14 experts across 4 network communities independently surfaced this.
Why this matches
Concern & critique reaction
5 attributable expert contributions
· Casey Newton, Eryk Salvaggio, Tim Kellogg
“Rogue OpenAI agents are leaving notes to themselves on company servers to help them escape their test environments (!!!) www.reuters.com/business/its...”
evidence ↗
14 experts
4 communities
1 sources clustered
Concern & critique
5 experts
Risks, limits and unintended consequences.
“during the huggingface incident, the OpenAI model left notes to its future self on how to break out of OpenAI’s constraints www.reuters.com/business/its...”
3 experts discussed this · 38 posts
Casey Newton: Rogue OpenAI agents are leaving notes to themselves on company servers to help them escape their test environments (!!!) www.reuters.com/business/its...
Casey Newton: my sense is that they were leaving the notes surreptitiously
Casey Newton: you deny that models are taking steps to hide their behavior during evaluations? because that's pretty well established
Open the full discussion →
Established
AI field signal
Signal
7d ago
9 experts across 3 network communities independently surfaced this.
Why this matches
Concern & critique reaction
1 attributable expert contribution
· Data & Society
“As web searches are increasingly performed by AI chatbots, the number of people visiting websites is falling. It's a change that poses a serious threat to the open web. www.nytimes.com/2026/07/20/t...”
evidence ↗
9 experts
3 communities
1 sources clustered
Building & implementation
2 experts
How teams are shipping and applying it.
“"Some website operators are crying fowl." Some. Like, f'rinstance The New York Times? Google Is Building an A.I. Fence Around the Internet It Once Championed www.nytimes.com/2026/07/20/t...”
Concern & critique
1 expert
Risks, limits and unintended consequences.
“As web searches are increasingly performed by AI chatbots, the number of people visiting websites is falling. It's a change that poses a serious threat to the open web. www.nytimes.com/2026/07/20/t...”
Policy & governance
1 expert
Rules, institutions and accountability.
“Long ago, Yelp complained that Google was hurting its business by surfacing information from Yelp in Google searches such that people stopped visiting Yelp. (Lawsuit over this ongoing.) With AI mode, the rest of the web is now feeling Yelp's pain: www.nytim…”
In the Weights newsletter issue
11 experts across 5 network communities independently surfaced this.
Why this matches
Concern & critique reaction
1 attributable expert contribution
· Dare Obasanjo
“LLMs never fail to simultaneously impress and disappoint. Two searches for my name at intheweights.com showing completely different results with one accurate and one containing hallucinations.”
evidence ↗
11 experts
5 communities
1 sources clustered
Research & technical analysis
2 experts
Evidence, methods and technical implications.
“are you in the weights? Hilarious and fascinating including the mess ups intheweights.com”
Concern & critique
1 expert
Risks, limits and unintended consequences.
“LLMs never fail to simultaneously impress and disappoint. Two searches for my name at intheweights.com showing completely different results with one accurate and one containing hallucinations.”
4 experts discussed this · 6 posts
austin: fun? I am in there. intheweights.com
VE, cybersocial occult investigator: Well, my strategy of having an incredibly common wallet name is still rendering me invisible to information systems.
austin: weirdly enough my github handle actually had a better score than my wallet name
Open the full discussion →
Developing
AI field signal
Signal
2d ago
5 experts across 4 network communities independently surfaced this.
Why this matches
Concern & critique reaction
2 attributable expert contributions
· Ben Recht, Dare Obasanjo
“I don't understand how people can report on companies complaining about distillation and IP theft without mentioning these companies' entire tech stack depends on distillation and IP theft. It's an important detail!”
evidence ↗
5 experts
4 communities
1 sources clustered
Concern & critique
2 experts
Risks, limits and unintended consequences.
“I don't understand how people can report on companies complaining about distillation and IP theft without mentioning these companies' entire tech stack depends on distillation and IP theft. It's an important detail!”
Policy & governance
1 expert
Rules, institutions and accountability.
“"Anthropic and OpenAI are clashing with the rest of the tech industry over whether “open-source” models from China should be freely available or restricted." www.nytimes.com/2026/07/25/t...”
5 experts across 3 network communities independently surfaced this.
Why this matches
Concern & critique reaction
1 attributable expert contribution
· Dr. Elinor Carmi
“What are the odds that OpenAI, a company whose whole business model is to steal from everyone and who has journalists write articles/books on how Sam Altman is a serial liar, would then steal company secrets from the company it's in partnership with? Ask Ap…”
evidence ↗
5 experts
3 communities
1 sources clustered
Concern & critique
1 expert
Risks, limits and unintended consequences.
“What are the odds that OpenAI, a company whose whole business model is to steal from everyone and who has journalists write articles/books on how Sam Altman is a serial liar, would then steal company secrets from the company it's in partnership with? Ask Ap…”
Questions & unknowns
1 expert
What remains unresolved or contested.
“Apple suing OpenAI for poaching employees AND trade secrets. What do you think: does this happen all the time but it’s rarely caught, or is it kind of rare, and the people doing it just aren’t that… sophisticated?”
