The name seems an odd choice. In Greek mythology, humans who tried to climb Mount Olympus were punished for their hubris.

Aside from the poor name choice, the other choice that looks pretty iffy is the size. Many, if not most, of the recent papers on the scaling of LLMs suggest that making them bigger does not necessarily make them better and often make them worse. Researchers like Emily Bender and Gary Marcus have been saying this for a while; last year DeepMind demonstrated that training smaller models on more data actually led to better results. It looks to me like in this case Amazon may have access to huge amounts of hardware, but they may not have huge amounts of insight. In that respect I guess they are sort of like most AI systems.

Setting aside the fact that the IDE is in fact $99 for personal use in the first year, and gets cheaper each of the next two years, and there's a very capable free edition which has all of the core features... there are plenty of tasks involved in coding that are tedious even in the best of languages. It's rare for any language to integrate an editor and have a debugger integrated into the editor. Good IDEs support efficient refactoring of existing code. Automated generation of unit test stubs, and one-click running of the tests, is not really something that is (or IMHO should be) a core language function. As a result developers benefit from a good IDE even when programming in easy-to-use languages.

So no, it's not a mixed message at all.

You beat me to it. JetBrains' "subscription" model pretty much the best of both worlds. You can effectively buy the current version, and a year's worth of updates, in perpetuity, for the price of a one year subscription. If they release something new that you really want then you can renew for another year, but you can delay that decision for as long as you want if there's nothing you feel you need.

True, but sadly Buffalo buffalo buffalo Buffalo buffalo Buffalo buffalo buffalo too, and so the cycle of abuse continues.

Supply of the new Pi5 is still very tight, since they are only just ramping up production. I just checked three US-based retailers (SparkFun, Vilros and PiShos.us; all three of them had 4GB Pi4 board in stock and only one limited the number, with a maximum of 5 per customer.

I know that there's a grand tradition of posting responses on Slashdot without knowing what you're talking about, but just in case anyone here wants to know what the law suit actually says before commenting, you can find the court filing here.

Actually US healthcare costs are so high because about 45% of the US healthcare spend goes to "the intermediaries—insurers, chemists, drug distributors and pharmacy-benefit managers (PBMS)—sitting between patients and their treatments". Most of this doesn't happen in countries where healthcare is viewed as a centrally managed social good rather than a source of profit.

According to the AP:

So not only did the prior felonies include violence and attacks on the police, but he was also well aware that he wasn't allowed to own a firearm.

They should replace it with synthetic hydrocarbon fuels made from CO2, water and energy from non-carbon sources, including nuclear as well as renewables.

The underlying technology for synthesising hydrocarbons as been around for 100 years and is well understood. Oil companies have huge numbers of chemical engineers and are perfectly capably of deploying it at industrial scale, if they have an incentive. Banning exploration would be an incentive to start. Banning new wells would be the next step, followed, eventually, by banning extraction altogether.

It seems to me that trying to get rid of hydrocarbon fuels is a big mistake. They are a very efficient way to store and transport energy and there are not many viable alternatives for aircraft. Moreover, there are 1.5 billion cars in the world; the overwhelming majority run on petrol or diesel and they are not going to be replaced with EVs any time soon. The problem the climate faces is the result of new CO2 being added to the atmosphere, not from the use of hydrocarbon fuels per se. Fuels that are made from CO2 that is taken out of the atmosphere using renewable or nuclear energy are net neutral.

Getting the oil companies to switch from pumping oil out of the ground to making oil from CO2 and H2O has the potential to make 1.5 billion vehicles carbon neutral without the trillions on capital expense to replace them all, which means that it could happen much faster. It allows existing infrastructure to be reused and, if done right, would make the oil companies part of the solution. While there are some people who think oil companies are inherently evil and should be destroyed, not giving them a role in solving the problem means that you will have trillion-dollar companies fighting against change tooth and nail, because their very existence depends on it. If instead you give them a path to making a profit without pumping new oil from the ground you can have all those resources working in ways that will help the planet, not hinder it. Banning new exploration for oil is the first step in incentivising that transition.

