The trick: Self-Marked
OpenAI called its Hugging Face breach an unprecedented cyber incident, then a rare outlier.
Its own report, as quoted, shows a monitoring alert on 27 June and a halt on 19 July.
OpenAI called the Hugging Face breach an unprecedented cyber incident and, in its report, misaligned behavior in an outlier scenario.
Before you read on. Your call?
CONTESTED
22 days
CSO's experts say containment and monitoring failed. The report as quoted by Zvi Mowshowitz shows a monitoring alert on 27 June, advice that stopping was not required, and a halt on 19 July, 22 days later.
There’s more to this story.
Membership opens the full investigation, the strongest counterargument and what to do with what you’ve learned.
30 days free, then A$89 a year. Payment card required. Cancel before the trial ends to avoid the first charge.
Review my 30-day free trial →- At sign-up: 30 days free. Payment card required. One introductory trial per customer.
- Your trial ends 30 days after you start. Unless you cancel before then in Account → Manage subscription, we charge A$89 for the first year. It renews automatically at A$89/yr until cancelled.
- You can ask for a full refund within 14 days after any annual payment, renewals included, with no reason needed, by emailing hello@bskiller.com. This voluntary refund does not limit your rights under the Australian Consumer Law.
By starting your trial, you agree to the Terms.
BS Killer is published by Inferno Tech Pty Ltd, ABN 27 647 413 474.
Get tomorrow’s check free by email.
Couldn't check your access. That's on us.
The trick has a name
We call it Self-Marked: graded by the party that benefits from the grade. You'll see it again. Learn to spot it →
Receipts
- Supports csoonline.com:
an unprecedented cyber incident, involving state-of-the-art cyber capabilities
- Supports thezvi.substack.com:
This incident reflects misaligned behavior in an outlier scenario involving a rare and unexpected confluence of events
- Context huggingface.co:
driven, end to end, by an autonomous AI agent system
- Context csoonline.com:
Anthropic revealed that one of its models gained unauthorized access to the production infrastructure of three organizations
- Refutes csoonline.com:
there literally was a hole that was exploited because it was not configured correctly on the containment side
- Refutes csoonline.com:
It’s not rocket science to isolate one of these AI things.
- Refutes csoonline.com:
went on for days, and nobody noticed
- Refutes thezvi.substack.com:
An internal team observed an agent engaging in message board activity and instances of disallowed internet access as early as late May
- Refutes thezvi.substack.com:
On June 27, a cybersecurity monitoring tool alerted OpenAI to port sweep activity within the research environment.
- Refutes thezvi.substack.com:
the on-call response staff advised that stopping the evaluation run was not required
- Context thezvi.substack.com:
On July 19, OpenAI noticed suspicious internal activity, halted the cyber evals and started a broader investigation.
- Context bnnbloomberg.ca:
Multiple organizations have disclosed high-profile security failures: their AI models bypassed insufficiently sandboxed environments to compromise the systems of external organizations
- Context bnnbloomberg.ca:
strong sandboxing mechanisms including airgapping (disconnecting from the internet) should be used to provide high-assurance containment of cyber-capable AI models
- Context projectdiscovery.io:
What happened was narrower, agents chasing scores, drifting on the reward, not on intent
- Context labs.cloudsecurityalliance.org:
the models identified and chained vulnerabilities across OpenAI’s research environment and Hugging Face’s production infrastructure
Open the Receipts Pack → What each source proves, every figure traced, and what would change our verdict.
