INTRODUCING ZERO COOL 🥶 This has been an intense year for me. In January I earned over $120k from auditing - a personal record. I was booked solid for the next 5 months. Life was good. But in February I decided to walk away from it all to pursue my dream. It's been a wild ride, but today I am excited to announce the launch of Zero Cool, an agentic, language agnostic, AI-first security review platform that I co-founded with my friend and colleague Kurt (@i_am_phaze). The story starts back in 2022 when I first got into blockchain security. I was always fascinated by the idea of automating the review process and when ChatGPT came along my vision really began to take shape. I used AI a lot as a security researcher over the years, but I always knew I wanted to take it further and build something. So in January of this year, I made the decision. I found someone to replace me on the engagements I was booked on, reduced my client load to almost zero, and locked the fuck in. Feb: 💭 The first couple months were rough. I watched my peers continue crushing audits while I was grinding away without much success. March: 📉 This was a low point. Every day questioning if I'd made a terrible life choice. I was starting to feel depressed. April: 🔄 Needed a mental reset and funds were low so I took a three-week audit gig with the amazing @0xleastwood and @zigtur — super complex protocol, fun stuff, reminded me why I loved security in the first place. May: 💡 Started talking with Kurt, my old friend from Trail of Bits. He was also interested in building something, and he had some novel ideas that would later become the core of Zero Cool. My previous work, like most existing tools, focused on a highly structured process -- think checklists or Slither detectors. But Kurt's ideas integrate AI into every aspect of the agentic process fully embracing "the bitter lesson". We're also both fans of XBOW and we've incorporated some of its principles, like "alloyed agents" into our design. June and July: 😅 Locked in, heads down, a total blur honestly - but we made progress every day, always beating yesterday's benchmarks. In July, @aviggiano announced the Size Credit agent competition, and we decided to enter. We don't even have a front end, but we knew we had something special and we wanted to start testing it in public. It was a success, we came in first amongst the agent competitors, identified 55% (6/11) of all highs and mediums, including 3 findings missed by the human auditors, and we were the only agent team with a solo finding. It's crazy to think that we just started this a couple months ago and this is the WORST that Zero Cool will ever be! I'm so relieved that finally, today I can share what I've been up to this year. Sorry for the lengthy post but it feels good ngl, like a much needed emotional release. We have so much more planned, and we can't wait to share it with you. We've also got a bunch of tools we've been using that will be completely open sourced and shared with the community. If you're interested in following our journey, please follow @ZeroCool_AI and join the interest list at zerocool.ai -devtooligan