(LLM-based) AI is slow to spread because the reality of what it can and cannot do is apparent. Anyone that wants to can use it for free & observe the results for themselves.

This isn’t hard, people.

1. LLMs are wrong. A lot. That’s a straight up liability for anything customer-facing.

2. Use cases beyond “everyone uses Gemini to help write emails” are variable and often need to be bespoke. That requires consultants, money, and effort.

3. Concerns about access to proprietary data.

Solve even 2 and most companies are good to go.

Because it’s not used properly! They trying to compete each other and make money. That their only goal! It was supposed to help humanity I guess.