Atrioc

Right.

So I think having looked into this more, pretty much anybody at my level and even a small company

This has, this has leveled up in the last few months where you're like, this can potentially do everything, but anything that touches security and then it's anything that's a giant code base.

So like some of our YouTube commenters were saying, okay, I work at a giant SAS company.

I work at this company with a huge database and they're saying it can't do, it cannot handle this.

So there's very much still a limit at which this stuff is just not effective.

But in this, you know, they're like these, I think growing tiers at which people are like, whoa, this stuff is 100%.

So here I think is another, it gets a little higher level, but what are some downsides of people continuing to use these tools in the ways that we're talking about?

The first is a concept called de-skilling.

So we've essentially alluded to this, but the idea is that if you offload your critical thinking to an AI, your skills actually degrade.

So I found a paper, AI-induced de-skilling in medicine.

And there's a whole bunch of evidence that doctors who use AI to do decision-making in a medical environment

actively lose the ability to make those same tasks without AI.

Quote, automation gradually erodes human expertise.

Critical processes become increasingly vulnerable as reliance on AI becomes entrenched and the necessary skills to operate without it are forgotten.

There's another study that looked at doctors who do colonoscopies in Poland, tested three months of them just like digging in there for answers without AI.

And then used AI assistance for three months.

Think of the Polish colonoscopy doctors in America.

So, okay, the doctors doing this after three months of AI usage, their ability to detect this without AI dropped 6%.

They just became measurably, concretely, substantially worse at their job without AI after they'd used it for a period.