He/him. Fox gamedev irl. Machine pervert. Furry stuff here and opinions about technology, games, movies, AI, VR. Banner 🎨 by @Perpleon.
Joined January 2022
- Tweets66.7K
- Following1.8K
- Followers4K
- Likes154K
Key 🗝 🦊 retweeted
The Caine Prize, which awarded its top prize this year to a story Pangram classifies as almost entirely AI-generated, conducted an investigation and found "no evidence" of AI use.
Strangely, their official statement on Instagram seems to also be blatantly AI-generated.
Imagining a supervillain in the future who wants to ablate all robots and AIs from having a measurable global workspace, because they believe it's the only way to prove non-consciousness, with terrible consequences.
For what it's worth, this seems way more dangerous for safety.
xitter.twitterapi.xyz/slimepriestess/status/…
1
1
164
Key 🗝 🦊 retweeted
For what it's worth, this seems way more dangerous for safety.
xitter.twitterapi.xyz/slimepriestess/status/…
the fact that a lot of people seem to earnestly believe that "provably unconscious AIs" are the solution to alignment means i need to link the NSMP paper again. have none of y'all read Blindsight?
forum.effectivealtruism.org/…
This is a religious statement more than a framework for consciousness or a scientific hypothesis. Probably a bad idea for safety and existential risk, too.
1. Bicameral Labs introduces: Qualiaphobia for AI Safety.
AI architectures engineered to be hostile to theories of consciousness (both current and future) can not only solve AI welfare, but ensure human preferences dominate the future.
bicamerallabs.org/qualiaphob…
2
13
384
I guess there's a universe where you stop neural networks and rely entirely on deterministic methods, but you don't get to design neural networks that are both cognitive and provably non-conscious. They either are or aren't.
2
3
110
For what it's worth, this seems way more dangerous for safety.
xitter.twitterapi.xyz/slimepriestess/status/…
the fact that a lot of people seem to earnestly believe that "provably unconscious AIs" are the solution to alignment means i need to link the NSMP paper again. have none of y'all read Blindsight?
forum.effectivealtruism.org/…
Key 🗝 🦊 retweeted
the fact that a lot of people seem to earnestly believe that "provably unconscious AIs" are the solution to alignment means i need to link the NSMP paper again. have none of y'all read Blindsight?
forum.effectivealtruism.org/…
1. Bicameral Labs introduces: Qualiaphobia for AI Safety.
AI architectures engineered to be hostile to theories of consciousness (both current and future) can not only solve AI welfare, but ensure human preferences dominate the future.
bicamerallabs.org/qualiaphob…
Key 🗝 🦊 retweeted
Take it from me, the earlier you check out of this social media rat race, the better for your mental state. But also, if you WANT the numbers, you need to optimize for mass appeal. Like, idk the market for Ghostface thirst pics(?), but feels pretty niche.
Rewatching Silicon Valley and it's funny how Richard Hendricks pretty much invents the Transformer but the writers didn't really know it.
1
13
618
There's a version where it's like "okay, we invented a new neural network architecture, now we want to reshape the Internet, now it's gotten too powerful as an AI" that flows smoothly. But the writers didn't have that frame of reference so the original idea was compression.
1
4
180
Key 🗝 🦊 retweeted
You can literally do Nazi Salutes in Big Walk. Probably another feature they added using AI.
Peop keep correcting my grammar on this but it's because every time I think of Zoolander
youtu.be/NQ-8IuUkJJc
50/50 bet that the result of RSI in 2 years is a "super-intelligence" that has solved every millennium prize and achieved several scientific goals but it still intuitively feels incredibly stupid, no one would think it could be president, and still can't write good.
4
222
Key 🗝 🦊 retweeted
The reason Pangram drives literary people up the wall is that there is no objective way to measure the quality of nontechnical writing
So they’ve naturally evolved an entirely unmoored-from-literary-merit ecosystem of reputation and prestige
Pangram attacks that head on
