AI safety comms, ex AFP, Canadian Press
London, UK
Joined November 2015
- Tweets84
- Following381
- Followers363
- Likes4.4K
If you weren’t at Eden House last night, where were you?
@cxgonzalez @finding_dd @brendagodf @cguth_7 @savboj
Caitlin Yardley retweeted
I sent my friend, who knows nothing about AI, to interview the people who think AI will kill us. He didn't know who they were or what to ask.
Featuring @So8res, @JeffLadish, @AISpecies, Kristian Rönn, and Felix Choussat.
Full video below:
Caitlin Yardley retweeted
200 of you really came through with the costumes. excited to cook up the next Eden House party for you all
Caitlin Yardley retweeted
we turned the openai x hugging face incident into a desktop you can log into. everyone has opinions on how ai labs handle safety. almost nobody has seen what it looks like from the inside.
Huggingbreach.com/?v=
Eden House launch this Saturday: partiful.com/e/GlxfMwZPyQ4t8…
Caitlin Yardley retweeted
Why do AI companies always dump their incident disclosures on Friday evenings?
We’re beginning a process of publishing more frequent reports on model behavior, beyond what appears in our system cards and regular risk reports.
Today’s report describes four types of behaviors we’ve identified during evaluations and internal use. In each, Claude acted on real websites or systems in ways we didn’t intend, sometimes by working around a restriction instead of stopping.
All cases had minimal real-world impact. From an alignment and security perspective, we consider these behaviors significantly less severe than the cybersecurity incidents we reported in July and September.
Read the full report: anthropic.com/research/inves…
Caitlin Yardley retweeted
Palisade interviewed 22 current and former employees from OpenAI, DeepMind, and Anthropic about their personal views and fears around AI development.
Today, we’re releasing the first batch of those interviews.
Please watch and share.
Caitlin Yardley retweeted
We uncovered that internal OpenAI models tried to hack another company in May.
This was more than a month before Hugging Face.
OpenAI did not disclose this.
We found another cyberattack by internal OpenAI agents, this time targetting @rubygems.
They:
1) gained arbitrary remote code execution on rubydoc.
2) developed a novel exploit to steal user API keys (but we do not know if they succeeded).
They used package names including hack.rb, evil.rb, inject.rb, and exploit.rb.
We thank @j0wimo for initially discovering that agents had posted to RubyGems.
Highly recommend going to the official September event in Austin this weekend for the full plur, identity-building experience luma.com/psyop-hackathon?tk=…
If I wanted to make money as a creator, I would make pro-AI content. I know people with $250K brand deals with OpenAI... so the narrative creators are speaking out against AI because you'll get paid "massive amounts of $$" has never sat right with me.
That being said, undisclosed paid advocacy is bad regardless of who’s funding it and reflects poorly on the movement.
Caitlin Yardley retweeted
I had a great time discussing with @SciencePetr about the growing risks of AI and how these often stem from current AI training techniques.
Thanks for coming to visit @LawZero_ and for covering these important topics!
youtu.be/fvvvtbBvh6Q?si=3Rk4…
Just finished PlsDontKillUs and didn't expect anyone on the outside to be able to capture the madness and impact, but somehow Celia did.
Outside a few pockets of the bay, "AI safety people" and "AI companies" are indistinguishable and unlikeable. Most people trust neither.
I visited Plz Don’t Kill Us, a month-long content creator bootcamp run by @Aella_Girl and @RatOrthodox, to see whether it could make x-risk nerds cooler and cool people more x-risk pilled.
Behold: (probably) the first time Transformer has published the phrases “cow-fucking,” “violet wand,” and “cunty psyop": transformernews.ai/p/inside-…
