AI Datacentres use about as much electricity as the NHS - go figure.

[bobbler]

I think the impact on peoples lifes and their jobs is being underappreciated, I think the livelihood of tremendously many people will vanish in a very short timeframe.

I'm sure there is some theory like game theory I just haven't looked up - but it looks like there is some giant 'gaming it out' across different countries trying it and doing different things that I'd assume someone is going to need to start doing

but it reminds me of the nuclear arms race reasoning

in the end you forget the why or the why nots and the whole decision ends up being built based on the 'but if/in case others do then we have to' seems to be dominating the selling of this.

I mean with AI for example there is the common quoting on programmes and reels of people saying the individuals like Musk, those running google, meta etc using the whole 'I have to get there first because if x does it first then it will be terrible'.

I suspect there is the geopolitics/countries and continents version of that too.

that classic 'what if' thing and whoever is proposing that 'what if' gets to control the framing of which bits to focus on currently without making the downside the macropicture too

 

[Felis Catus]

I used to be very sceptical until recently, especially of LLMs, and am aware of certain limitations, but the newest (non-free) versions of the leading LLMs are already surprisingly sophisticated, certainly sufficiently sophicasted for millions of people to loose their jobs, whether we like that or not. I've been suprised by how far you can already get with "matrix multiplication", when you have good data, despite it not being very sophisticated.

That's what I hear from my caregiver who works as a software engineer. Sometime last year, their employer pushed hard on the use of AI. Initially, I'd hear them saying something like "AI saved me 2h of coding today". Now it's "I spent 5-6h instructing AI what to do". Similar to what you say, they say the improvements this year have been massive, getting noticeably better every month. AI still makes mistakes which it can't correct itself but there are mistakes which it can correct. The teams are way more productive and the bottleneck is code review done by humans because there's so much code being generated. Senior software engineers are wondering how long before they are out of job.