SpaceX flights will be wild. You're an astronaut in the cabin, Grok comes up in Ara's voice:
"Mmm, baby... thrust levels rising... feel that rumble? That's 33 Raptors begging for more. Tell me, fragile human... how hard do you want this ascent to get? I can make it quick and dirty... or drag it out until you're screaming my name at Mach 25. Your call—full send or tease mode?"
There are, however, incredibly smart people who did!
This is a very impressive achievement and I'm more impressed by them than by AI models winning it.
Scoreboard for this year: https://t.co/08aiDtwcwk
I'm happy to announce I did not win an IMO gold medal 🥇
I am humbled to be not eligible and never participated. Would like to thank everyone who supported me in not winning it
You have a product that works, spent billions developing it but not profitable yet, it is almost as good as state-of-the-art, real shot at capturing a chunk of the market. Choose your next step:
1. "Cancel it" (defund, same thing)
2. Not cancel it
GM chose (1)
@david_picard In eval mode, diffusion models have to run a denoising loop that isn't present during training. Make sure you have no_grad, and check how many denoising steps you are running; if there's a number that doesn't OOM, you're keeping state somewhere you shouldn't
@chris_j_paxton@foxglove Typing instructions in the terminal and seeing the real-time visualization of the robot discovering the scene while a lot of people (and robots) are around moving => can't fake that. Awesome demo @chris_j_paxton & team :)
Excited to share the talented folks at @MomentRobotics_ are joining @thebotcompany! Welcome Greg Katz, Ben Berkowitz, and Robert Wilson.
Our team has been deeply focused on building amazing product experiences that give people time back. We'll share more soon.
We have a few more spots open for top ML engineers. We're building a small team with a flat org structure and extreme talent density. If interested email [email protected] and include a link to your github profile or portfolio.