Bitsec's multi-model agent stack found 160+ exploits, beating a single 'superhuman' model
markjeffrey · x · 2026-09-16
On the Twist Startups podcast, Bitsec explained why it outperforms Fable at vulnerability discovery: models have distinct personas, each good at certain bug classes. By combining a buffet of models with specialized agents and rigorous benchmarking instead of relying on one "superhuman" model, their approach (Bittensor SN60) reportedly found 160+ vulnerabilities in a client codebase, including 5 criticals and 10 highs that Fable missed.
More from coding & agent
- AI Evals FAQ grows to 48 Q&As: sensitive data, huge traces, and stale gold datasets — HamelHusain · 2026-09-16
- One prompt to a real object: AI agent models in Blender and 3D-prints the result — anthara_ai · 2026-09-16
- Open-source crew of 8+ AI agents manages your Obsidian vault, email triage included — tom_doerr · 2026-09-16
- AI agent gets user's $70 Amazon credit after weeks of runaround, one tap needed — armand_ruiz · 2026-09-16
- 3 weeks through Stanford CS329A: the generator has outrun the verifier — le_james94 · 2026-09-16
- Generator outruns verifier: 250 samples lift DeepSeek-Coder to 56% on SWE-bench — le_james94 · 2026-09-16