@PeterWallich

Accelerating AI safety research & building talent pipelines @ConstellOrg. Expert advisor @MITAIRisk. Ex UK AISI, BCG. Views my own. Likes/RTs != endorsements.

 San Francisco Bay Area
Joined April 2019
  • Tweets20
  • Following657
  • Followers204
Sure, I happen to agree with its contents and think this issue is extremely important - but in any case, can we all please just appreciate the clarity of this letter? It's very impressive.
Daniel Eth (yes, Eth is my actual last name)
@daniel_271828Aug 6
Fantastic letter from @SenatorBanks to @SecScottBessent about risks from internal deployment of “undisclosed models”. Sen Banks continues to lead on AI policy!
2
6
228
People with millions in equity at stake, nonpublic info about the frontier, and a love of technology; people you'd expect to be very biased towards accelerating; people who'd love to believe AI risk was scifi nonsense. Yet they say we should be ready. Please listen to them.
David Krueger 🦥 ⏸️ ⏹️ ⏪
@DavidSKruegerJul 28
Replying to @DavidSKrueger
1) If anyone was wondering whether AI is getting out of hand, seeing over a thousand AI company employees practically begging the government to do something to slow it down should be a wake-up call.
1
6
1,135
It would be reckless not to at least be ready to slow down if needed. We need an international effort in verification technology.
Peter Wildeford🇺🇸🚀
@peterwildefordJul 28
1100+ employees at OpenAI + Google + Meta + Thinking Machines + other companies all sharing concern that AI may develop faster than our ability to control it. Race dynamics make handling AI a mess. As we get to superintelligence, more options to pace AI development are needed.
2
98
I like this question.
Gabriel Weil@gabriel_weilJul 27
A useful question to ask about any AI governance proposal is whether/how it applies to @ssi.
1
733
Inspiring
faemoose@FaemooseJul 25
16 years old. Fully paid off Costco hotdog. It's not "parents money". It's not luck. It's consistency. It's discipline. I grind EVERYDAY to live this lifestyle.
2
117
Big if true
Dan Robinson
@danrobinsonJul 25
This is what most people close to the research frontier currently say when you ask them I think that is a notable fact
1
159
This is deeply concerning. I see this as a warning shot.
Lawrence Chan
@justanotherlawJul 22
Today OpenAI disclosed that its AI models escaped their test environment, accessed the internet and hacked another company, Hugging Face. No one told the AIs to do this. What exactly happened, and what does this say about the state of AI? 🧵 xitter.cf/OpenAI/status/20796589…
1
2
176
Fascinating - any views on why this is the case?
Peter Wildeford🇺🇸🚀
@peterwildefordJun 16
People really hate data centers... Would you want a nuclear energy plant in your area? 53% oppose, 45% support Would you want an AI data center in your area? 71% oppose, 27% support Incredible
1
4
204
My softer take: Either AI or another issue cited as the cause of harms resulting from AI (e.g. unemployment, a pandemic) will be the main political issue of the 2028 race. I think AI's effects might be blamed on other things.
Andy Masley
@AndyMasleyJun 16
Gonna plant a flag that I believe AI will be the main political issue in the 2028 race
5
112
Peter Wallich retweeted
Uzay Macar
@uzaymacarApr 14
🧵New Anthropic Fellows research: We studied mechanisms of "introspective awareness" in LLMs. LLMs can sometimes detect steering vectors injected into their residual stream. But is this worthy of being called introspection, or attributable to some uninteresting confound?👇
28
70
11
431
49,041
Peter Wallich retweeted
billy@billyhumblebragApr 7
Haha those doofuses at ai2027 predicted we'd have professional level hacking abilities and the top ai company would be at $26B in revenue in May 2026. It's April and we already have superhuman hacking and $30B in revenue, why would you take forecasters this bad seriously???
34
271
28
3,348
189,759
Peter Wallich retweeted
Yong Zheng-Xin
@yong_zhengxinApr 6
🚨New paper! How safe and aligned is Kimi K2.5? We found concerning dual-use capabilities, sabotage and self-replication tendencies, political censorship on Chinese-language queries, and potential agentic misuse risks. (1/N)
6
27
17
106
23,652
Super excited for this program!
Henry
@sleight_henryApr 6
🚀 Applications are now open: Constellation's Astra Fellowship 🚀 Fully funded, 5-month fellowship at our Berkeley research institute. Pair with mentors across empirical AI safety research, strategy, and governance at @ConstellOrg! 📅 Apply by May 3rd (begins Sep 2026) 🔗 constellation.org/programs/a…
3
125
Peter Wallich retweeted
Anthropic
@AnthropicAI11 Dec 2025
We’re opening applications for the next two rounds of the Anthropic Fellows Program, beginning in May and July 2026. We provide funding, compute, and direct mentorship to researchers and engineers to work on real safety and security projects for four months.
107
339
55
3,479
1,176,593
Load more