Which tasks have become newly feasible, cheaper, more reliable, or more autonomous?
The provider says AI was used across reconnaissance, exploitation and exfiltration, sometimes through multi-agent workflows. Defenders need faster adaptive loops, but the evidence remains provider-observed and selectively disclosed.
DeepMind and external evaluation partners report a way to test a proprietary model without revealing either its weights or the evaluator's private prompts.
Three counts describe three different parts of the incident. None supports the claim that 700 agents successfully hacked Hugging Face, but together they expose a wider evaluation-control boundary.
