CEO of AnchorWatch Rob Hamilton reported that on August 8th, he began integrating OpenAI's cybersecurity program capabilities into his work with Bitcoin Red Team. However, the very next morning, he encountered an access restriction.
Yesterday I began fully integrating the @OpenAI trust cyber program into my Bitcoin Red Team efforts.
— Rob Hamilton (@Rob1Ham) August 9, 2026
This morning I woke up to see this.
I am now being blocked from doing additional analysis on a codebase which I've already responsibly disclosed to, and have received... pic.twitter.com/Ain5RbOBDN
As a result, Hamilton was forced to revert to Chinese open-source AI models.
"For clarity: this happened after I had already completed KYC and finished the onboarding process months ago to use the cyber capabilities OpenAI offers," he clarified.
According to Hamilton, due to the restrictions, he was unable to continue auditing code changes and checking if other potential vulnerabilities remained. He added that this approach by American AI companies cuts defenders off from the most powerful tools, while malicious actors face no such constraints.
When asked about Anthropic's policy regarding trusted access to its most powerful models, Hamilton replied:
"Haven't had problems yet. But I'll probably be downgraded from Fable to Opus, so it's kind of the same thing."
Bitcoin Red Team is a volunteer group that uses AI and manual review to analyze hundreds of open-source repositories related to Bitcoin. As of August 9th, the team of 25 participants had found 1,288 critical and high-level vulnerabilities in the Bitcoin ecosystem.
🚩 Bitcoin Red Team Update
— calle (@callebtc) August 8, 2026
We're continuing a large-scale security review of the Bitcoin open-source ecosystem.
25 Bitcoin developers around the world have been working on this task non-stop for 108 hours. pic.twitter.com/CwNXHgRmdn
In one tweet, Hamilton wrote that he and another initiative participant under the nickname calle are entering a "very Chinese period of life."
The Bitcoin Red Team stated that OpenAI and Anthropic's cyber models deliver "incredible results," but the majority of their cumulative AI spending still goes to Kimi K3 — nearly 75%. The share of Qwen 3.8 is also growing.
Our AI spend remains consistently high.
— calle (@callebtc) August 8, 2026
Our cumulative spend has passed $58K as we add more inference sources. Notably, we've started using cyber models from OpenAI and Anthropic.
These models are producing incredible results, but the vast majority of our spend (74%) still goes... pic.twitter.com/DGPn2aElxC
It is worth recalling that in July, 25 American companies urged US authorities not to impose restrictions on Chinese AI models.
The Dark Side of Kimi








