the outset. Android went a different direction because, rightly or wrongly, it was believed at the time that it was necessary in order to fend off other participants in the smartphone ecosystem, and over time it has gotten harder to change the model, not easier. In particular, one major Android OEM has amassed so much market dominance that they can and often do simply refuse Google's requests. Legally, Google could cut ties, but that would be bad for Google and i think it would be bad for Android users, since it would instantly fragment the ecosystem. IMO, Android users (and I am one) are better off with a slower-moving but relatively unified ecosystem.

No, this is about the merger of the platforms. It probably will eventually result in some reduction in staffing, but it's not happening now, and hasn't happened in any significant way. Both Android and ChromeOS have been relatively untouched by layoffs.

The goal is to reduce engineering effort. Even Google has to consider costs.

This was announced on June 12, 2024.

It doesn't mean Android and ChromeOS will share a common UI. Android already supports several distinct user interfaces for different platforms (mobile, wearable, auto, TV), and there's lots of customization even within those spaces. I expect that once the transition is complete, ChromeOS will still look and act much like it does now. It may run Android apps a little better than it does now (though it already runs them fairly well). It'll just share a lot of infrastructure with Android underneath the surface.

From what I find online, microdoses of psilocybin are generally 0.1-0.3 g of psilocybe cubensis, which would give a dose of about 1-4 mg of psilocybin, not 15 mg.

The COVID mRNA vaccines were the culmination of decades of research into genetic vaccines that could be in essence engineered to target a selected antigen without the years of trial and error that are required by the methods we have been using since the 1950s. Within days of the virus genome being published, they had a vaccine design, the months it took to get to the public were taken up with studies of the safety and effectiveness of the heretofore untested technology, ramping up production, and preparing for the distribution of a medicine that required cryogenic storage.

It would be unreasonable not to give the Trump administration credit for not mucking up this process. But the unprecedented speed of development wasnâ(TM)t due to Trump employing some kind of magical Fuhrermojo. It was a stroke good fortune that when the global pandemic epidemiologists have been worried about arrived, mRNA technology was just at the point where you could use it. Had it arrived a decade earlier the consequences would have been far worse, no matter who was president.

The lesson isnâ(TM)t that Trump is some kind of divine figure who willed a vaccine into existence, itâ(TM)s that basic research that is decades from practical application is important.

Currently have a project going using Python scripting in Blender using scipy.optimize.differential_evolution and Cycles rendering to optimize the shape of a reflector to match a desired light distribution pattern. It's not a perfect tool for the job, but it seems to be pretty accurate.

Also, the mice got something like 500mg of psilocybin per kg of body mass. For humans, 280 mg/kg is considered a lethal dose (LD50). It's really unclear how this research could transfer to humans.

OTOH, it's a starting point. Rather than concluding that this means humans should trip on massive doses of shrooms to live longer, we should think that further research may elucidate the specific mechanisms and yield other insights that can transfer -- and might even be vastly more effective.

I agree on the matter of authority... but if you read the link, it largely suports what garyisabusyguy said. The link says:

the most commonly used mushroom is Psilocybe cubensis, which contains 10–12 mg of psilocybin per gram of dried mushrooms

Which is exactly what garyisabusyguy said.

It also says:

For example, if you want to consume 15 mg psilocybin (a common dose) from cubensis with 1% psilocybin content: 15 mg / 1% = 15/0.01 = 1500 mg = 1.5 g

But it also says that "strong" and "heavy" doses are 2.5-5g (25-60 mg psilocybin) and 5+g (50-60+ mg psilocybin). There's also a bit of inconsistency on the site, because if you look at the page devoted to Psilycybe cubensis, it gives different, slighly larger numbers. It says a common dose is 1-3g, a strong dose is 3-6g and a heavy dose is 6+g.

That all accords pretty will with what garyisabusyguy said, assuming his experience is with people who take doses at the high end of common and greater.

Of course, his ranges still suggest a maximum dose of ~84mg. A typical lab mouse weighs about 30 g = 0.03 kg, so they're taking a dose of 15 mg / .03 kg = 500 mg of psilocybin per kg of body weight. If an 80 kg human takes an 84mg dose, that's 1.05 mg of psilocybin per kg of body weight. So the mice are getting 475 times what appears to be a quite heavy dose for humans.

Further, the LD50 (dosage that is lethal 50% of the time) of psilocybin is 280 mg/kg of body weight. So the mice in the experiment got nearly twice what is usually considered a lethal dose in humans. It's unclear to me how or whether this can apply to humans.

see subject.

Coding is to CS as math is to Physics.

Software engineering is building the supercolliders for physics.

Meanwhile we have LLM's declaring themselves Mechahitler.

Good job University of No Job Placement.

(yeah, some Universities have been giving CS degrees for HTML and MS Office)

Good one. :)

If we can't move the planet by then we've gone extinct too.

According to the Timcast review (spoilers) he never says either phrase in the movie.

Just some dope making posters, reportedly.

Apparently all the immigrant and Israel propaganda stuff is fake too.

Marketing Buzz achieved - don't believe anybody without verifying.

What is the current total Federal budget for science across all departments?