News for hackers. Stuff that matters.
Re:AI industry insiders launch site to poison the data that ... by lukan on Sunday January 11, 2026 @01:07PM (#46577973)
"They could even use an LLM to detect if the data has been poisoned."

And for extra safety, you can add another LLM agent who checks on the first .. and so on. Infinite safety! s/

Re:AI industry insiders launch site to poison the data that ... by archerx on Sunday January 11, 2026 @01:11PM (#46578023)
People already do this with multi agent workflows. I kind of do this with local models, I get a smaller model to do the hard work for speed and use a bigger model to check its work and improve it.
Re:AI industry insiders launch site to poison the data that ... by lukan on Sunday January 11, 2026 @01:22PM (#46578131)
The tech surely has lots of potential, but my point was just, that self improvement does not really work yet unsupervised.