[chistev]

If it says no, you move on to a competing model that will say yes. These companies with their models are always competing. There will always be a model willing to fill in the deficiencies of others because of... Money.

For example, ChatGPT refuses certain sexually explicit prompts, or certain NSFW prompts that are not sexual, but Grok will do as it is told.

[Jang-woo]

That's a good point.

I think you're right that at the model level, competition pushes toward "always say yes."

What I'm wondering about is whether control needs to exist at a different layer — not in the model itself, but in the system that decides whether actions are allowed to execute.

In other words, even if a model is willing to say "yes," the system using it might still need to decide whether execution is permitted.

Otherwise, it feels like we're relying entirely on model behavior for safety, which seems fragile in competitive environments.