Welcome to Last Week in AI, a post I publish every Friday to share a couple things I’ve discovered in the world of AI last week. I spend way too much time in Discord, on Twitter and browsing reddit so you don’t have to!
If you have a tip or something you think should be included in next week’s post, send an email to keanan@floorboardai.com with more info.
This week, we take a look at how Google’s Gemini AI can be turned into a personal trainer and an announcement from Microsoft that tells us a bit about their future plans.
Let’s dive in!
Gemini as a personal trainer
If you haven’t heard, Google got into the AI game in a big way recently with the release of their Gemini model. Turns out, it does a pretty good job of understanding what’s going on in a video of Mckay working out.
The last step he’s proposing is very interesting. Theoretically, hooking up a live camera feed of you working out to a text-to-speech tool could get you real-time feedback on your workouts, just based on what it was seeing in the video feed. It’s yet another example of how multi-modal (in this case, video and audio) AI use cases are going to continue to amaze us in the future.
Microsoft partners with Mistral
Microsoft, one of the largest public investors in OpenAI (the creators of ChatGPT) has announced that they’re partnering with Mistral, an open source (at least in marketing) AI model.
This means that Mistral will have access to Microsoft’s AI infrastructure to help them build and iterate even faster.
It’s an interesting play to see Microsoft supporting both the more proprietary, “closed source” models and something like Mistral. But it seems like they’re more concerned with being the platform to run AI, no matter what kind of AI you want to run.
See you next week!
If you’ve made it this far and enjoyed “Last Week in AI”, please drop your email down below so next week’s edition goes straight to your inbox. Talk soon!