Gemini 1.5 Was Put Through A Workout With Exercise Videos 

Google announced and released Gemini 1.5 Pro to select developers and enterprises and some of the testing results are coming back.  One of which was putting the upgraded Generative AI to the test with many different use cases.  

Everyday there are fun experiments that people run with GPT, Gemini and other LLMs. Today, this one caught my eye as it shows something I've seen others try and build products around.  Today it was put to the test with analyzing Exercise videos and providing the feedback loop other computer vision apps, wearables and other hybrids have tried to turn into products.  

McKay Wrigley, posted a video on Twitter where he recorded himself lifting weights and then fed each video into Gemini 1.5 Pro and asked it to write JSON for each exercise's name, set count, rep count, weight, and to generate form critiques.  

THE RESULTS - Watch The Video - 4.5 STARS 

Each videos was summarized

• Exercise Name
• Set Count
• Rep Count 
• Weight Used
• Form Critique in written format 

Checkout The Twitter Post and Video Here 

This is just one use case that one quickly was able to dream up.  The idea here was could Gemini provide the same feedback that other applications or a trainer could and then imagine if it was done in real-time.  There are many different ways this can be done and many new ideas people will play with. 

What ideas do you have?

• Upload a video of you playing sports for feedback and analysis 
• Upload a video and ask it to produce my next exercise clip video 
• Upload a video and have it cut and tag your long form into short form videos 
• Upload your podcast audio and have it cut into clips, analyze for insights, etc. 

There are some companies who provide services like this today and these tools are putting the power back into your hand.  You might be on the negative side of this but you can also use it as there is always room to innovate with your business model and many other areas.  

The bottom line here is that you should be experimenting with using these tools to see how it can improve your business, your products and your life!  

The World of Generative AI - A Snap Shot of Gemini 1.5 PRO

1.5 Pro context window capacity is far beyond the original 32,000 tokens for Gemini 1.0.  You can now run up to 1 million tokens in production.  This means that 1.5 Pro can process vast amounts of information in one go, including 1 hour of video, 11 hours of audio, codebases with over 30,000 lines of code and over 700,000 words in documents.  And reports show this can go as high as 10 million tokens.

Bottom line is Gemini is enabling an entirely new set of capabilities and will help developers build much more useful models and applications.  If you want more information on what and read about the MoE architecture that allows this to all happen you can go read their release here.

Mike G. Hansen,

25 Year Health, Fitness and Technology Expert

Contact: www.MikeGHansen.com