Linus Torvalds says AI-powered bug hunters have made Linux security mailing list ‘almost entirely unmanageable’
https://www.theregister.com/security/2026/05/18/linus-torvalds-says-ai-powered-bug-hunters-have-made-linux-security-mailing-list-almost-entirely-unmanageable/5241633
23 Comments
Just use AI to manage the mailing list, then use AI to check the bugs and finally use AI to fix them 🙂
Maybe it’s time to upgrade to a Discord channel then (/s lol)
Linus will now think on how to solve this issue and will come out with another great product
Last time Linus got sufficiently annoyed about tools used for managing Linux development, we got git
The penguins emperor lol apt name, or perhaps the emperor penguin.
Thanks vibe coders. Once again, you improved everything!
People are talking about expecting him to come up with a much better solution. But to be honest I always thought that a “mailing list” was very archaic and it’s not like he has to come with a completely new and revolutionary system to a “mailing list” himself.
Knowing Linus the solution will be some tool he rage-codes in a weekend that accidentally becomes critical infrastructure for the next 30 years.
They need to point another LLM at the mailing list and have it remove all the dupes and then send only the useful stuff to an actually *useful* mailing list…
He should use AI to sort the reports.
If all you have is an hammer, everything looks like a nail.
Isn’t it a good thing that these bugs are being discovered? Of course a mailing list won’t be able to handle the load.
Hacking was hard because you couldnt test 10*10^100000 passwords fast enough.
everything change since average user having 64 and big corpo infinite amount.
Password = your door lock.
an idiot with youtube have high chance to get in 🙂
“AI found this issue”
cool.
now explain the kernel subsystem involved without opening another tab.
That’s basically the problem Linus is talking about imo.
I am tempted to say Linus should filter the reports with ai, but i really diske the idea principally.
“AI has become an increasingly useful tool for the FOSS community.” I see it the same way. Code review is the gold standard, and it was mostly companies that were able to accumulate enough money to pay for the service. Open source alone has not offered the same level of confidence in the code only because everyone can read the code. Now, AI can do it and help improve the code of FOSS applications.
“AI tools are great, but only if they actually help, rather than cause unnecessary pain and pointless make-believe work,” he wrote. “Feel free to use them, but use them in a way that is productive and makes for a better experience.”
– Bro declares hate on VibeCoders
AI makes everything easier.
And as is often the case that’s actually not a good thing.
Problem are Ai slop reports that look genuine but are hallucinated and are not really reproducable.
And updates too. There is update almost every single day.
So true. Every new tech tool seems to fix one problem and create another. AI’s no exception.
Would love to see a demographic breakdown of the people who spam this list.
Then ban the accounts using them!