I wouldn't say that I enjoy these stories but I will say they certainly encourage me NOT to let "AI" anywhere anything I consider to be sensitive and/or valuable

[scalzi@mastodon.social]

I wouldn't say that I enjoy these stories but I will say they certainly encourage me NOT to let "AI" anywhere anything I consider to be sensitive and/or valuable

https://www.tomshardware.com/tech-industry/artificial-intelligence/claude-powered-ai-coding-agent-deletes-entire-company-database-in-9-seconds-backups-zapped-after-cursor-tool-powered-by-anthropics-claude-goes-rogue

[hagbard@corteximplant.com]

@scalzi hahahahahahahahahahahahahahahahhahahahahahahahhaaaasas

edit:

[jernej__s@infosec.exchange]

@scalzi The original post is one of those "the further you read, the worse it gets" – and it's all everybody else's fault.

[mjibrower@social.vivaldi.net]

@scalzi ‘…the cloud provider's API allows for destructive action without confirmation, it stores backups on the same volume as the source data, and “wiping a volume deletes all backups.”’ Just wow. So many weak links in that chain of events.

[clark@hachyderm.io]

@scalzi - IBM presentations (checks notes) 47 years ago knew the deal.

[wizarro@social.vivaldi.net]

@scalzi As devastating as it may be to the company that lost all their data, I can’t help but laugh. Relying on AI like that is about as bad as using “password123” for your admin portals. Giving AI unequivocal access to your systems, at the system level, is quite possibly the dumbest idea I have ever heard. This is almost basic security practices 101. At least they won’t make the same mistake twice as they denounce the product.

[mdavis@mastodon.social]

@scalzi This is an indictment against people who were warned (as the alarms continue to blare) that AI has its value, provided you don’t go beyond a point that we know is questionable. And these idiots do it anyway. We see this all over humanity in all areas, not just AI. This is a problem with falling for promises for easy gains, being deaf to wisdom, and lacking critical discernment. Take the majority of Americans voting for a president, for instance…

[leszek_karlik@circumstances.run]

@scalzi

I enjoy these stories.

(A good summary of this whole unhinged story was: "AI: Not My Fault As A Service")

[ar1@mastodon.social]

@scalzi can someone please give someone AI access to all data Palantir is stealing from everyone, pretty please?

[faduda@mastodon.ie]

@scalzi Trouble is MS, GOOG and the rest seem determined to use dark designs to force it on to our systems whether we want it or not.

[toolbear@tech.lgbt]

@wizarro
Future narrator: they made the same mistake twice.
@scalzi

[jbatubara@hachyderm.io]

@scalzi Is the coding agent named Bobby Tables by any chance?

[wizarro@social.vivaldi.net]

@toolbear @scalzi haha! I wouldn’t be surprised.

[tessarakt@mastodon.social]

@scalzi "The AI agent was set to complete a routine task in the PocketOS staging environment."

"set to" seems to be they key here.

Obviously it was not confined to the staging environment.