And fission plants definitely aren't cheap.

Especially if you include the unsolved issues such as waste barrels stranded on the beach.

"But but no true Scotsman would put barrels..."

Just shut up.

I guess that proves that even this site is using AI to write their summaries these days.

Regarding b. They get 21% of their revenue from government. However, they get 61% of their revenue from Starlink. Now, I could pay myself 100% of my salary and thereby earn double what I am earning now... except that's preposterous. So, if we don't count SpaceX paying itself, it's still getting more than 50% of its revenue from government contracts. (because if we take out that 61%, then you have 21% from government contracts and 18% from other so, out of that remaining revenue the majority is still government contracts).

Stephen Miller is the grandson of Pogrom survivors, and yet...

I think you're conflating Elon Musk with Tom Mueller. Where are these supposed whitepapers that Musk himself authored?

OpenAI was remotely watching his work as he used their tool, and when they determined his stuff was interesting, they grabbed it and published first. They basically piggybacked on the work of a mathematician, for 99 percent of a marathon, and then used their massive compute advantage to sprint the last 100 yards to take the first place medal. Absolutely. No. Surprise. Too bad for the guy - ouch hard life lesson.

To be fair, I think most people are just token predictors.

I'm just going to quote that here. I could write out a history of electric cars to scrutinize it, but what's the point really?

They did say a biology researcher. Whether that's corporate or university, they probably do have the computing resources available and probably shared among many researchers. Of course, you do run into the issues that, these days, those resources are probably running on cloud servers. So still quite possibly not actually local.

Yeah. Even just doing casual research you run into troublesome censorship of information issues with these LLMs. For example, I was interested in the whole process of production of pennicillin and I had a long discussion going over all the principles from scratch: how to find, isolate, assay, and force evolve the most productive strains of pennicillin looking at how things were done historically, how to grow in aerated flasks with continuous feed nutrient systems (with discussion about corn steep liquor being a traditionally ideal feedstock because it both feeds it, but also keeps it under stress so it ramps up production), carefully decontaminating all inputs and isolating the output through a drip system through an oil barrier, using amyl alchohol, and sufuric acid for extraction along with how to actually produce those, etc. Basically it was a very interesting and enlightening session. However, then I tried to more or less restart the conversation on another occaision to clarify some points and I started off with questions incorporating essentially a summary of the first session. The LLM basically refused to because I was asking so directly and it was worried that I was actually trying to set up an at-home penicillin production lab.

So, basically, while you can get information bit by bit from these things, a whole web of probably mostly liability-based safeguards get in your way. That is one of the real concerns (to add to hallucinations, etc.) with AIs taking over as the interface to human knowledge: they will censor it. Not necessarily for sinister reasons (though I have caught LLMs directly presenting corporate press releases to certain questions rather than generated answers, which I consider pretty sinister), but just for the same reason people who frequent certain heavily corporate forums have started to say "Yahtzee" and "unalived", etc. rather than the real terms even outside those forums. Basically, we have to worry about access to information becoming a compartmentalized, dumbed down, lowest common denominator situation.

How is OpenAI morally right?!?

Best case for OpenAI is that they heard rumours that a researcher was closing in on a NS breakthrough using a specific method, so they instantly set a team on the problem giving them with a massive amount of resources. Their internal team then specifically guided the model to use the same very specific direction the researchers was using, something they (nor the model) would ever have known to do without the rumours. A detail that OpenAI conveniently left out of the narrative.

And that's the best case scenario. The worst case is that because the researcher was using ChatGPT to work on drafts of the paper and those drafts ended up in the training data for the model. So rather than discovering anything the agents just implemented / finished off the work they already had access to.

It's insanely sketchy. The NYU researcher is honestly being extraordinarily calm about the whole thing, typically in a situation like this people would be getting close to a fist fight.

Obviously, if you approximate a complex system containing 1e30 quantum particles using a couple of differential equations and a few boundary conditions, you're not going catch every detail.

The actual two main factors that cause oil prices to be high are the war with Iran, and the simple fact that we're nearing the end of oil as the dominant power source. Technically, we're already past traditional "peak oil" as it was originally defined. Because of advancing technology the peak oil target has moved, of course, so there's plenty of accessible oil, but the actual average difficulty of both extracting the oil and processing it to be useful has increased. We're at the point where new wells in the US need a price point of close to $70 per barrel to be economically viable, and that's just going to go up. There are places in the world where the price point for new wells to be profitable is lower, but a good portion of those are in the Middle East. Of course, the cheap cost to extract and process Middle Eastern oil is a bit of an illusion since those states actually absolutely rely on a much higher price point to actually fund their governments since they are critically dependent on oil. If we look at Venezualan oil, which Trump has pointed to as some sort of economic savior of oil prices, then it gets even worse with new Venezuelan wells requiring a breakeven oil price of something like $80+ per barrel. Meanwhile there's a demand crunch coming as we get more and more cheap renewables displacing oil in the energy market. Since Oil is a high capex industry, it can't scale down easily. So a demand reduction really, really hurts the industry. Basically, the industry knows this, but it's in denial, but it still means that interruptions like the Iran war can cause the industry (and therefore oil prices) even more damage.

In other words, a perfect candidate for a pardon from this admin, as long as her husband’s family coughs up the appropriate “donation”.

The two studies aren't very representative of what I consider mainstream users. Both studies appear to be run on a Mac Mini running Firefox 146 (current version is 155). CPU and memory usage appear to be more dependent on the OS than the individual browsers. Very little data is available for Brave Desktop on Windows, or Brave for Android on mobile devices.

Claiming a program needs a specific amount of memory for their tests also seems OS dependent. My running copy of Firefox with multiple tabs and private windows is only using half of the claimed 1.65GB for Firefox (847MB). Open windows in Amazon, Slashdot, and Google don't need the memory amount their tests claim is needed for their test page.

Granted, I am running both uBlock Origin and Privacy Badger so some of my memory usage might be reduced by those tracking and ad blockers, but I think most users of any browser will have installed similar blocking software when available.

If their claim is the built-in blocking features of Brave are better than the clean installed versions of Chrome, Edge, and Firefox, then the test results should be a little more specific about their conclusion.