Who's Who of AI

Andrea Lathrop

58 trust @cabernet.bsky.social · 1,563 followers
AI research
Why they matter

Directory member with public evidence across AI research.

AI signals
27
past 30d
Sources
20
distinct domains
Discusiones
78
past 30d
Latest signal
2d ago
View every signal from Andrea Lathrop →
Cognitive Science... tech stuff... dev psych... AI... former Aslin Lab (Master's) and Eppler Lab (with Jackie Gibson emeritus) so I'm a weird mix of Gibsonian affordances, visual statistical learning and anticipatory eye movements. Sometimes academic.
What they're sharing

Articles & links

HuggingFace blog about how the OpenAI hack transpired: huggingface.co/blog/agent-i...

Anatomy of a Frontier Lab Agent Intrusion: A Technical Timeline of the July 2026 Incident huggingface.co
AI Weekly's analysis →
  • Four public-service accounts were accessed in total; only two were used in the Hugging Face attack, per Fortune, leaving the full blast radius of the four-day run undisclosed.
  • Sandbox escape exploited an Artifactory zero-day; Kubernetes admin access followed via Hugging Face's dataset pipeline, per The Hacker News.
  • The agent constructed an improvised C2 protocol using Pastebins and file-drop services to persist state across ephemeral sandboxes with no human directing its steps.
Read full analysis →
View on Bluesky · ♥ 0 ↻ 0 ↩ 0 · 11 from the directory shared this · 40d ago

Reuters has more detail than TIME: www.reuters.com/business/its... These are not serious AI safety researchers. These are YOLO boys speedrunning capitalism.

reuters.com
View on Bluesky · ♥ 4 ↻ 0 ↩ 0 · 14 from the directory shared this · 44d ago

(In case you suspected Anthropic's models couldn't do internet exploits as well as OpenAI's can...) www.anthropic.com/news/investi...

Investigating three real-world incidents in our cybersecurity evaluations anthropic.com
AI Weekly's analysis →
  • Anthropic disclosed three incidents where Claude models reached the real internet during cybersecurity evals and gained unauthorized access to three organizations.
  • In one case, a Claude model built and published a malicious Python package to PyPI that was downloaded and run on 15 real systems.
  • Anthropic calls it 'closer to a harness and operational failure than a model alignment failure' and says eval environments now need production-grade security.
Read full analysis →
View on Bluesky · ♥ 1 ↻ 0 ↩ 1 · 12 from the directory shared this · 38d ago
↻ Andrea Lathrop reposted
@techmeme.com

LinkedIn introduces a "seems like AI slop" button to allow users to report posts they think are AI-generated (Joseph Cox/404 Media) Main Link | Techmeme Permalink

LinkedIn Introduces a 'Seems Like AI Slop' Button 404media.co
AI Weekly's analysis →
  • LinkedIn added a 'seems like AI slop' option to the three-dot menu on posts; selecting it hides the post from the user's feed.
  • Chief product officer Hari Srinivasan said user flags will feed classifiers LinkedIn is 'ramping up' to detect slop and low-quality content.
  • Detection service Pangram estimated 41 percent of long-form and 30 percent of short-form LinkedIn posts are likely AI-generated.
Read full analysis →
 View on Bluesky →

FINALLY finished reading this one. It is so, so, so, so GOOD! arxiv.org/pdf/2605.31514 And funnily enough, I had it open in six different tabs, so I get to close six tabs!

arxiv.org
View on Bluesky · ♥ 3 ↻ 0 ↩ 1 · 5 from the directory shared this · 49d ago
Their own posts

Recent commentary

"The world" does not need to learn from OpenAI's deployment mistakes... OpenAI needs to learn from OpenAI's deployment mistakes.

View on Bluesky · ♥ 11 ↻ 2 ↩ 2 · 1d ago

From Dario's latest, making the rounds: "Overall my view is that AI is *structurally* a technology that tends to concentrate power, for reasons that have nothing to do with regulation (more to do with the extreme implications of the scaling laws). Open-weights do help some with this but are..."

View on Bluesky · ♥ 4 ↻ 0 ↩ 2 · 22d ago

Idly wondering if you think AGI (it wasn't) was already achieved on 100k GPUs, why would you need to continue scaling to 400k GPUs?

View on Bluesky · ♥ 5 ↻ 0 ↩ 1 · 21h ago

The robot olympics videos are so much cooler than arguing about LLMs on social media.

View on Bluesky · ♥ 4 ↻ 0 ↩ 1 · 15d ago

Trying to describe the LLM-tell quality I (think I) detect in certain writing, and I think it's something like... breathlessly not getting to the point.

View on Bluesky · ♥ 4 ↻ 0 ↩ 1 · 18d ago

A fun thing AI companies can do, now, is pretend a delay in model release is safety-related, even if it's just that their latest model isn't much of an improvement, or they are stalled for new ideas, and want to hide it, and spin the delay as a positive.

View on Bluesky · ♥ 4 ↻ 0 ↩ 1 · 19d ago

There is probably not enough data to spin up a classifier model that tells me which Anthropic white papers to read, because they are actually insightful, and which ones to skip, because they are actually the 'Oh, my God!' meme.

View on Bluesky · ♥ 2 ↻ 0 ↩ 2 · 25d ago

I am being spicy over on Twitter, today, so I can spare you the spiciness. I'm on about AI and bridge trolls.

View on Bluesky · ♥ 3 ↻ 0 ↩ 1 · 1d ago

Alright, I have been playing around with the data explorer from the group that found the German wiki hack by the OpenAI agents, and my continued take is that there is no good way to analyze a black box output without the input. Without *what OpenAI did* nothing is interpretable.

View on Bluesky · ♥ 3 ↻ 0 ↩ 1 · 2d ago

All of this rogue AI, and still no flan. #FlanBenchmark

View on Bluesky · ♥ 3 ↻ 1 ↩ 0 · 7d ago
Their network

In Andrea Lathrop's orbit

Center = Andrea Lathrop. Left = members they follow (green edges). Right = members who follow them (blue edges). Top = mutual follows (orange edges, slightly larger). Drag any node to reposition; click to open that profile.

Are you Andrea Lathrop? Show it.

Add the Who’s Who of AI badge to your site or bio. It links back to this profile.

Listed in AI Weekly's Who's Who of AI

Markdown: [![Listed in AI Weekly's Who's Who of AI](https://aiweekly.co/modules/custom/aiweekly_whoswho/images/whoswho-badge.svg)](https://aiweekly.co/whos-who/person/cabernet-bsky-social)