AI notkilleveryoneist • Medical Doctor • Land Of Lisp Author

Mill Valley, CA
The problem is that the immediate things you notice about a game (graphics, animation, surface mechanics) are now extremely cheap to create- but the deep, rewarding mechanics are still as difficult as always This makes it much harder now for a quality game to find an audience
making games easier to make does not make it easier to make good games
1
10
720
Here is the latest chart today, the problem is the benchmarking sites are broken right now for the frontier models (artificial analysis is clearly undervaluing openai models, and epoch ai hasn't rated the latest models yet) With better data, I think openai could still compete
when are we getting more of these?
1
7
319
My pet theory as to why Jensen Huang is so poorly informed about anything that veers even a little beyond his core business is that he is just a crazy 99.999 percentile workaholic He literally does absolutely nothing except solve business problems and spend time with his family
1
9
280
the only way you could ever get him to have any insight about AI xrisk is if you gave him a problem for his business that could only be solved by having such an insight
1
90
I didn't realize until recently why pigs were such a big deal in human history: It's that they're the only farm omnivore, and before refrigeration everybody had lots of revolting food garbage, but a pig will eat anything turning your nasty garbage into yummy cutlets
55
130
6,978
94,900
It would be interesting to try to use AI to do experimental research into counterfactual history, for instance testing the "great man theory" What happens if you change some of the leaders involved in WW2- Does the same stuff happen, or something wildly different?
1
2
339
To be clear: I agree you would never be able to simulate WW2 exactly, but you might be able to simulate a vaguely similar war, but then you could make small counterfactual changes and see how it affects the war
1
1
227
and of course frontier models have lots of safeguards against military-style prompts, which you'd have to contend with
1
155
I say we rename it API (Asshole President Intelligence)
3
266
Pro tip: When you take a drink from the starbucks mobile order station, it's a good idea to pretend to read the label, as if it's a drink that you ordered and you're checking the name
1
5
486
I'm surprised no recent billionaire has ever simply built a giant pyramid in the desert, with the hope people will still remember them 5000 years from now
5
2
31
1,338
I hope any OpenAI employees currently in Australia get arrested
Australia has been hacked. 'And today, I spoke with the CEO of OpenAI, Sam Altman, to express Australia's extreme concern about this incident. And I also expressed my disappointment that it took the company way too long to inform the government what had occurred, and the nature of the way that that notification occurred as well was unacceptable.'
10
779
Continuous money (versus discrete money) would be fun: Imagine if your job paid $100/day, but your cable bill is $4/day, so in the money system you could just subscribe to cable, and your residual job stream then just changes to $96/day, without ever a discrete payment event
12
72
5,524
Obviously if you go to the store and buy a bagel for $5 dollars, the system would then still need to fall back to using a simple discrete transaction. I did a search, and several people in crypto have already created some implementations of such continuous money
1
12
825
How does everyone who was a coauthor on that cursed "stochastic parrot" paper have such weirdly convoluted writing
Replying to @mmitchell_ai
1. Assuming that LLMs should be the basis of AI/AGI is worth questioning. They are indeed stochastic, which can mean they are harder to predict and to control. From a risk/harm standpoint--including x-risk--this is critical to wrap your head around. Questioning LLMs as the basis of modern "AI" has also been articulated by people who have worked in AI much longer than many at Anthropic, including @drfeifei, @ylecun, and ofc consistently @GaryMarcus.
1
4
684
I'm heartened to see that "superior intelligence" is winning the vote right now If I wanted to come up with a rebranding of "AI" that maximally emphasizes the dangers of the tech, I don't think I could come up with a better term
4
21
866
I saw "Fear and Loathing in Las Vegas" in the theaters, and during the drug trip scene the projectionist somehow had loaded that reel of film to play upside down and backwards It took ~10 minutes until all of us people in the theater finally figured it must be a mistake
A close friend once told me he listened to the audio book of Blood Meridian and thought it was extremely surreal and avant guard, only to learn his audiobook player had a "random" feature for chapters and that it had switched on.
1
6
654
Tech progress for normal people was absolutely bonkers until everyone had a color TV and a wall phone- which happend around the 70s when I was born. Then, tech mostly stalled (the computer was the only blip) But from now on, tech acceleration is going to be inconceivably fast
21
1,043
Conrad Barski retweeted
Replying to @AndrewCurran_
3
12
182
3,906
Conrad Barski retweeted
Mathematician Terence Tao: "we have to slow down AI. the pace is insane, and there's no reason to be this fast — no reason at all" It's amazing how willing we are to change everything without any idea what happens afterward These are extremely nonlinear dynamics
885
1,031
7,460
7,510,596