Still no moat.

The reason someone would use this vs. third-party alternatives is still the fact that the $200/mo subscription is markedly cheaper than per-token API billing.

Not sure how this works out in the long term when switching costs are virtually zero.

I think at this point the aim is less about moat, and more about getting an advantage that self-sustains: https://www.rand.org/pubs/research_reports/RRA4444-1.html

> Not sure how this works out in the long term when switching costs are virtually zero.

All these not really helpful, but vendor specific, "bonuses" sounds like a way to try to lock people in, to try to raise the switching cost.

I'm using, on purpose, a simple process so that at any time I can switch AI provider.