Again, chasing nonsensical hype. We will see more of this as local models get better. Of course; Anthropic, OpenAI need to heavily subsidize their token pricing for growth and are still losing billions doing that.
At some point they are going to either raise prices on their APIs and / or go after compoevery other "AI business" to justify their valuations. Expect Anthropic to cut off some of their long-term customers from using Claude.
Right now everyone (except Google) building AI bot infrastructure that doesn't make money on top of the debt that was borrowed to fund it will cause an unpleasant crash.
Where the heck did that come from? I've been toying with openclaw and it's pretty fun, but also so fragile and messy. It also suffers from what might be a terminal trade-off: either you trust it to ruin your life and it acts on your behalf, or you keep it very restricted and it's just a LLM with a heart beat and cron jobs. I can't imagine how anyone thought this would have been a profit-driver, there's so little tech there.