Autonomous code analyzer beats all human teams at OSS zero-day competition
1–5 of 5 posts
Re: Autonomous code analyzer beats all human teams at OSS zero-day competition
#2This is the kind of advertisement-blog-post that should not have been published until after disclosure. These are quite extraordinary claims and that demands extraordinary evidence.
Re: Autonomous code analyzer beats all human teams at OSS zero-day competition
#3This is the kind of advertisement-blog-post that should not have been published until after disclosure. These are quite extraordinary claims and that demands extraordinary evidence.
Why should thy have waited further? Their claims about their performance in the competition are backed by the competition announcements.
Re: Autonomous code analyzer beats all human teams at OSS zero-day competition
#4This is the kind of advertisement-blog-post that should not have been published until after disclosure. These are quite extraordinary claims and that demands extraordinary evidence.
Why should thy have waited further? Their claims about their performance in the competition are backed by the competition announcements.
In the commenter's defense, it's reasonable to be skeptical about the level of autonomy claimed in the post.
We are very eager to share more evidence (including the raw inputs and output artifacts for these bugs) and will absolutely do so as soon as we can.
Re: Autonomous code analyzer beats all human teams at OSS zero-day competition
#5This is the kind of advertisement-blog-post that should not have been published until after disclosure. These are quite extraordinary claims and that demands extraordinary evidence.
This is fair, and we will gladly share the extraordinary evidence as soon as we can.
If you're curious, we have already released the full traces of finding a sqlite3 0day with an early version of Xint Code (submitted to the AIxCC competition and now open sourced): https://theori.io/blog/exploring-traces-63950