RT @ecommerceshares: The largest gross profit pool in the whole world? Financial services, with $6.5 trillion. “Pure tech” (semis + softwar…

A new technique to fine-tune a diffusion model. Faster than Dreambooth >

Multi-concept Customization of Text-to-Image Diffusion https://t.co/LKtDO10cJy

1/ Now, of course, if @StabilityAI releases additional open building blocks, like a large language model (via @carperai) and a text-to-speech model, then the community could attempt to build the next-generation #AI assistant (terrible name…) to compete against Alexa, Siri, etc.

@StabilityAI @carperai 2/ It would be especially interesting to see an open #AI assistant that works on desktops.

Can a large language model control an operating system?

I have ideas about applying computer vision to the desktop for certain tasks, but maybe there’s a lot more that could be done.

Amazing how @liuliu continues building @drawthingsapp at a furious pace and gives it away to everyone for free.
Custom models are a superb addition.

(if it could only work in background!) https://t.co/wsXKXd8lVW

A great question. The answer is not surprising:

“For many teachers [reluctance is down to] a lack of confidence about how to teach something which they think they’re not good at themselves”

Why is personal finance not being taught more in schools? https://t.co/4dA29qBCrn

Incredible quality. If you want to do see what #stablediffusion 2.1 can do, follow @rainisto https://t.co/jOnWMTrxO5

I wonder what @philbarden thinks about this? 🙂

https://t.co/oyKVoRdTzn

RT @benedictevans: Chidlren online https://t.co/O4azsLJJ7y

@Merzmensch All established artists I know are experimenting with #AI or using it, not condemning it.