I keep seeing a huge red flag when speaking to my network and friends at big tech companies.

It's either complaints about being _"forced to use AI and be more productive,"_ or bewilderment that _"GitHub Copilot is so good."_

I accept that working in a greenfield startup is different from an enterprise trying to change.

However, it's crazy to see that they're only now adopting technologies I was using before ChatGPT.

_They don't know what's coming..._

This helped me define what "Safe AI" means to me.

**We have to - not just need to - integrate superintelligence into our society. We have to leverage its capabilities to the max. At the same time, we cannot become overly reliant on them. Cognitive decline is not acceptable.**

This is a very delicate balance to strike, but we'll figure it out.

Education, at every level, will be where it starts.

_P.S. I might turn this into a full essay one day. Follow my [RSS feed](https://olshansky.info/index.xml) or [Substack](https://olshansky.substack.com) if you want to catch it._