Conversation

Replying to

5) twitter.com/evancharles/st Very early in Alameda’s life, Bitcoins on Japanese exchanges were trading 10% higher than BTC on US exchanges. So you could make 10% with the following trade: bank1 → Coinbase → BTC → Bitflyer → JPY → bank2 → USD → bank1 → Coinbase

Quote Tweet

We probably took "do things that don't scale" too far. It was absurd. But there was a major upside to doing so many orders ourselves: we understand the details:

Show this thread

1

15

57

6) The core of the trade, of course, was: buy on Coinbase, sell on Bitflyer. But you wouldn’t make money if that’s how you thought about it. Because as it turned out, most of the difficulty was in the banking. The intersection of banks and crypto has always been a sore spot.

2

28

7) Back then, Alameda had a Bank of America account. And each weekday at 10am two team members would walk to the local branch. They’d be there for hours: waiting in line, reading out confirmation codes, talking to the manager, requesting higher limits, signing forms…

1

3

27

8) Wire transfers closed at 1pm. If the wires weren’t out by then, we didn’t make money that day. And sometimes they weren’t. Sometimes 3 hours sitting in a bank wasn’t enough; so we started showing up at 9am. 8 people-hours spent sitting in a bank, every day.

1

31

9) In some sense, this obviously didn’t scale. (If only there were some way to transfer wealth online and permissionlessly!) But if we’d never done it, we’d never understand how wire transfers worked.

1

1

33

10) And in some senses it did scale. You can do the math: 10% arb; once every weekday; multiply by capital base…. Unfortunately that trade only lasted a few weeks before the bubble burst. But it taught us a lot.

1

2

27

11) About ourselves, and markets, and banks, and companies, and how you get there from here.

1

23

12) twitter.com/evancharles/st Ask pretty much any startup, and they’ll tell you the same thing: you have to go all-in. You can’t half ass it.

Quote Tweet

A few months in, we explicitly decided to reset our lives, move in together, and devote 24/7 to building DoorDash. After YC, we made the same commitment again. A business like this doesn't get built without sacrifice. I can only imagine what it's been like to do this for 6 years.

Show this thread

2

1

32

13) The pure number of things you have to do is staggering. You have to form entities and get bank accounts and build a brand and hire and fire and manage and raise and invest And, of course, you have to actually build the business.

1

31

14) And if you fuck up any one of those too badly, there’s nothing left. You can’t skip management day. Early on, I didn’t know what to do when things went sideways with people. We had a few unhappy employees. And I… well, I kept running into the same walls.

1

25

15) I’d try to talk through our differences, and present evidence, and argue when I thought my approach was right. And often that worked! But unhappy employees come when that doesn’t work. And if it didn’t work the first 10 times, it probably won’t work the 11th.

11:37 PM · Dec 12, 2020Twitter Web App

Replying to

16) I was out of ideas. And things festered. Festering is the right analogy. It spread. Eventually it infected half the company. Had I acted sooner, and more thoughtfully, I probably could have fixed it. But by the time I did act, amputation was the only option left.

1

21

17) Over the last few years, there are a number of times we’ve faced a decision. And often, that decision is: do we want to go out on a limb? When we founded FTX, we had no idea how to get customers. We didn’t know if we would ever know how.

1

17

18) The thing that convinced us to try it was: Value of FTX = (probability we get customers) * (probability we build a good product) * (value of top exchange)

1

24

19) And so we wrote down some numbers: 20% * 75% * a lot When we first started FTX, we thought it would probably fail. We did it anyway, because 15% of a lot is… a lot.

1

2

32

20) In the end, we got there, because we tried hard, and learned what we could, and put ourselves in the best position we could to succeed. And because 15% isn’t actually all that low.

1

25

21) With Serum, there was something similar: will an ecosystem actually develop here? projectserum.com/tokens

1

26

22) Ok, so those are a few anecdotes. But how about more generally? Almost all startups fail. Is it really correct to think big, even if you’re risk neutral?

2

16

23) And this, really, is the heart of the matter. 15% is a lot different than 0.15%. What can you do to have a 15% chance of success?

1

1

22

24) I’ve known a lot of smart people. Many have gone on to do good things with their lives. Some have gone on to do great things. They were bright, sure; but that wasn’t the only thing they shared. They also all made the same choice.

1

2

24

25) They all chose upside over certainty. There are a number of people in the crypto community that I respect. But there’s no community I have more respect for than the effective altruism (EA) community.

2

6

69

26) And the most impressive parts of the EA community are the ones full of people who chose, very intentionally, to end up there. nytimes.com/2007/12/20/us/ Together they’re discovering big problems; scoping out approaches; and generating hundreds of millions of dollars for them.

2 Young Hedge-Fund Veterans Stir Up the World of Philanthropy (Published 2007)

Holden Karnofsky and Elie Hassenfeld rank charities by analyzing the numbers in much the same way they did at their investment management company.

3

1

39

27) They’re impressive people funneling their lives into doing as much good as they can. The amount of good each person can do is enormous, and once they chose to maximize, the rest of their stories aren’t actually all that improbable.

2

3

33

28) And we’ve seen the same thing at FTX: the people who think the biggest and try the hardest to get there and give the most to it often do amazing things.

1

4

31

29) Because the truth is, most people just don’t try that hard; maybe it doesn’t make sense to, if you don’t care much about the tail cases.

1

2

38

30) And so if you do push yourself--you do motivate yourself to do everything you can, and you do think strategically, and you do identify massive opportunities--maybe the odds of success aren’t all that low.

Quote Tweet

Linear wealth utility functions (conveniently driven by effective altruism!) are a form of edge in markets where most are log... twitter.com/SBF_FTX/status…

2

4

43

31) And so that brings us back to the Bank of America lobby. It was a detail, a mundane task that anyone could have done, theoretically. But in fact, very few people did it. Most of our competitors didn’t.

Quote Tweet

One of our most memorable lessons from YC was "do all the things." We came with a list of 20 ideas for how to grow, and asked the YC partners which to prioritize. I think it was @paultoo who said something like "How would I know? Do all the things."

Show this thread

1

3

36

32) And you might think it was because we thought to hire a low-level employee to sit in a bank all day. But as it turned out, it was our COO. Maybe it “shouldn’t” have been him, I don’t know. What I do know is that if you think that way, instead it’ll be no one.

4

3

49

33) Sometimes the only thing standing between what is and what could be is the will to get there, whatever it requires.

8

8

107