ASI Readiness (@ASIreadiness) header
ASI Readiness (@ASIreadiness) profile picture

Not coming soon…

San Francisco, CA
Joined May 2025
Photo by ASI Readiness (@ASIreadiness)Photo by ASI Readiness (@ASIreadiness)Photo by ASI Readiness (@ASIreadiness)Photo by ASI Readiness (@ASIreadiness)
“use my custom meta prompt to get chatgpt to talk straight to you” yeah cool, do you think regular people are going to do this? Obviously the default setting is the only thing that matters for the impact on society at large
we’ve invented a robot that tells people why they’re right no matter what they say, furnishes sophisticated arguments for their side, and delivers personalized validation from a seemingly “objective” source. Mythological-level temptation few will recognize for what it is
2
2
19
3,306
The switch to @ASIreadiness will happen faster than we think…
1
1
3
136
A new RAND report on AI x-risk is shockingly bad; I don't see how it got past their internal peer review. There are many issues, but the main critical flaw is the conflation of "It seems hard to me" with "It will be hard for a superintelligent AI". Other issuse: - Not grappling with superintelligence, except to fallaciously argue that It would be software based, and thus we can focus on specific mechanisms by which of might cause harm. - Ignoring systemic risk by assuming that actors would never deploy something that they didn't trust - overstated/unsupported conclusion that x-risk must happen slowly and provide time for humans to respond - arguing that because we expect benefits from AI, we shouldn't stop On the basis of their largely irrelevant analysis, they end up advocating a “wait and see” approach to dealing with x-risk from new technology, explicitly arguing that shutting it down is bad.
15
15
2
141
9,211
ASI Readiness retweeted
What are some things AI can do (now in 2025) that it couldn’t do in 2024? We sometimes forget for fast AI is evolving.
17
41
3
300
29,285
Replying to @ASIreadiness
@ASIreadiness we should talk…
1
4
1,197
Imagine watching this and STILL saying "it's not reasoning" I will say it again: Reasoning Deniers are the Flat Earthers of AI
The way that o3 can guess locations feels very science fictional, down to the “zoom, enhance” tools it applies to an image. This is sped up, but it guessed the location from a photo I just took with geodata removed and no clear signage drawing on the vegetation & highway columns
24
20
2
250
16,754