Why are people always speedrunning learning all of ML or 100 days of writing kernels why is there no speedrunning learning art or 100 days of socializing with your peers or blog posts on contributing to society
You are lucky, we’ve had this one ticket open with Azure for almost 4 months now that they continue to not be able to figure out but every time I follow up they hit me with the “our team is investigating” and then come back with irrelevant shit
I actually did this during an interview and explained how to automate what they wanted to hire for and the CTO loved it and asked me to explain to their analyst then they didn’t hire me
There are guys w <100 followers who competitively but openly write the worlds fastest gpt training for fun that you can just send offers to but instead you are reading resumes and tweeting about how it is “incredibly difficult to find top talent”
Unironically a couple random conversations I had with @a1zhang and @m_sirovatka in the popcorn channel of GPU MODE last year on synthetic kernel data sent me down a trajectory of doing increasingly cool things and finally joining PI
I spoke with world class researcher and KernelGuy @m_sirovatka how he wrote such a beautiful environment and his response blew my mind
He told me, he read Verifiers
That’s right, he just read the code
Who would have EVER thought to do that??
So cracked, so alpha
Imagine
All the new people following me now bc PI and now over analyzing my tweets are in for a rude awakening when they find out I just be saying shit on the TL
Why are people always speedrunning learning all of ML or 100 days of writing kernels why is there no speedrunning learning art or 100 days of socializing with your peers or blog posts on contributing to society
not to take away from spikes spotlight but i merged this pr in a waymo on the way to the dataologyxarceexprime event and i think that deserves something