Elon Musk's Grok 1.5 vs GPT-4's OpenAI

Let's compare Grok 1.5 Vision Preview and GPT-4’s OpenAI…
Compare their abilities to write code based on a diagram, calculate calories and explain a dumb meme. Can it put us to sleep (in a good way) with a creative bedtime story?! Watch GPT-4 take approximately 43 days to translate an image to CSV, but ultimately figures with a little help.
Grok and GPT-4 are exciting and skilled contenders in the AI landscape in their own ways, but ultimately GPT-4 wins due to it’s overall performance.
Timeline:
00:25 - Grok writes code from diagram
00:50 - GPT-4 writes code from diagram, joy ensues
01:50 - Calorie counting
02:30 - Create a bedtime story
03:46 - Explain a meme
05:10 - Convert table to CSV
07:54 - Cheat question!
09:06 - Which object is larger?
09:43 - Driving a car
10:08 - Space for car to drive?
10:25 - Identify dino direction
11:43 - THE FINAL WINNER
(Shockingly GPT-4!? Although I had some hope for Grok...)
#XAI #VisionPro #ElonMusk #coding #chatgpt #claude3 #geminiai

[@Mazz1983]

I'm very glad I watched the whole thing cause I could make up my own mind. These judgements are so flawed imo. I don't think your value system here matches practical applications in the real world - which is what your examples are. So it made me a bit restless in my seat seeing these points get assigned.

[@joshsowin]

How would you rank them differently?

[@Mazz1983]

@@joshsowin ideally I would weigh a lot more towards usability. That table data not having capitalised is insignificant compared to missing whole stats. Decisions are made upon data. More data mean better decisions. Missing even 1 datum makes it lose automatically. You tied them.

[@joshsowin]

@@Mazz1983yes I tied them because grok also lost data - lowercase and uppercase are completely different characters. The numbers are also lost data, probably because it’s seen as an emoji not a number. So they both didn’t do a perfect job. I could also be convinced they both lose instead of tie 😅 but anyway these are just my opinions!

[@Mazz1983]

@@joshsowin maybe I should have watched on a bigger screen!