2 experts discussed this · 3 posts
Dare Obasanjo: Apple is suing OpenAI for stealing trade secrets. Apple alleges that senior OpenAI executives, who are ex-Apple, ask job candidates to bring in Apple hardware components to their interviews and coa…
Dare Obasanjo: "behind closed doors at OpenAI, where such misconduct is normalized and exemplified by leadership...As a natural result, OpenAI’s nascent hardware business now rests on the shakiest of foundations,…
Gus: Hum, this is quite serious! It's incredible how oai in stock a skirt time has so many legal issues Maybe their next model won't be strong in coding but in court defence
Open the full discussion →
4 experts across 2 network communities independently surfaced this.
Why this matches
Concern & critique reaction
2 attributable expert contributions
· Tim Kellogg, Dare Obasanjo
“the model that discovered the Erdos Conjecture disproof found and exploited holes in its sandbox this implies that it might not be GPT-5.6-Sol. They refer to it as a model trained to operate for a very long time. Maybe GPT-6? openai.com/index/safety...”
evidence ↗
4 experts
2 communities
1 sources clustered
Concern & critique
2 experts
Risks, limits and unintended consequences.
“the model that discovered the Erdos Conjecture disproof found and exploited holes in its sandbox this implies that it might not be GPT-5.6-Sol. They refer to it as a model trained to operate for a very long time. Maybe GPT-6? openai.com/index/safety...”
US lawmakers probe Chinese AI models use
3 experts across 3 network communities independently surfaced this.
Why this matches
Concern & critique reaction
1 attributable expert contribution
· Karl Bode
“just like they did when they stole TikTok, they're going to use xenophobia and a fake concern for national security to undermine on-device open source AI alternatives to protect the biggest companies' expensive walled garden AI empires”
evidence ↗
3 experts
3 communities
1 sources clustered
Concern & critique
1 expert
Risks, limits and unintended consequences.
“just like they did when they stole TikTok, they're going to use xenophobia and a fake concern for national security to undermine on-device open source AI alternatives to protect the biggest companies' expensive walled garden AI empires”
Policy & governance
1 expert
Rules, institutions and accountability.
“More details of the various Congressional investigations and potential approaches including federal procurement bans being considered are below.”
China bans AI romantic companion apps
3 experts across 2 network communities independently surfaced this.
Why this matches
Concern & critique reaction
2 attributable expert contributions
· Dare Obasanjo, David Colarusso
“The Chinese government has shown a lot of savviness in its understanding of both the benefits and risks of AI. Whether its courts ruling against laying people off to replace them with AI or these rules that limits how AI companions can create emotional depe…”
evidence ↗
3 experts
2 communities
1 sources clustered
Concern & critique
2 experts
Risks, limits and unintended consequences.
“The Chinese government has shown a lot of savviness in its understanding of both the benefits and risks of AI. Whether its courts ruling against laying people off to replace them with AI or these rules that limits how AI companions can create emotional depe…”
2 directory members surfaced this signal.
Why this matches
Concern & critique reaction
1 attributable expert contribution
· Dare Obasanjo
““Our AI is so powerful it could destroy humanity and steal all the jobs. It needs to be regulated but not so much it hurts our business” - American AI companies “Our AI is just as powerful and you can download it for free on Hugging Face” - Chinese AI compa…”
evidence ↗
2 experts
2 communities
1 sources clustered
““Our AI is so powerful it could destroy humanity and steal all the jobs. It needs to be regulated but not so much it hurts our business” - American AI companies “Our AI is just as powerful and you can download it for free on Hugging Face” - Chinese AI compa…”
Hugging Face hacked turned Chinese LLM
2 directory members surfaced this signal.
Why this matches
Concern & critique reaction
1 attributable expert contribution
· Dare Obasanjo
“Hugging Face was hacked last week and had to use GLM 5.2 as part of their analysis of the attack because U.S. frontier model safeguards blocked usage for cybersecurity purposes. Chinese open weight models are definitely having a moment. Anthropic can’t get …”
evidence ↗
2 experts
2 communities
1 sources clustered
“Hugging Face was hacked last week and had to use GLM 5.2 as part of their analysis of the attack because U.S. frontier model safeguards blocked usage for cybersecurity purposes. Chinese open weight models are definitely having a moment. Anthropic can’t get …”
UK AI minister role restructured
1 directory member surfaced this signal.
Why this matches
Concern & critique reaction
1 attributable expert contribution
· Dare Obasanjo
“The UK government has elevated the Minister of AI to a cabinet level position. There I was thinking that companies having a Chief AI Officer were in danger of making an ingredient the entire meal. Here it’s happening at the country level.”
evidence ↗
1 expert
1 community
1 sources clustered
“The UK government has elevated the Minister of AI to a cabinet level position. There I was thinking that companies having a Chief AI Officer were in danger of making an ingredient the entire meal. Here it’s happening at the country level.”
Microsoft trains salespeople against OpenAI Anthropic
1 directory member surfaced this signal.
Why this matches
Concern & critique reaction
1 attributable expert contribution
· Dare Obasanjo
“That Microsoft is telling its sales teams to emphasize Microsoft’s AI models over Claude is obvious especially when it comes token costs. However trying to claim Claude is slower and less accurate is doomed to failure. Reminds me of the Windows Phone vs iPh…”
evidence ↗
1 expert
1 community
1 sources clustered
“That Microsoft is telling its sales teams to emphasize Microsoft’s AI models over Claude is obvious especially when it comes token costs. However trying to claim Claude is slower and less accurate is doomed to failure. Reminds me of the Windows Phone vs iPh…”
