Tuesday, March 12
Daily News Stuff 12 March 2024
Chipi Chipi Edition
Chipi Chipi Edition
Top Story
- Elon Musk has announced that his AI company - xAI - will be releasing its AI - Grok - as open source. (Notebook Check)
Exactly how open that is remains to be seen, but there is nothing at all open about OpenAI (creator of ChatGPT) so the bar is pretty low.
- Meanwhile Midjourney has banned all Stability AI staff from its service over alleged attempts at data theft. (The Verge)
Stability AI is the home of the open-source Stable Diffusion image generation model. Its CEO has stated "It wasn't us and if it was it was an accident."
Tech News
- The US must move "decisively" to avert an "extinction-level" "threat" from "AI", according to a government funded "report". (Time)
"Current frontier AI development poses urgent and growing risks to national security," the report, which TIME obtained ahead of its publication, says. "The rise of advanced AI and AGI [artificial general intelligence] has the potential to destabilize global security in ways reminiscent of the introduction of nuclear weapons." AGI is a hypothetical technology that could perform most tasks at or above the level of a human. Such systems do not currently exist, but the leading AI labs are working toward them and many expect AGI to arrive within the next five years or less.
The only problem with this statement is that none of it is remotely real. None of the leading AI labs are working toward AGI. They are working on redefining AGI so that they can claim to have achieved it.
- Lenovo may be releasing its Legion Y700 tablet to the global market next month. (Notebook Check)
In fact, something I didn't know is that it's already sold in Japan by NEC as the Lavie Tab 9.
Anyway: I'll buy two. There is literally no competition right now.
- We haven't checked in on QNAP in a while and oh they're on fire. (Bleeping Computer)
Still.
- How do AMD's integrated graphics stack up against the best graphics cards made with Chinese chips? If you guessed AMD's built-in graphics are twice as fast, well, yes. (Tom's Hardware)
Some of that will be software limitations, that the Chinese makers will resolve over time. But right now a high-end China-only card gets stomped by even the cheapest previous-generation card from AMD.
- Europe has found Europe guilty of breaching Europe's data protection laws. (Tech Crunch)
Well, there's a thing.
Disclaimer: Place water into a Glowstone frame for a Hostile Paradise.
Posted by: Pixy Misa at
05:48 PM
| Comments (2)
| Add Comment
| Trackbacks (Suck)
Post contains 412 words, total size 4 kb.
1
My view on these claims of AI and 'extinction level' is that I believe that very few of the people making these claims have made a serious effort to study the challenges involved in totally exterminating humanity.
I have actually studied this last problem a bit.
Among other things, it is related to the problem of totally exterminating any one national population.
Which is reasonable to study given such things as the AGW hypothesis, or Russia's theory of denazification in it's "sphere of influence". The latter implies that Russia and Mexico are in America's sphere of influence, and that the American government can legitimately pursue denazification.
Anyway, the AI experts talking about extinction either started out insane, or became insane by hanging around with insane people, or maybe even were never truly expert in AI.
(Among other tests... Humans do not need electrical power to survive as a species. AI necessarily needs electrical power. Humans are necessary for electrical power supply. Beyond that there is no evidence of breakaway synergies, the problem of automating infantry is hard, and there is no reason to expect AI to qucikly solve all the problems needed to solve to exterminate humanity and also to continued to function while and after doing so. If the AI fails early for lack of solving those, then humanity survives by default. The challenges run right through the heart of engineering. So, I would either want to see successful tests of automation replacing most engineers, or I would want to see successful tests of killbots, or I must conclude that the folks calling the alarm are mystics or idiots. )
That is them being crazy even by my standards, and I have been pretty nuts for a long time.
The alternative explanation is that political idiots are desperate, and all of the AI noise was meant to pay off on them playing the 'extinction level threat card' and getting everyone to back down, and the AI experts are simply lying.
I have actually studied this last problem a bit.
Among other things, it is related to the problem of totally exterminating any one national population.
Which is reasonable to study given such things as the AGW hypothesis, or Russia's theory of denazification in it's "sphere of influence". The latter implies that Russia and Mexico are in America's sphere of influence, and that the American government can legitimately pursue denazification.
Anyway, the AI experts talking about extinction either started out insane, or became insane by hanging around with insane people, or maybe even were never truly expert in AI.
(Among other tests... Humans do not need electrical power to survive as a species. AI necessarily needs electrical power. Humans are necessary for electrical power supply. Beyond that there is no evidence of breakaway synergies, the problem of automating infantry is hard, and there is no reason to expect AI to qucikly solve all the problems needed to solve to exterminate humanity and also to continued to function while and after doing so. If the AI fails early for lack of solving those, then humanity survives by default. The challenges run right through the heart of engineering. So, I would either want to see successful tests of automation replacing most engineers, or I would want to see successful tests of killbots, or I must conclude that the folks calling the alarm are mystics or idiots. )
That is them being crazy even by my standards, and I have been pretty nuts for a long time.
The alternative explanation is that political idiots are desperate, and all of the AI noise was meant to pay off on them playing the 'extinction level threat card' and getting everyone to back down, and the AI experts are simply lying.
Posted by: PatBuckman at Tuesday, March 12 2024 11:19 PM (r9O5h)
2
AI expert: someone who plays candy crutch all day.
Extinction level event: being without your smartphone for more than five minutes.
Extinction level event: being without your smartphone for more than five minutes.
Posted by: normal at Wednesday, March 13 2024 12:17 AM (zx18t)
53kb generated in CPU 0.0128, elapsed 0.2691 seconds.
58 queries taking 0.2599 seconds, 341 records returned.
Powered by Minx 1.1.6c-pink.
58 queries taking 0.2599 seconds, 341 records returned.
Powered by Minx 1.1.6c-pink.
