Happy Monday and Welcome Back!

Here’s what we got today:

• What’s Zuck been working on? 🤔

• GPT-4 🤝 U.S. Medical Exam

• Check out Ameca 🫣

What’s Zuck been working on? 🤔

Zuck and some of his boys are taking a step out of the Metaverse to work on AI.

Well, sort of.

Andrew Bosworth, CTO of Meta, recently had an interview with Nikkei Asia about the artificial intelligence work going on behind the scenes at the tech giant.

According to the interview, top executives such as CEO Mark Zuckerberg and CPO Chris Cox are spending most of their time working on artificial intelligence.

“We just created a new team, the generative AI team, a couple of months ago; they are very busy. It's probably the area that I'm spending the most time [in], as well as Mark Zuckerberg and Chris Cox.”

Bosworth said some of the work would benefit the Metaverse. “So previously, if I wanted to create a 3D world, I needed to learn a lot of computer graphics and programming. In the future, you might be able to just describe the world you want to create and have the large language model generate that world for you. And so it makes things like content creation much more accessible to more people."

Meta is expected to debut commercial applications using AI, particularly for the company’s ad business this year. 👀

Check out the full interview here.

GPT-4 🤝 U.S. Medical Exam

Remember when GPT-3 passed the U.S. Medical Exam? If not check it out here.

This time GPT-4 is being put to the test.

Dr. Isaac Kohane, a physician and computer scientist at Harvard, teamed up with colleagues to test the limits of GPT-4 in a medical setting in their forthcoming book, “The AI Revolution in Medicine”

In the book, Kohane says GPT-4 answered 90% of questions correctly on the US medical licensing exam. The book also touches on GPT-4’s ability to translate, a particular example involves the chatbot translating discharge information for a Portuguese-speaking patient.

How well does GPT-4 diagnose conditions? The book touched on a clinical thought experiment, based on a real-life diagnosis from a few years ago. The chatbot was given a few key details from a physical exam, ultrasound, and hormone levels. GPT-4 was able to correctly diagnose a 1 in 100,000 condition called congenital adrenal hyperplasia. 🤯

Bottom Line: We don’t see AI taking a physician’s job anytime soon, however, it looks to be a powerful tool in medical settings. Certain issues involving reliability and ethics need to be addressed, however, we’re still betting big on AI in medicine.

Check out Ameca 🫣

Ex-Machina is looking more realistic everyday.

Ameca, the viral robot made by Engineered Arts has been making the rounds recently, after a video of it speaking multiple languages was released.

Do you think the human face is really that necessary? 😳

How does it work?

• GPT-3 is used for conversation and translation

• DeepL is used for language detection

• Amazon Polly Neural is used for voices

📝📝📝

According to Engineered Arts, “The main purpose of Ameca is to be a platform for developing AI. We love designing and building robots, we'll leave it to you and all the other amazing (naturally intelligent) brains out there to create the AI and machine learning alogrithms and see how far we can progress the technology together.”

The robot is built on Engineered Art’s Tritium software platform, which will be released as a public beta soon.

Check out Ameca here.

That’s it for today, we’ll see you back here tomorrow!