Andrea Lathrop reposted
Fable is being taken down? IMMEDIATELY?? www.anthropic.com/news/fable-m...
Who's Who of AI
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
Statement on the US government directive to suspend access to Fable 5 and Mythos 5 anthropic.com
View on Bluesky →
It Is Trivially Easy to Use Reddit to Manipulate AI Search, Research Suggests 404media.co
View on Bluesky →
Opinion | Dear A.I. Companies: The Doom Trolling Needs to Stop nytimes.com
Opinion | Bernie Sanders: A.I. Is a Public Resource. You Should Own Half of It. nytimes.com
Introducing Claude Tag anthropic.com
No, Artificial Intelligence Is Not Conscious theatlantic.com
Project Glasswing: what Mythos showed us blog.cloudflare.com
[2605.22763v1] Advancing Mathematics Research with AI-Driven Formal Proof Search arxiv.org
[2606.19317] Explaining Attention with Program Synthesis arxiv.org
China may have accessed Mythos theverge.com
No, Artificial Intelligence Is Not Conscious theatlantic.com
If LLMs Have Human-Like Attributes, Then So Does Age of Empires II arxiv.org
Articles & links
Andrea Lathrop reposted
New: A tiny snippet of user-generated text as short as 13 words is often enough to manipulate the AI agents that power tools like ChatGPT and Google’s AI search. Shows trivially easy for companies/brands to manipulate ChatGPT results by posting on Reddit, Quora, etc www.404med…
"Let’s call this strategy 'doom trolling.' It’s one of the defining and most arresting properties of our current A.I. moment, and I’ve come to believe that it’s morally indefensible." www.nytimes.com/2026/06/17/o...
Are you serious? www.nytimes.com/2026/06/01/o...
LINKS: Anthropic blog post on Claude Tag: www.anthropic.com/news/introdu... Narayanan's thoughtful response on Twitter: x.com/random_walke...
Forwarded this earlier, and have now read it, in its entirety. This is EXCELLENT from Ted Chiang: www.theatlantic.com/philosophy/2...
Finally got around to reading this Cloudflare blog post about how they harnessed Mythos for their Project Glasswing preview: blog.cloudflare.com/cyber-fronti...
Haven't read it, yet, but apparently the Deep Mind gang solved 9 more Erdös problems... arxiv.org/abs/2605.227...
I have not read it, yet, but this looks like it might be interesting. (One of the authors posted it to Twitter.) arxiv.org/abs/2606.19317 They replaced attention heads with human-readable Python code snippets(?)
AI Weekly's analysis
→
- Under 1,000 LM-generated programs reproduce attention patterns in GPT-2, TinyLlama-1.1B and Llama-3B with above 75% average IoU on TinyStories.
- Replacing 25% of attention heads with the synthesized programs raised average perplexity by only 16% while keeping downstream question-answering performance.
- The pipeline computes a head's attention matrices, prompts a pretrained LM to write Python that reproduces them, then re-ranks by held-out accuracy.
Read full analysis →
"... may have..." www.theverge.com/ai-artificia...
Forwarded this earlier, and have now read it, in its entirety. This is EXCELLENT from Ted Chiang: www.theatlantic.com/philosophy/2...
Andrea Lathrop reposted
Everyone should read this, just to see how they implemented a perceptron (a very basic neural network) in Age of Empires II arxiv.org/abs/2605.31514 The artificial neurons pass activation energy in the form of goats.
AI Weekly's analysis
→
- Adrian de Wynter argues LLM anthropomorphic attributes are empirically non-unique, using Age of Empires II as a counterexample.
- De Wynter proves Age of Empires II is functionally and Turing-complete, making it a formal stand-in for LLM behavioral claims.
- The paper proposes a 'null assumption': treating LLM non-uniqueness as the experimental baseline rather than assuming human-like attributes.
Read full analysis →
View on Bluesky →
Their own posts
Recent commentary
My Twitter timeline is full of Google guys announcing they are leaving Google for Anthropic, Anthropic guys going to OpenAI, OpenAI guys going to Google or Meta, and Meta guys going, "Save us, we are in hell!"
Do I detect a shift in the Force from AI back to cybernetics? 🤔🧐
One flavor of AI hype I strongly dislike (and is happening on Twitter, today, to some extent. This is a subtweet) is in which a known insider makes a sensational, breathless assertion about the current state of AI capability, which then gets credulously re-broadcast, assumed true... gets pushback...
I really need to stop letting Anthropic press releases capture my attention in any way.
I am doing this backwards, reading the Dehaene and the Eleos group's responses to the Anthropic J-space claims, before I go read the Anthropic paper, and that's probably not a good way to do it, but I am.
My Twitter timeline is full of people announcing new AI or AI-related startups.
"We believe the future of AI is a collective of modules interacting and coordinating..." Yeah, that's also the past of AI. That's just Minsky.
Alright... I was planning to sort of live-tweet my thoughts on the Anthropic J-space thing, after reading the responses, but... sorry... is the 'global workspace' they 'found' just the residual stream, which is, by definition, a... global... workspace...? (Or am I missing something?)
I've lost the link (I'll find it and add it to the replies below) but someone on Twitter made this banger phrasing I want to luxuriate in: "What has really happened is VC built the wrong AI."
I don't see how selling AI inference retail, and then selling a small % of vastly overvalued stock back to the government aligns with the stated goal of 'benefiting all of humanity.'
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.