Google and Janelia released the complete fruit-fly connectome on 3 September. Wired to the Daedaluzz mazes beside Echidna and Wake, how far does it get?
Research
Articles and research reports from the team.
Latest articles
From the engineers who build Wake and run the audits.
Google and Janelia released the complete fruit-fly connectome on 3 September. Wired to the Daedaluzz mazes beside Echidna and Wake, how far does it get?
ack3 is verified under OpenAI's Trusted Access for Cyber and holds Daybreak Blue access. What the program is, and why verified access to OpenAI's cyber models matters.
Monad's Metropolis hackathon runs six weeks from September 1. ack3's part: an AI scan for every finalist project and a mentor through the build window.
Can deterministic static-analysis MCP tools improve audit results from non-deterministic AI? Three GPT-5.6 models reviewed the same Go codebase with and without them.
Three EVM fuzzers ran eight hours per maze against 153 reachable Daedaluzz assertions; an agent-written harness ran once. The deciding variable was not speed.
Sherlock launched the Audit Engine with ack3 among its AI auditors. What the engine is, and where else the ack3 AI scan runs.
How good are frontier models at finding real vulnerabilities? Nine models measured against human auditors on five audits that never appeared in training data.
NEST automates LDO buybacks funded by Lido DAO treasury earnings. Our security review covered the contracts, the fixes, and the verified mainnet deployment.
Browse by topic