NEWS OF THE WEEK

🤖ChatGPT Can See and Talk Now

The new version of ChatGPT can understand images and early testers posted great results on X.

This will be super useful for:

• asking how to fix something by taking a photo of the problem.

• generating images with DALL-E 3 then letting it provide feedback and refine them.

• taking a screenshot of a website and having ChatGPT recreate the code.

Here’s a collection of early use cases.

Additionally, it will feature a conversation mode where it will respond in voice – think of it as a Siri that works.

One more thing. ChatGPT will get back it’s ability to browse the web, a feature which OpenAI took away after users misused it to access paywalled articles. This means ChatGPT will be able to give you up-to-date info on any topic.

These updates will roll out to Pro users in the next two weeks.

OTHER IMPORTANT NEWS

OpenAI’s Sam Altman and John Ive want to create an Iphone for AI

In a universe-altering collaboration, Sam Altman and John Ive, who created many of the iconic Apple products alongside Jobs, have been brainstorming a device that would provide a better user experience interacting with AI than current smartphones.

Nothing is certain yet, but Softbank’s CEO, Masayoshi Son, said he would be ready to invest $1 billion into the project.

Spotify AI translates podcasts

Powered by OpenAI’s technology, Spotify has started rolling out a new voice translation feature to a select group of podcasters. The tool takes the podcaster’s voice and uses it to recreate the podcast in different languages. Spotify is slowly rolling out pilot episodes in Spanish, German, and French.

I gave it a listen, and I was floored. I can’t believe how well it translates and how natural it sounds.

Check it out here: Lex Fridman Podcast – “Interview with Yuval Noah Harari” in Spanish

Meta’s AI-Powered Chatoverse

Meta is rolling out chatbots to WhatsApp, Instagram, and Messenger. They will have a generic chatbot running on Llama 2, plus 28 celebrity chatbots. Expect to bump into Mr. Beast or Snoop Dogg in your DMs.

Amazon & Anthropic: AI's Next Power Couple

By far the best competitor to ChatGPT is Claude 2 from Anthropic. The company, founded by ex-OpenAI employees, is a solid second to GPT-4 and even outcompetes it in certain tasks.

Last week Amazon has announced a commitment of $1.25 billion and is open to investing up to $4 billion in the company. With the backing of a giant like Amazon, Anthropic aims to develop the next generation of its AI model, which they claim will be 10 times more effective than the current version. Clearly, the race to develop next-gen AI models is intensifying.

Amazon’s investment in Anthropic aligns with its ambition to position AWS, its cloud service, as a dominant force in the AI sector.

Getty Images launches an image generator AI that respects copyright

The company with one of the biggest library of stock images in the world, created its own model. They promise to pay the creators whose pictures they trained on and offer a lawsuit guarantee to anxious customers.

Getty Images sued Stability AI earlier this year for scraping its content and training Stable Diffusion on it without permission or payment. A favorable outcome for the company could make it a big player in the AI generated stock image field.

AI Art of the week 🎨

Lex Fridman released an hour-long podcast with Mark Zuckerberg in the Metaverse. It’s an early demonstration of how we might attend Zoom meetings in a couple years.

IMO you can still tell these are computer generated avatars, but the facial gestures were surprisingly well simulated. We are living in the future!

🌊🏄🌊🏄🌊🏄🌊🏄🌊🏄🌊🏄🌊🏄🌊🏄🌊🏄🌊🏄🌊🏄🌊🏄

That’s it folks!

If you liked it, please share this hand-crafted newsletter with a friend and make this writer happy!