What's insane to me is the fact that the best interfaces to utilize any of these models, from open source LLMs to open source diffusion models, are still random gradio webUIs made by the 4chan/discord anime profile picture crowd.
Automatic1111, ComfyUI, Oobabooga. There's more value within these 3 projects than within at least 1 billion dollars worth of money thrown around on yet another podunk VC backed firm with no product.
It appears that no one is even trying to seriously compete with them on the two primary things that they excel at - 1. Developer/prosumer focus and 2. extension ecosystem.
Also, if you're a VC/Angel reading my comments about this, I would very much love to talk to you.
Automatic1111, ComfyUI, Oobabooga. There's more value within these 3 projects than within at least 1 billion dollars worth of money thrown around on yet another podunk VC backed firm with no product.
It appears that no one is even trying to seriously compete with them on the two primary things that they excel at - 1. Developer/prosumer focus and 2. extension ecosystem.
Also, if you're a VC/Angel reading my comments about this, I would very much love to talk to you.