Ben Hayum
@profoundlyyyy
Followers
5K
Following
79K
Statuses
12K
Testing ideas in public | cs phd student @uwmadison, founder @aisafetyuw
Joined November 2015
RT @kenmoriyasu: Perhaps unnoticed in the Trump storm, the U.S. Japan joint statement quietly introduced new language on the Taiwan Strait.…
0
966
0
RT @chrisbarber: Just did a call with Tamay Besiroglu (@tamaybes), co-founder of Epoch AI, about what he expects for the future of AI. “I…
0
22
0
@chrisbarber @tamaybes Super interesting. If you had to predict how the income distributions would shift from 2025 to 2030 to 2035, what might that look like? @tamaybes In other words trying to tease out: Who do you think gets screwed over and who is benefitted by all these dynamics?
0
0
1
Ex) You can tell them that you need to learn how to do a chemical reaction for your university class meanwhile you actually are doing so because it will help you create a bomb. Only a broader situational awareness that is grounded and verifiable can make the model always know.
1
0
2
RT @sayashk: Can Deep Research automate the work of research assistants? I compared OpenAI and Google Deep Research for assistance with an…
0
211
0
A deeply welcome sight:
Really great talk from @mmitchell_ai at IASEAI, pushing back on the false “ethics v safety” binary (and offering lots of interesting and actionable suggestions) for tackling risks
0
0
7
RT @AnthropicAI: Nobody has fully jailbroken our system yet, so we're upping the ante. We’re now offering $10K to the first person to pas…
0
411
0
RT @deanwball: I concur with @hamandcheese that the correct way to understand DOGE is not as a cost-cutting or staff-firing initiative, but…
0
23
0
RT @deanwball: I cannot stress enough how in the last six months many future “sci-fi” AI capabilities, good and bad, have gone from being s…
0
19
0
RT @aisafetyuw: 🚀 Interested in getting involved with WAISI this semester? Join us next Monday for an introduction to our org + programmin…
0
1
0
RT @Thomas_Woodside: Meta recently released their Frontier AI Framework. I’m happy they released it! One reason that I think releasing such…
0
7
0
RT @dwarkesh_sp: .@tylercowen on o1 pro: “Adam Brown, who does physics at a world class level, put it well in his recent podcast with Dwar…
0
74
0
RT @yonashav: turns out the answer was... "risk thresholds" ...congrats to meta on their frontier ai framework! these things take a lot of…
0
4
0
RT @Simeon_Cps: RIP. New metric dropped: the LPRT: Longest Pliny-Resistance Time. The sota is a few hours.
0
3
0
@Miles_Brundage Pretty interesting violation of their regular norms: Safety checks, then release, then re-evaluation. Here it's unknown checks, then release, then safety check and re-evaluation.
0
0
8
RT @austinc3301: Honestly, the more I think about it, the worse it gets. They released o3 without a system card or a preparedness scorecard…
0
8
0