Lots of the decisions described in this paper I had to face myself last year. It’s good to see them captured here.

Lots of room for expansion as one could also discuss the business model options that inform/constrain the release decisions. Probably not for this paper. https://t.co/wTXyr4tz1K

One thing we haven’t seen yet is the use of #GenerativeAI to manipulate stock market performance. Especially around earnings season. That will be a very interesting example of mass-disinformation. One that is way more enticing for many more actors than political influence.

If you are really keen to see how it feels to use a #ChatGPT-like AI to do search, you don’t have to wait for Microsoft to unlock Bing for you. You can try @perplexity_ai and their Chrome extension. I’ve been using it for a few days and it generates interesting results: https://t.co/BfIIOqwqG1

Last November I was in Iceland for a personal trip and my biggest regret is that I didn’t have the time to squeeze in a meeting with the @vivaldibrowser dev team in Reykjavík. I really wanted to pay a drink to the people that make the best browser in the world. https://t.co/SOp82sOf85

Looks familiar… https://t.co/KuPNUmVdMe

RT @Replit: Without any previous ML experience, @coding398 used 40,000 images to train a Convolutional Neural Network on detecting specific…

👀 https://t.co/ciZmx7NNEW

👀 https://t.co/wTXyr4tz1K

@wavymulder What model did you use for this one?

@GaryMarcus @ylecun The transformation is not yet complete: now *you* should start saying that Galactica is amazing to confuse the hell out of everybody 😉