👀 https://t.co/npu0mf8Dzf

If you don’t integrate with #ChatGPT you are a loser 🫠

Speaking of #AI tools and large language models today: I really like how @perplexity_ai is evolving their Chrome extension. I now can use it not just to ask questions but also to summarize specific pages. https://t.co/xV8ZBDMME7

Wonderful news: @LLo_ai, maker of the must-have-or-I-won’t-use-my-macOS tool #BetterTouchTool has implemented an integration with #ChatGPT. Well, in fact, two.

Which gives me a great excuse to mention it in this week’s issue of Synthetic Work.

https://t.co/N3YjKONlKg https://t.co/wttOgWWgnr

One of the most foolish thing I heard in the last few years is that the #AI opportunity is unlocked by supporting NVIDIA GPUs. 1. it’s picking up nickels in front of a steam roller 2. Google, Meta & others don’t want to be locked in a mono-vendor relationship. PyTorch is the way.

And if not PyTorch, then another cross-hardware framework that allows Google’s expertise in GPUs maintain an edge that wouldn’t have in a NVIDIA-only world.

Do you really think that any big tech vendor that manages to invent the ultimate chatbot, like the one in “Her”, adopted at planetary scale, would ever want to depend on NVIDIA to make performance and efficiency quantum leaps?

Synthetic Work Issue #1 Splendid Edition, out last week, was dedicated to how #AI is changing the #Education industry and the impact it’s having on teachers.

If you know a teacher that might be interested in reading that content, let me know: I’d be happy to send it for free. https://t.co/Clh8T5rt62

.@lewingtonpitsos has published the simplest-to-understand and clearest explanation I have ever heard of how the noise function in #stablediffusion works.

This man is so good at explaining very complicated concepts. You should really follow his work.

https://t.co/jXb0mjukFi

I don’t know how to tell you this, but hands are happening (thanks to #ControlNet):

https://t.co/9aZgrfvOlu https://t.co/ZzvbW0E9hB