2024-11-02
Exciting work from @GoogleDeepMind and Project Zero: perhaps the first public vulnerability discovered using an LLM agent in a widely-used software component. Early days, but shows the promise of LLMs to find and fix vulnerabilities before they ever make their way into production
PCMag
Google researchers detail Project Big Sleep, formerly Project Naptime, which used an AI agent to discover a previously unknown and exploitable bug in SQLite
The company's experimental AI agent finds a previously unknown and exploitable software bug in SQLite, an open-source database engine.
2024-08-30
Noteworthy new research here on APT29 use of old NSO and Intellexa exploits. In my opinion, this is yet more evidence that the proliferation of commercial hacking capabilities — sometimes into the hands of apex teams like APT29 — poses a serious threat to the web.
TechCrunch
Google says Russia-linked APT29 is using exploits that are “identical or strikingly similar” to those previously made by spyware makers Intellexa and NSO Group
«Russian government hackers found using #exploits made by spyware companies NSO and Intellexa: … Frederic Jacobs / @fj@mastodon.social : “Our latest n-day exploit reporting shows t...
2024-08-29
Noteworthy new research here on APT29 use of old NSO and Intellexa exploits. In my opinion, this is yet more evidence that the proliferation of commercial hacking capabilities — sometimes into the hands of apex teams like APT29 — poses a serious threat to the web.
TechCrunch
Google says Russia-linked APT29 is using exploits that are “identical or strikingly similar” to those previously made by spyware makers Intellexa and NSO Group
Google says it has evidence that Russian government hackers are using exploits that are “identical or strikingly similar” …
2023-08-10
Can AI help secure open source software projects? We'll find out! Awesome @DARPA initiative being announced this week at #BlackHat2023. @Google and @GoogleDeepMind are excited to sign on as partners. Congrats to @perribus and team. https://whitehouse.gov/...
Engadget
The US debuts a DARPA contest to build AI systems that can proactively identify and fix software flaws, with help from Anthropic, Google, Microsoft, and OpenAI
a competition to automatically find/fix software vulnerabilities. Details: https://www.darpa.mil/... [image] Charley Snyder / @charley_snyder_ : Can AI help secure open source soft...