Well there's some psychopathic phrasing right there...

So what's your plan for getting the next generation of true experts?

"By the end of this week, you and Sam will be the most hated men in America. If you insist, so it will be."

Oh Elon, don't fret. There's plenty of hate left over.

As long as it requires not one, but two unobtainable 5090 cards, I think you'll find approximately 100% of gamers will be left with a game that looks like crap.

Can't have "Cloud" without "C" *taps head*

Presumably, just not brought it up in court.

I got my new gaming PC about 6 weeks ago. It has only Linux on it. Right now I'm playing through Kingdom Come Deliverance II (it was on 50% off last week) but I've also been playing Elite Dangerous, Sniper Elite 5 and Sniper Elite Resistance, Terra Invictus, WARNO and The Witcher 3. To be fair I only played the Witcher for a few minutes as I want to finish KCD before starting a proper re-play of the Witcher.

But the point is that it was no more difficult to install most of those than on PC. The exception being The Witcher as GoG is not currently Linux native, but Heroic allows you to use your GoG account and then points your games towards Proton so the only extra step for The Witcher was to install Heroic rather than GoG.

Whoosh

Better still, there's an entire *novel* about it: Ender's Game. I don't remember the SG1 episode you refer to but I would bet the book is better.

"More fundamentally, AI models may not understand 'stakes' as humans perceive them."

Of course they don't. They don't *understand anything*. They just predict which word(s) is statistically most likely to come after all these other word(s).

But Microsoft will happily rent you a CoPilot service that does. All hail the cloud slop!

Oh come on, it's only been about 2 weeks ... and 4 years.

But Meta owns both the ends. They donâ(TM)t need to break the encryption to spy on you.

I don't think that analogy is directly comparable. There were very clear benefits to the truck compared to the horse. Though you are right about maintenance, there was also a clear path toward fixing them - more mechanics, more petrol/diesel fueling stations, this is just a matter of scale. It was also clear that basic reliability would improve with further engineering.

There is no such clear path forward with LLMs/Generative AI. People can lie to you, but when they do, they know they're telling you a lie. LLMs don't "lie" to you. They literally have no concept of truth. They don't have a concept of *anything*. Statistically, this word is likely to come after that one, and that's what the LLM spits out to you. You can't tell an LLM to stop lying when it does know what a lie *is*.

LLMs are a dead end. They do have some legit uses -- blown way out of all proportion in this ridiculous bubble -- but simply chucking more words at ChatGPT isn't going to suddenly make it become sentient.

If (not when) we finally create an AGI, it won't be a giant version of any of the current models. It will be something else entirely. It might have have an LLM as some sort of subfunction, but it won't be simply an LLM x 1000.

No, the humans who OWN the AIs will never have to sell their waking lives just to exist. The rest of us will continue to serve at their pleasure.

What could possibly make you believe that AI will break the cycle of rich people getting exponentially richer at the expense of everyone else? We already produce more than enough food for everyone on earth to eat well, yet millions starve because giving food to the poor doesn't make any money.

There will always be scarcity because the rich people *make it scarce*.