korrents

korrents · The Pragmatic Engineer Podcast

Scaling Uber with Thuan Pham (Uber's first CTO)

Thuan Pham · 1h 39m · youtube.com ↗

24 korrents from this recording

1h
Thuan Pham did not write this page.

Every claim below is a statement made in this recording, quoted word for word and linked to the second it was said, so you can hear it rather than take our word for it. The wording comes from the transcript published alongside the recording; the sentence above each quote is our reading of the claim, not their wording.

  1. 0:11:21 · watch on youtube.com↗

    Our reading

    A technology being right is never enough on its own: whether it wins turns on whether the world is ready for it and whether the economics work at that moment.

    Their wordsIt's not about just a technology. It's about whether the world is ready for it, whether it's economically feasible.
    Interesting
  2. 5 min later

  3. 0:16:45 · watch on youtube.com↗

    Our reading

    Growth without profit is not yet a company: fast growth only makes a business durable once it has been converted into sustained profit.

    Their wordsThere has to be sound business model that makes sustained profit, right? Growth and profit. Growth alone eventually burn, you know, money and that's not good.
    Interesting
  4. 1 min later

  5. 0:17:41 · watch on youtube.com↗

    Our reading

    Today's AI boom will end the way the dot-com boom did — consolidation, a wave of failures and a few genuinely great companies — and customers, not argument, will decide which is which.

    Their wordsThere may be a lot of AI things that's going on right now, right? Uh eventually some of these things will consolidate, some will go under, some will become really awesome solutions and all that stuff. And so, but the the market will sort it out.
    Interesting
  6. 9 min later

  7. 0:26:48 · watch on youtube.com↗

    Our reading

    A career is not built by searching for jobs but by accumulating a reputation slowly, so that when you become available the opportunities come to you.

    Their wordsif you try to do a really good job at every company you've been working well with all the people that you work with, including your own team, your peer, whatever it is. Over time, very slowly, you accumulate a decent reputation in people's mind.
    Interesting
  8. 10 min later

  9. 0:36:58 · watch on youtube.com↗

    Our reading

    A technical leader should let engineers find the flaw in their own design through leading questions rather than announce it, because a discovered flaw is acted on and a reported one is argued with.

    Their wordspretty soon by asking those leading question, the engineer now discovered a flaw that this thing would not scale,
    Interesting
  10. 1 min later

  11. 0:37:48 · watch on youtube.com↗

    Our reading

    The right way to specify a rescue rewrite is a couple of scaling constraints and no new features at all, so that hardware can absorb the growth the team has no time to design for.

    Their wordsNo new feature necessary. Just make sure that we can do that. And then that allow the business the company to just pour a whole bunch of hardware behind that and it will scale.
    Interesting
  12. 2 min later

  13. 0:39:18 · watch on youtube.com↗

    Our reading

    How big a system will eventually have to be is not worth speculating about; the only useful question is how long the one you have has left before it hits the wall.

    Their wordsthe question about how big it can possibly grow nobody knows really. But it's actually not fruitful to pontificate on that.
    Interesting
  14. 1 min later

  15. 0:40:46 · watch on youtube.com↗

    Our reading

    Repeated rewrites are the correct answer under fast growth, because a system specified to scale forever would take so long to build that the company would die before it shipped.

    Their wordsLet's say if my requirement for the engineer was build a system that will scale infinitely to the test of time, it might take a year. We never get there. We'll die before then.
    Interesting
  16. 4 min later

  17. 0:45:00 · watch on youtube.com↗

    Our reading

    Launching the hardest case first beats the cautious order of smallest-to-largest, because after the hardest one everything that follows is downhill.

    Their wordsthat was the most brilliant thing because by doing the hardest thing first once you launch that everything else is downhill from there
    Interesting
  18. 3 min later

  19. 0:48:14 · watch on youtube.com↗

    Our reading

    A functionally organised engineering team seizes up at around a hundred people, because every feature then has to be negotiated against every other team's bandwidth.

    Their wordsevery feature that we want to put out has to be queued up on mobile development bandwidth dispatch bandwidth and it become impossible to navigate tradeoff because every feature you want to do you have to go negotiate with so many team
    Interesting
  20. 2 min later

  21. 0:50:30 · watch on youtube.com↗

    Our reading

    Uber's thousands of microservices were not an architectural preference but a survival decision: under that growth only independent, unblocked teams could build fast enough to stay alive.

    Their wordsnone of us wanted to go through that extreme but lots of time when you are under a lot of pressure and no time to react other than just to survive that scale that keep on coming at you, you have to make uh decision that increase uh speed and velocity because speed and velocity allow us to build quick enough to survive
    Interesting
  22. 1 min later

  23. 0:51:23 · watch on youtube.com↗

    Our reading

    Breaking up a monolith takes years rather than months not because the work is hard but because the business keeps adding to it faster than you can pull pieces out.

