It's all the rage these days. Let's collect the ones we've heard about so far.

Text

• GPT-4, the latest large-language model (LLM) by OpenAI.
• GPT3 (known as ChatGPT), the previous generation LLM by OpenAI.
• Falcon 40B, FOSS LLM. Video review by Sentdex.
• https://chat.lmsys.org
• https://github.com/declare-lab/flan-alpaca
• https://scifilogic.com/foss-llm-that-you-can-run-your-pc/
• https://www.technologyreview.com/2022/07/12/1055817/inside-a-radical-new-project-to-democratize-ai/
• https://www.databricks.com/blog/2023/04/12/dolly-first-open-commercially-viable-instruction-tuned-llm
• Elicit AI-powered science paper search tool (not FOSS, behind sign-up wall)

Image

• Scribble Diffusion, turn a scribble (and an associated text prompt) into a refined image. Powered by ControlNet. Simple to use, generated images 512px square and easy to download. Seems to work well in Chromium, less so in Firefox.
• Midjourney, generative AI service that creates images from language prompts. How to use Midjourney, Tom's Guide.
• https://drawfast.tldraw.com, via @simon@simonwillison.net (service has been shuttered!)

I'm well aware there are like a 100 more, will add to this list as I get time.
On a related note, it's too bad Shaarli does not allow others to edit or at least comment on posts (but see this post regarding developments).
But if you really want, I guess you could open the post on its own page (see permalink below) and annotate that page using Hypothes.is.

Links and notes

• https://www.wired.com/story/how-chatgpt-works-large-language-model
• https://changelog.com/podcast/578 What exactly is Open Source AI?

This unofficial site let's you explore the In Our Time archive by topic and year.
Via kottke.org.