Hacker News Reader

Comment by ein0p | Hacker News Reader

Google was a powerful place to work at 10 years ago. At Microsoft you had to convince people to give you quota for a few cores and wait for days or even weeks, if they even agree at all. At Google you could spin up a distributed batch job processing terabytes of data during your orientation codelab, literally on the first day. The contrast was very stark. Google was a very low friction environment where things just happened, and the environment was set up such that you couldn’t help but ship something. Note the past tense.

View on HN

Internal research had low friction external product not so much. They even had a tool, ariane, to help you navigate the product launch friction.

IMO Ariane was pretty straightforward and reasonable, all things considered. In fact it’s preferable to navigating the usual wobbly maze of humans each of which takes days to respond.

That's in contrast to what OpenAI's David Luan "Why Google couldn’t make GPT-3" (https://www.latent.space/p/adept):

  And it turned out the whole time that they just couldn't get critical mass.
  So during my year where I led the Google LM effort and I was one of the
  brain leads, you know, it became really clear why. At the time, there was a
  thing called the Brain Credit Marketplace. Everyone's assigned a credit. So
  if you have a credit, you get to buy end chips according to supply and
  demand. So if you want to go do a giant job, you had to convince like 19 or
  20 of your colleagues not to do work. And if that's how it works, it's
  really hard to get that bottom up critical mass to go scale these things.
  And the team at Google were fighting valiantly, but we were able to beat
  them simply because we took big swings and we focused.”

The whole episode is very interesting.

Yes. Hence the past tense. There are more reasons why they fell so badly behind, from bureaucracy to eye wateringly decrepit, overengineered, and legacy ridden training code, to slow and hard to debug AI infra, to deliberate forking and siloing of critical projects, etc. Let’s just say Google is now very far from where it was a decade ago. I’m mildly surprised they released anything competitive at all. I’m not surprised that they failed to beat OpenAI (and therefore ironically Microsoft - how do you like them turntables?)

thanks for sharing it!

This video [0] was from 2010s, and seems like they had similar issues even back then. (never worked at Google)

[0] https://www.youtube.com/watch?v=3t6L-FlfeaI

Reply on news.ycombinator.com