    Their wordshad we freeze time, that piece of code could be decomposed in a matter of 3 to 6 months. But it took us 2 years to do that because as we peel out a piece of code, the business keep on going forward, right?
    Interesting

  25. 0:51:45 · watch on youtube.com↗

    Our reading

    The organising rule under extreme growth is that no team may block another — even when honouring it means adding to the very monolith you are trying to dismantle.

    Their wordsthe philosophy we all operate at the time was no one should be blocking anybody else. No one can block anybody else.
    Interesting
  26. 4 min later

  27. 0:55:22 · watch on youtube.com↗

    Our reading

    Depending on open source you cannot fix is the frightening kind of outage, because at your own scale there may be nobody in the world you can pay for an answer.

    Their wordsI don't mind if we think we can do something of our own problem it's terrifying when we have a major problem and we depend on somebody else
    Interesting
  28. 7 min later

  29. 1:02:36 · watch on youtube.com↗

    Our reading

    A company's staff-engineer bar should be set against the best companies in the industry rather than against its own history, which is what makes title inflation a real cost.

    Their wordswe have a very clear definition and expectation of what it is at the staff engineer level because we benchmark ourselves to all the great company out there Google, Facebook and all that
    Interesting
  30. 2 min later

  31. 1:04:19 · watch on youtube.com↗

    Our reading

    A company should manufacture fresh internal challenges for its best people, because the alternative is that they go and find those challenges at another company.

    Their wordsI look at um the talent base and I think it is best for us to create opportunity for people to keep on growing with fresh new challenges within the company because if we don't do that they would leave the company
    Interesting
  32. 7 min later

  33. 1:11:45 · watch on youtube.com↗

    Our reading

    Engineering offices abroad should be opened where world-class talent already lives, not where labour is cheap — you bring the work to the talent rather than moving the talent.

    Their wordsWe didn't go to other places because of cost savings like that. We go there because we have need and we have world class talent
    Interesting
  34. 15 min later

  35. 1:27:08 · watch on youtube.com↗

    Our reading

    Directing a swarm of coding agents is a different, multi-threaded way of thinking that costs more cognitive load, and the engineers who make the switch have doubled their output.

    Their wordsit required a different way of thinking and the cognitive load might be a little higher but the output is dramatic and we have seen our best engineer double their output.
    Interesting
  36. 1 min later

  37. 1:27:41 · watch on youtube.com↗

    Our reading

    AI has already made sweeping mechanical changes cheap; the unsolved problem is getting the same productivity when adding features to an old, entangled codebase.

    Their wordswe're trying to crack the next frontier which is how we get that level of productivity increase and output building new features on top of a code base that are older,
    Interesting
  38. 2 min later

  39. 1:29:24 · watch on youtube.com↗

    Our reading

    AI tooling has not compressed the gap between great and average engineers: it is still two or three times, and for the same reasons as before — curiosity and appetite for the bleeding edge.

    Their wordswe see the difference between the great engineer and an an average engineer is still two 3x in terms of their capability. They're more inquisitive. they're at the bleeding edge more, they're more innovative,
    Interesting
  40. 1 min later

  41. 1:30:29 · watch on youtube.com↗

    Our reading

    Standing still is now falling behind: as the world speeds up, complacency is the one professional habit that is fatal.

    Their wordsComplacency is death. I mean like every the world will move faster and faster and the moment we stand still we are falling behind.
    Interesting
  42. 1 min later

  43. 1:31:43 · watch on youtube.com↗

    Our reading

    Talent density is self-protecting: once a team is made of A players it will only hire more of them and will not tolerate anyone who is not performing.

    Their wordswhen you have team A, team A will just want to hire more A level players and yeah, they just intolerant of anybody who's not performing,
    Interesting
  44. 1 min later

  45. 1:33:12 · watch on youtube.com↗

    Our reading

    At every level of management your job is to see further ahead than the people who report to you, because they are busy with the near term and nobody else will do it.

    Their wordsyour job is to see a little bit further out than your folks, right? Because your folks are busy working on the near-term things,
    Interesting
  46. 3 min later

  47. 1:35:59 · watch on youtube.com↗

    Our reading

    A company that stops hiring new graduates because AI makes juniors look unnecessary will have no senior engineers in a few years' time.

    Their wordsthe best of the best still get offer from us because if we don't hire those folks today, what senior engineer will we have for yeah years from now, right? You have to feed the talent pipeline
    Interesting