Im guessing these are people that got in trouble for trying to submit incorrect AI slop and were specifically told that they are no longer allowed to use AI because they keep fucking up with it.
I’m not convinced they actually exist. Even the very conservative implanted medical device company I recently left, for reasons unrelated to AI, is adopting it.
But they’re doing it right. Generative AI is only used for internal tools and automation. The software that runs on clinical devices is still written using a rigorous manual process where every module, function, and operation is planned out, reviewed, and documented before any of the code is written.
The only clinical use of AI is in data analysis because (non-generative) AI can be really good at finding patterns in massive piles of data. Those results are then presented to physicians, along with the raw data, for review.
But the AI bros don’t understand the ideas of “appropriate context” and “safe adoption.” Since “move fast and break things” attracts startup capital, it must be the right approach for everyone everywhere in everything, right?
Where are these mythical tech companies that don’t allow AI? Where are the ones that aren’t forcing their engineers to use it?
Im guessing these are people that got in trouble for trying to submit incorrect AI slop and were specifically told that they are no longer allowed to use AI because they keep fucking up with it.
I’m not convinced they actually exist. Even the very conservative implanted medical device company I recently left, for reasons unrelated to AI, is adopting it.
But they’re doing it right. Generative AI is only used for internal tools and automation. The software that runs on clinical devices is still written using a rigorous manual process where every module, function, and operation is planned out, reviewed, and documented before any of the code is written.
The only clinical use of AI is in data analysis because (non-generative) AI can be really good at finding patterns in massive piles of data. Those results are then presented to physicians, along with the raw data, for review.
But the AI bros don’t understand the ideas of “appropriate context” and “safe adoption.” Since “move fast and break things” attracts startup capital, it must be the right approach for everyone everywhere in everything, right?
My company is currently quite restrictive on ai. They probably don’t even know and I won’t tell them.