6 Comments
User's avatar
Jon Qian's avatar

Why push a researcher so hard on the exact timing? Feels like we want to cool the bubble or something. Is that the right use of both your time? Isn’t what are the remaining research problems a better conversation to have if you want to gauge out progress and challenges

Expand full comment
Max Conradt's avatar

I wish HRT published the order book dataset as a challenge like the Jane Street one.

The flow matching technique Sergey described for modeling continuous actions is also relevant in deep learning for trading, so it was a fun coincidence to see the HRT bit shortly after.

Expand full comment
Immanuel Giulea's avatar

This was an amazing interview!

Expand full comment
Abikaasa LENSE's avatar

Yeah 👍🏼

Expand full comment
Advait Patel's avatar

“what does it mean for a robot to be out there?”

“It means that there is a robot that is actually doing a thing competently enough that other people want done”

check out Dyna robotics. We are already deploying foundation models onto our robots into the real world, for real customers, doing valuable tasks.

Expand full comment
Harry Buck's avatar

Sergey's work was an inspiration to the robots in my book. And he was one of the amazing "eastern European AI mafia" that I rolled into my lead character, Stash Novak.

Super excited to watch this one!

Expand full comment