Ethan Mollick
Remember how a couple days ago us educators were feeling comfortable that ChatGPT produces B essays & can't do citations & makes up stuff all the time?

I had the new Bing AI create an assignment and rubric. Then I had it write an essay in response.

Not perfect, but a huge leap. https://t.co/7PpEuTa7dn
tweet

Hugging Face
RT @matthieurouif: @julien_c PhotoRoom Generative AI works on the logo you love in 10s

All powered by @huggingface, @StabilityAI and @photoroom_app https://t.co/gprCuJAYMi
tweet

John Nay
LLMs as ML Task Priors

Priors extracted by multi-shot prompting an LLM can be incorporated into any learning setting where natural lang description is available

-Causal inference
-Feature selection
-Reward shaping for safe reinforcement learning

Paper https://t.co/OGedKPPjJg https://t.co/10eL68cra9
tweet

illustrata
gn!

Turn on notifications to stay up to date!

#aiia https://t.co/CpQvcltcco
tweet

trash
this concludes the “brownies edges are mid” batch. https://t.co/dmGyPs23Wo
tweet

Robert Scoble
RT @heyBarsee: NeRFs of the Louvre made from a handful of short iPhone videos shot using AI.

Those cinematic shots 🤯

Lovely work by @paultrillo
Made with @LumaLabsAI

What's your thought on this? https://t.co/HRMchjt56d
tweet

Jordan Burgess
RT @RazRazcle: These three steps are an emerging pattern for LLM self-improvement used in all these recent papers:

‣ Toolformer
‣ STaR — LLMs that are better at reasoning
‣ Constitutional AI — Harmless AI from AI feedback

and the whole process can be repeated multiple times! https://t.co/tECMiGylEq
tweet

Ethan Mollick
Bing guardrails are much stricter than ChatGPT, prompts that work well in ChatGPT do not always fly with Bing. It wants to do search and analysis, it has no time for your fun & games.

And it gets pretty "mad" at you if you keep trying to cross that line. I actually felt guilty! https://t.co/SsUa7onJ7L
tweet

Linus (●ᴗ●)
RT @mindykoch: @LinusEkenstam Ooh, I have had too much fun with this one! https://t.co/vjMOmJMFdH
tweet