#49: 🤖 AI helps Argentina seal World Cup victory, Meta VR drama, AI art controversy, and more

Greetings, fellow humans. This is Not A Bot - the newsletter about AI that was definitely not written by AI. I’m Haroon, founder of AI For Anyone, and I’ll be sharing with you the latest news, tools, and resources from the AI space.

Welcome back!

So…you wanna learn what’s happening in AI, eh?

Well good news - for the 55 thousand of you that were subscribed to the All About AI newsletter, rest assured that the fun will continue. Sure, we have a new look, but the motive is the same: to help you stay plugged into what’s happening in AI.

And btw, there has never been a better time to jump back into the what’s what of AI. It’s hotter than ever. What’s the opposite of an AI winter? AI summer? Well, whatever it is, we’re in it now.

Thanks for joining me on this journey. Now let’s get into it…

⚽ AI in the FIFA World Cup

After 28 exhilarating days, the 2022 FIFA World Cup has finally come to a close.

What. A. Tournament.

Yes, sure, the football — sorry “soccer” fans — was great. In fact, many people are calling the “best World Cup in history” (me…’people’ is me).

But it was also the most AI-driven WC in history.

AI was literally used everywhere in the stadium, from controlling stadium temperature to making on-field decisions, like the controversial onside call that ultimately led to Argentina’s third and final goal.

By justifying the onside call, AI played a tiny role in helping to cement the legacy of, arguably, the GOAT. That’s right. Lionel Messi.

One of the ways in which AI was leveraged during this WC was in tracking on-field calls. For the first time in WC history, a motion sensor placed inside each match ball, tracking precise location data 500 times per second.

This led to some incredibly controversial calls during the tournament, like the one that ultimately decided to uphold Japan’s game-winning goal to knock Germany out of the tournament.

The call was so controversial, that FIFA (the organizing body of the WC) ended up justifying their decision on Twitter by releasing a 3D-rendering of where the ball was exactly when it was played in.

Some say that the charm of the sport is being ruined by the use of technology like video assistant referees (VAR) and ball sensors.

I say, let the technology flourish! Yes, it can lead to controversial calls, but so controversy is inherent to sport. At least now, we can fight with the robots instead of the referees (not that that shielded referees from the line of fire).

Better to make the correct call in the end, than to leave these things to chance. But, hey, that’s just my opinion. 🤷‍♂️

Some other ways that AI was used in this World Cup include:

• An advanced cooling system to “spot cool” only parts of the stadium that contained fans

• 15,000 cameras to track people’s movements (yeah…not sketchy at all 👀)

• An algorithm from the Alan Touring Institute that predicted the winner (spoiler alert: it was wrong)

The next FIFA World Cup in 2026 takes place in North America and I’d bet we’re going to see some mind-blowing applications of AI there.

Silicon Valley, you’re up!

🗞️ Other AI news

• OpenAI has released a new version of it’s embeddings model. The new embeddings model will better facilitate a computer’s ability to grasp the connections between those concepts

  • Explain embeddings to me like I’m five: An embedding is like a special code that helps a computer understand and recognize things. Imagine you have a bunch of pictures of different animals and you want a computer to be able to identify which animal is in each picture. You can give each animal a special code, or "embedding," that the computer can use to recognize which animal is in each picture. For example, you might give the elephant an embedding of "1," the giraffe an embedding of "2," and so on. Then, when the computer sees a picture of an elephant, it will recognize the embedding "1" and know that it is looking at an elephant. Embeddings help computers understand and recognize things just like we do!

• Lensa, the AI art app that’s been generating the AI selfies you’ve been seeing all over your Instagram feed, has been raising concerns over the ethics of AI art. Some artists have even identified the remains of artist signatures on the Lensa portraits, as signs of stolen artwork. Others have spoken out in favor of AI art.

• Tim Ferriss, renowned author of The 4-Hour Workweek and a bunch of other stuff, is launching an AI art competition, where three winners will each win $2K USD. You can learn more on Tim’s blog.

• Virtual reality titan John Carmack has resigned from Meta, and has shared an openly critical letter to his Meta colleagues, and now the rest of the entire Internet. You can read it here.

  • tl;dr: The full post explains how the author had been frustrated with how things got done at Meta, and believed the "derivative of delivered value" was positive in 2021, but had turned negative by 2022. There was a gap between John and Mark Zuckerberg on strategic issues, so John decided to leave and pursue building AGI at Keen Technologies.

🔥 Trending Tools

Here’s a list of my five favorite AI tools & resources of the week

• Huberman AI: Huberman AI uses the latest models from OpenAI to search every episode with Google-level accuracy and summarize the results. Huberman Lab podcast fans rejoice!

• Flikia - turn text into videos. I’ve dug into text-to-video a little bit and can confirm that it’s a very challenging problem space. Excited to see the Flikia team go for it!

• DoNotPay (beta) - a chatbot that will negotiate your Comcast bill for you

• Sendo - AI-generated instant replies to your emails

• Another great prompt engineering guide

🤔 Thought-provoking tweet

Sam Altman, CEO of OpenAI (the organization behind GPT-3, DALL-E, and a bunch of things) points out the double standard of AI critics in this tweet.

And that does it for this week’s issue.

Welcome back, thanks for reading, and see you next time. ✌️

How often would you like to receive an issue of Not A Robot? Let me know in the poll below :)