While companies like BP and Shell make most of their money from fossil fuels, they see the writing on the wall and most oil companies are investing heavily in renewables and alternative fuels. So the key question is what sort of university research are they spending this money on? If they were spending it on finding new was to frack or suck more oil out of the ground then they would be shameful hypocrites. If, as I suspect, they are spending it on finding better ways to capture renewable energy or synthesise alternatives to fossil fuels then it could be hugely beneficial to the planet.

From a climate point of view, the problem we have is not the use of hydrocarbons as fuels but the extraction of more and more carbon from the ground. Getting people in rich countries to buy new electric cars is fine, but there are nearly 1.5 billion cars in the world already. If we want to accelerate the move away from fossil fuels then we need to find a way to power most of these vehicles with compatible fuels that come from non-fossil sources. Banning new internal combustion engines will take a very long time to have an impact. Telling oil companies that you're going to progressively stop them from pumping oil out of the ground, so they better put their thousands of highly skilled chemical engineers to work on switching their refineries to fuel synthesis would have more of an impact more quickly. It would also provide a path for these politically influential companies to be allies in the effort rather than enemies. If they want to pump money into universities to do that then I'm all for it.

These days most industrial titanium production is done with the Kroll process, where TiCl4 (obtained as above) is reacted with liquid magnesium in an inert atmosphere, and the resulting magnesium chloride drawn off in liquid form. The bulk of the energy used in the whole process is consumed converting the resulting MgCl2 back into magnesium and chlorine. Current estimates of the end-to-end energy consumption of this process that I could find online suggest between 295 to 420 MJ/Kg. This compares pretty dreadfully to the numbers for the Bayer process for extracting aluminium from bauxite, which Wikipedia puts at between 7 and 21 MJ/Kg. So per unit mass, titanium is 20 or 30 times worse than aluminium.

Having said all that, if you look past the marketing headline at what Apple actually announced, you will see that the iPhone 15 Pro isn't made from titanium. It has an aluminium chassis with a titanium band around the outside for strength. From their press release:

I suspect that the total mass of titanium in the device is pretty small, but provides useful strength. Coupled with using recycled metal for the aluminium, it might even have a lower carbon footprint than previous models.

This sort of use of face recognition, and the failure of the ' "very poor" police work' that the chief referred to, are a classic case of users of technology not understanding the technology.

There are around 650,000 people in Detroit. If facial recognition has a false positive chance of (very optimistically) 1 in 10,000 then you still expect 65 hits. A similar thing happens when the police scan DNA databases; even with a supposed 1-in-a-million chance of error, any test will throw up 350 hits across the population of the USA.

If the use case is "Jane Doe is our prime suspect, she was seen in the area and she had opportunity and motive", and then you run a test of a photo/DNA sample from the crime scene against the suspect, then a hit is pretty compelling. If you have no idea who the suspect is and you trawl the entire city/county/country for a hit then the chances of it being wrong are very high indeed. Sadly it seems the lazy police officers sometimes mistake confirmation tools for search tools. When they do, this sort of stuff happens.

Aside from the magical thinking that somehow they will be able to create a mechanism that allows access only by a "competent authority" (which they won't), the issue that the UK government seem to be totally missing is that in an international market, they are not the only authority. Warrants won't just be issued by some fine, upstanding justice in the Crown Court, they will be issued by CCP endorsed judges in Hong Kong, oligarch owned judges in Russia and Taliban appointed Sharia mullahs in Afghanistan. British citizens and British government workers travelling abroad will be subject to warrants from these sources and many others, and they will be silently enforceable.

While the stated goal of this legislation is admirable, it's not going to work, and it's going to weaken the security for everyone, including the UK Government itself.

From the WSJ article:

Any semblance of credibility that they had (and it was slight to start with) evaporated when I learned that their qualifications are successful quests for loose change and unsuccessful quests for the Holy Grail. That said, they deserve credit for the ingenuity of their publicity stunt for the TV shows.