@benjamineskola@hachyderm.io so yes, that's the crux of the matter. The issue is, critics of LLMs are probably better informed than proponents, and to learn what happens in such an imbalance we only need to look at social networks and the relative market shares of the terrible, toxic ones (everything Meta, X, etcetera) and the reasonable ones (where we are now and others). Ignoring Facebook doesn't make a dang difference, building and using Mastodon does. I think I'm advocating a similar approach. Ignoring OpenAI and Anthropic won't work, building on open weights (or better yet, fully open source) models to see how we can have decent, non-coin-operated tooling is probably the better approach. It's rooted in a deeper notion that boycotts pretty much never work (that's more a fact than an opinion as far as I'm aware). It all hinges on your answer to the question "is this stuff useful to begin with?" and my answer is now "yes". If yours is "no", well, that changes the picture of course :-)