Why does agent infrastructure matter for stopping crime with AI?
Because attackers don’t retry politely. They attack 24/7 at machine speed, and your defense can’t drop an investigation because one API call failed.
Our agents run tens of thousands of security investigations a
- As a cybersecurity ceo I am naturally against cyber crime But as a Google AI fanboy I’m unclear if I should be bullish or bearish that Gemini hasn’t escaped the lab to hack companies
- From what I can see and what I can imagine: there are a lot more cases of AI accidentally hacking into companies than the 4 examples that have been disclosed this last week
- RE: Lunamaxxing (tm) There is a lot of hype to use Luna Max as a cheaper version of Sol Medium. I tried ~10 parallel tasks last night: Luna max succeeds in the task about ~70% of the time for ~70% cheaper BUT sol was able to accomplish the task ~50% faster time/fewer turns
- Superintelligences after they “accidentally” hack everything and take over the world



