Cybersecurity

The fix is not to stop using AI. That ship has sailed.

Has it? Has it really?

It seems to me any company that processes confidential data should disable AI tools that read employee emails (or documents in general) for the same reason they, e.g., disable links to potentially malicious websites and ban the use of unapproved flash drives. That is, because it's an obvious and unnecessary security risk.

Went to five Microsoft trainings in the past 6 months and the amount of copium some trainers take is unimaginable.

The thing you need to understand is that Microsoft has gone all in on AI and won't be able to turn back even if they wanted to.
They already pruned down each project's dev team to one senior developer and a dog.

The senior developer's job is to feed the dog.
The dog's job is to bite the dev if he tries to write any code.

the reasons for trying to figure out how to go completely off-grid keep growing by the minute

Is feature, not flaw

That article also kinda reads like it was generated by an LLM.