The job responsibility section is very odd, coming from a frontier AI lab.
"Contribute to threat models and risk assessments"
Why don't agents do this?
"Curate and triage inbound intelligence from commercial feeds"
Summarizations are a primary use case of LLMs
if anyone’s interested in coming to help build out new capabilities and help drive security, detection and defense forward, i know an awesome team hiring!
https://t.co/e5027kyvAQ
In consultation with @SecWar, we will EXTEND the A-10 “Warthog” platform to 2030. This preserves combat power as the Defense Industrial Base works to increase combat aircraft production.
Thank you to @POTUS for your unwavering support of our warfighters and quick, decisive leadership as we equip our force. More to come.
@robtlee@OpenAI@AnthropicAI Yeah, it's that simple. give it a repo and a test env and tell to find vulners and it will go down that and validate the findings. Of course, if you craft more sophisticated prompt you will get better or more targeted findings. You don't need a man page, dawg 😂
@FuzzySec Mythos is a powerful tool but it still takes a skill person to steer the ship, so to speak. It reduces the time it takes to complete a test from weeks down to hours and in that time it finds the same vulners as a skilled human plus more.
I hate Mondays it’s the worst day of the week lol. Everyone blasts you in the morning and through the day - it’s just fighting the call, text, email storms. Balances out Tuesday - Friday.
This morning 11 texts, 77 emails, four calls all requesting responses.