Conversation

8) c) A US business wants to send 100m to a European business. So you initiate a wire transfer. A day later, your bank sends it out. Over the next week, 3 other banks touch that wire transfer, and sometimes it gets stuck in the middle, waiting for you to rescue it somehow.

566

9) Oh, and you pay 1% in FX fees, plus $50 in wire transfer fees. d) You are living paycheck to paycheck, and get an overdraft fee. You are surprised, and investigate why. It turns out your paycheck takes 4 days to deposit, your cash won't hit until the next business day.

557

10) Meanwhile your credit card just batched together a few weeks of expenses and deducted them from your account, and your landlord finally cashed a check from a few weeks ago. So all your deposits are pending and all your withdrawals just hit.

477

11) The moral of the story is: payments are hard. There are many reasons for this, but they all circle around the same core problem: What, exactly, does it mean to send someone money? What does it mean for that money to "settle"?

681

12) ACH and credit card transfers can take a day to a month to take effect. And then for the next month, they can, maybe, be recalled. (Even if there's no reason!) There's no way to send USD to a non-US bank unless they are whitelabeling one; there are tons of intermediaries.

460

13) So, how does blockchain help this? Well, for every one of these scenarios, you could replace them with: a) Bob creates a wallet b) Alice creates a wallet c) Bob sends the assets to Alice's address

630

Account JAEcyoWQVj1EctpTqqfyGRSrSXQFAUk2KBNa1n5TdMnu

14) Just to demonstrate this: I initialized Bob with $99.75 and some SOL (from FTX) (solscan.io/account/CHEr3k). I then created a second brand new wallet, solscan.io/account/JAEcyo ('Alice')

solscan.io

View transactions, balances and other info of JAEcyoWQVj1EctpTqqfyGRSrSXQFAUk2KBNa1n5TdMnu

438

Solana transaction details | Solscan

15) So then I went to send $50 from Bob to Alice. I clicked 'send' at 8:19:33 am. By 8:19:45, when I tabbed over to Alice, the $50 had already landed. The fee I paid was $0.0002: around 2% of a penny.

solscan.io

Solana detailed transaction info for signature 4oeqVbHWxXyCk1xBE4VDGenAcqjz6sW4tNBJwHQBt14XG4nj2KZ9HfBR7zJ76xP9P5cwmH4SGPbMHwcA2vdDocDw

716

16) Blockchains allow anyone to create a wallet and use it to send and receive tokens--including USD pegged stablecoins. Those payments take seconds to process, cost fractions of a penny, and are finalized in less than a minute. No long wait, no account balance uncertainty.

758

1:04 PM · Jul 16, 2022Twitter Web App

17) Ok, so how about market structure? On January 28th 2021, most major retail brokers shut down. Users were unable to buy; sometimes they were unable to sell, too. And on some platforms users got liquidated. The weird thing is that there was basically no leverage!

Replying to

18) This was, of course, the day that Gamestop, AMC, and other memestocks skyrocketed. Why were users shut off? It wasn't because of leverage: users tried to deposit $50 to buy $25 of stock, and were denied. And it wasn't because stocks were "out of stock" or anything.

426

19) Instead, the system broke down because of settlement risk. See, when the typical retail trader tries to buy a share of AAPL, they don't send an order to a stock exchange. Instead, they send an order to a stock _broker_, like Robinhood or Schwab. (Or, now, FTX US!)

526

20) The dollars they send in are routed to the broker's bank account. Their order is in turn routed to a PFOF firm, like Citadel or Virtu. That firm, in turn, might buy the stock on an ATS from another trading firm, which would finally bid on NASDAQ, or NYSE.

449

21) And then, two days later (!!!), comes settlement time. The share of AAPL is settled through the DTCC via something like: NASDAQ --> PFOF#2's clearing firm --> ATS --> PFOF#1's clearing firm --> broker's clearing firm, with the DTCC in between every transfer.

481

22) Meanwhile, the dollars used to purchase the stock flow through banks. So in order for your friend to buy one share of AAPL, 11 different entities have to settle with each other over the next few days in 10 payments. And in theory, _any one of those settlements could fail_.

510

23) And on January 28th 2021, there was a lot of retail trading volume. Which meant tens (hundreds?) of billions of dollars of pending settlements between tens of counterparties which would take days. As GME rose in price, so did the potential loss if settlement failed.

426

24) And eventually that risk got too big for the brokers, which had to partially shut down. How does crypto help this? Well, first of all, with a full stack product, settlement is way simpler. Here's what it takes to buy and settle on ftx.com/trade/BTC/USD:

403

25) On FTX, everyone can send orders straight to the exchange; in equities, only the wealthiest traders can. Which means you can bypass all of the settlement risk from intermediaries. But there's another reason that blockchain helps.

538

26) Say you tokenize stocks. Instead of waiting 2 days to settle, you can just swap AAPL-token <> USD-token on a blockchain. Which, remember, takes about 10 seconds and costs about $0.0002 in fees. No remaining settlement uncertainty or risk.

150

966

Risk Management: Avoiding the Next “LME Nickel” Market Incident

27) So blockchain can create simpler, more equitable, and less risky market structure and settlement. It can help us avoid problems like Gamestop day. And, for that matter, LME Nickel: ftxpolicy.com/posts/risk-man.

ftxpolicy.com

In early March, a Chinese investor with a short position on nickel through derivatives suffered losses when the cost of nickel began to increase substantially. Along with other commodities, nickel...

656

28) So how about social media? Right now, if you tweet something, and your friend pulls out Facebook, they can’t see your tweet. Social media networks are isolated, not interoperable.

564

29) This means that everyone has to manage 10 different apps, and our networks and conversations are fractured. In addition, all of your network, and friends, and data, are tied to a specific platform.

457

30) The largest networks have pseudo-monopolies, with gigantic network effects that prevent new competitors. And how does censorship work? Well, basically, the leaders of the top few social networks have to make unilateral decisions.

479

31) Let’s say that, instead, we put messages on a blockchain. So if you used Blockchain-Twitter (BT): --You type the message in BT’s interface --BT posts the message on a public blockchain --Your friend pulls out Blockchain-Facebook (BF) --BF reads your message and displays it

632

32) By using an underlying public chain for messages, we’ve made different networks compatible. You can use any platform, and still talk to all your friends on every platform. Your messages, and network, are yours: you can move platforms and keep them.

407

33) New platforms can read/write from the chain, inheriting the network effects, allowing real competition. And each platform can make moderation decisions, so at least there’s diversity of opinions.

328

34) I could see blockchain having a large impact on payments, remittances, market structure, and social media. And this list isn’t exhaustive; there are tons of areas blockchain can innovate in. I haven’t really touched at all on DeFi, or web3 gaming.

499

35) But taking a step back: how many of these areas has crypto revolutionized so far? I think the answer is “not really any of them”. It’s starting to impact some, but not in a widespread way yet. So what could get us there from here?

427

36) (i) Technology. Right now the fastest chains can handle 5-50k TPS. We probably need to scale that up to ~1m TPS, whether it’s through L1s, L2s, lightning, rollups, etc. I’m guessing we’ll get there in ~3-7 years.

572

37) (ii) Regulation. There needs to be regulatory clarity in order to get mass adoption, and the major platforms have to become licensed. I’m really impressed by the progress the US has made in the last year!

468

38) (iii) Great products. Builders have to focus on building great products, not on monetization as the core product.

550

39) (iv) Network effects. Many of these use cases don’t work unless two people both use crypto: buyer/seller; sender/receiver; tweeter/reader. That means we have to get up to ~30% adoption by person for (30%^2 =10%) adoption by transaction.

534