Because we know the Earth to be slowing in the long term, so an increase is speed has to be a short-term glitch. How short term? I don't see it stretching beyond a second , unlikely to stretch to tens of seconds, and a UTC/UT1 difference of even 10 seconds is unlikely to be problematic.

On the other hand, we've only had clocks accurate enough to measure this for under 100 years, so saying that we understand Earth's rotation might be a stretch.

The digital world is chasing something.

Unfortunately that something is mostly decided by marketing and fetish.

How many millions of lines of code are required to render this post on your screen? Apply that to all webified apps, even the ones that run locally.

We are so far removed from the metal at this point, the only rational direction is backwards.

It remains to be seen if AI-branded computers will outsell anything.

If it doesnt have an AI key then you need to replace it anyways

No, my argument is that the reduction is so tiny that it's meaningless. It's temporarily reducing the rates from, say, 3% to 2.96%. In reality, it should be reducing the fees from 3% to 0.5%.

Even if it only took a couple months for someone to be proficient with training and tuning models, that's too slow for a lot of the companies making these hires. It's a gold rush to be on top of this "next big thing", claims need to be staked out immediately. It doesn't matter that there might not be that much gold in the ground. A late mover isn't going to get any of it, so companies have to scramble to be there.

That's the entire point. They would, but they can't, and this settlement didn't do anything to fix the problem. Or any problem, really. The 0.04% reduction in rates will make literally zero difference to anyone.

The moderator must have the power to insert advertisements because a fee just aint going to work.

It was a limit set on what was supposed to be a temporary UI at a time when drive sizes were measured in megabytes, not gigabytes.

Prompt your way into a useful modernized UI layer in a language that you do not know.

Sounds tricky.

These things are in the first stage of being investigated, and we are still in the stage of asking, "Is this a good thing to do?". This research comes down on the 'possibly not' side, with a solid 'But more research' caveat.

We know what we should do, and that's to stop burning stuff. That isn't happening because burning stuff earns people money and makes people's lives more convenient.

For a more practical take, try https://what-if.xkcd.com/162.

If the second crash hadn't happened, Boeing could have got away with it. If the Ethiopian pilots had the good luck not to have been at a high throttle level when MCAS triggered, or had recognized it as a trim system failure abnormally quickly, Boeing could have created and pushed out a software update a few months later and saved their shareholders a heap of money.

This would have allowed them to continue on as they had before, which is preferable in this journalist's eyes.

... and do the corrective action straight away. It only took a few seconds delay for the situation to develop beyond what the simple remedy covered.

What I conclude happened - the stuck sensor caused a sudden stick shaker and stall warning, masking the sound of the trim actuating. The airplane felt out of trim, so the pilot applied trim manually, setting the badly programmed system up to push the plane way out of trim.

MCAS was programmed to only apply a little downward trim, but that reset if the pilot applied manual trim. So MCAS trimmed 3 down, pilot applies a bit of up trim, then pauses to see if he has done enough. MCAS puts in another 3 (really fast), the pilot trims up again, maybe 1. MCAS puts in another 3, etc. Soon you are way out of trim.

So all of this leads to a delay in identifying a trim system failure, and so a delay in applying the cut out switches, while the plane speed gets out of hand (Note, you are hoping they'll pull the throttles back while they are still hearing a stall warning!). Pilots overworked until the situation is beyond their training, and then you are relying their skills as test pilots.

Intel has never been a successful player in the foundry market. They've made some prior attempts to enter the market, but they all failed. Their most recent attempt was in 2014 with "Intel Custom Foundry", but the only major customer they secured was Altera... who Intel purchased in 2015. They gave up on that attempt by 2018.