🔥 OpenAI launches ChatGPT API

PLUS: Can Stable diffusion read minds?

Hey hey what’s up?

Welcome to the 210+ new subscribers who joined since the last edition. This was a massive week in AI with lots of new developments making it feel like Christmas every day. With OpenAI being Santa 😉. This is a long issue (you can thank the new ChatGPT API), so strap in!

What's cooking in today's edition? 🧪

  • OpenAI’s launches the ChatGPT API

  • Corridor Crew just changed animation forever

  • Can AI really read minds?

👋 Psst! New here? Subscribe to receive the top AI news 2x/week + exclusive content. Follow us on Twitter!


This is the big news for this week ladies and gents. Every single GPT-3 based AI product out there just got a big shiny upgrade, and you know what the best part is?

It’s 10X cheaper than the previous model, and devs are loving this.

Here’s a quick tl:dr on the API

What’s under the hood? The model is gpt-3.5-turbo, which is the same model used in the ChatGPT product.

Cost? It is priced at $0.002 per 1k tokens, 10x cheaper than the existing GPT-3.5 models or approximately 750,000 words or 5 Harry Potter books.

Who can use it? Anyone, head over there!

OpenAI also launched their Whisper API, their speech-to-text model which is icredbible. Hasn’t gotten as much love as the ChatGPT API, but this is very soon (if not already) powering Alexa, Siri and other Voice UI.


Corridor Crew launched an animated youtube video with ControlNet and Stable Diffusion. And it’s f*cking awesome.

First, who's Corridor Crew? A very popular youtube channel that is the top dog of 3D animation. They experiment with AI often (very early into deepfakes and Stable diffusion) and create awesome videos.

In the video, they use AI models to convert a regular video into an anime.

The video is incredible—120 shots done by a small team of 3.

How did they do this? I’ll give you a quick ELI5.

  1. They train an AI model over the actor’s faces (Dreambooth - Remember Lensa?)

  2. Shoot a video with all the props in a greenscreen

  3. Use ControlNet to convert the individual video frames to an anime style

  4. Combine the frames into a video

  5. Add some VFX and tada!

Creating anime videos is hard, and new AI models like ControlNet reduce the effort by 10x and unlock lots of new opportunities for creators. I'm excited to see what's next!


Well..sort of. A team from Japan published a paper where they showed the 'presented images' to a group of humans. The 'reconstructed images' resulted from fMRI output to Stable Diffusion

Is this legit? Yup! The paper was accepted at a top AI conference.


  • Typeface emerges with $65 million to implement generative AI in businesses. That's a big fat round! (link)

  • Microsoft launches Kosmos-1, a new LLM model. Their model can understand images, text, images with text, OCR, image captioning, and visual QA. (link)

  • YouTube’s new head teases AI tools that can virtually swap creators’ outfits and locations (link)

  • Microsoft brings its new AI-powered Bing to the Windows 11 taskbar (link)

  • Robotics startup Figure unveils its humanoid robot. Things just got more exciting baby (link)


  • Dashworks - The AI search assistant for work.

  • Planmytripasap - Creates a custom, tailored trip itinerary.

  • Paxo - Creates meeting notes in minutes.

  • Goldie - Your AI sales co-pilot.

  • Auralbyte - Builds Short-Form videos with zero camera or design work.

  • DemoTime - Helps you turn demos into deals.

  • GetChunky - Trains your personalized smart assistant by using Google Sheets.

  • Polycam - Share your 3D scans and collaborate with anyone, everywhere.


  • Training your own custom Cohere models without machine learning skills (link)

  • How to build a GPT-3 Slack bot with Zapier (link)

  • Jaq n Jil + Zapier: Getting Started with AI Automations (link)

  • LangChain Beginner's Tutorial for Typescript/Javascript (link)

  • How to use Additive Prompting framework to generate characters/models, design sets, & more (link)

  • Redesign your room in seconds with AI! 100% free and open source (link)

  • All the different ways you can jailbreak ChatGPT! (link)

  • All about foundation models (link)

"soup dogg"

That's a wrap for today. Stay curious & see ya tomorrow! If you want more, be sure to follow on Twitter (@sudu_cb)

Loved this issue? Share it!

What did you think of today's edition? Reply to the email with 

  • 🔥🔥🔥 - This is f*cking awesome

  • 🔥🔥 - Meh, could be better

  • 🔥 - I want a refund.


PS: Move this email into your primary inbox! Don't want you to miss a beat!