There are a lot of components of Blast that I’m excited about and would be interested in engaging with people on. That said, we at Paradigm think the announcement this week crossed lines in both messaging and execution. For example, we don’t agree with the decision to launch the…
I know moderation at scale is very hard, but you would think that after 13 years they could implement basic heuristics like “don’t let verified accounts change their name and profile picture to match a different verified account”
Uniswap v3 is now public, after eighteen months of research and development:
I think it’s the best DEX design on the planet.
If you've heard me cryptically address a problem by saying that "Uniswap v3 fixes this," now I can finally explain why.
1/
Someone accidentally locked up some tokens in an Ethereum smart contract.
@gakonst
and I thought we'd found a way to recover them.
We learned that the mempool is a very creepy place.
If your project is using token incentives, you should probably try removing them and see what happens.
If your response is "but we're worried users wouldn't stick around"... then you should DEFINITELY try removing them.
The Uniswap v4 whitepaper is now public:
Uniswap v4 continues the evolution of Uniswap into a flexible platform for DEX innovation, with its support for permissionless deployment of new pools that use custom “hooks.”
1/5
If you work with Ethereum data, this tool might change your life:
You can enter any contract address, and it looks up the ABI and generates queries that can be run on BigQuery to create parsed tables for each event or function call on that contract:
First page of every ZKP introduction: “Imagine you have a Where’s Waldo book and want to prove you know where Waldo is with a piece of cardboard…”
Second page of every ZKP introduction: “If we replace FRI with pairing-based Kate commitments, recursive composition becomes…”
I don't think there's much alpha left in designing new AMM invariants
The next generation of DEX features are going to be about fair execution and tx cost minimization, not new shapes for reserves curves
Introducing Blend!
@blur_io
wanted a lending protocol with:
* Arbitrary collateral, including NFTs
* No oracles
* No expiries
* Market-set interest rates
So
@transmissions11
and I worked with them to design a new mechanism
Here’s how it works 🧵
1/ Introducing Blend: the Peer-to-Peer Perpetual Lending Protocol for NFTs.
Built in collaboration with
@danrobinson
and
@transmissions11
at
@paradigm
, Blend enables 10x higher yield opportunities than current DeFi protocols and unlocks greater liquidity for NFTs.
Here’s how 👇
Uniswap is arguably becoming the primary market for ETH
New research from
@gordonliao
and me shows that Uniswap v3 has deeper liquidity in ETH/USD and ETH/BTC than leading exchanges like Coinbase and Binance
Uniswap has been fighting for the core principles of crypto since they launched. To me, they stand for the idea that immutable peer-to-peer financial technology is not only possible, but inevitable. That it can win.
It’s unfortunate that the SEC is forcing them to take that…
Today
@Uniswap
Labs received a Wells notice from the SEC.
I’m not surprised. Just annoyed, disappointed, and ready to fight.
I am confident that the products we offer are legal and that our work is on the right side of history. But it’s been clear for a while that rather than…
This thread is the purest instance I’ve seen of a very common misunderstanding of how AMMs work.
It’s just wrong. The math in it is wrong, the mechanics are wrong, and as a result, the conclusion is wrong.
0/ Uniswap only works because Ethereum is slow, *and* because Uniswap LP fees are expensive (30 bips)
Normally, when you combine two bad features, you get something extra bad
But in this case, negative * negative = (almost) positive
How can this be?
A thread
Matt and Fred took a chance on hiring me when the idea of embedding a protocol researcher at an investment firm was a weird experiment.
Now we’re doubling down on the mission of building an industrial research organization within
@paradigm
.
1/
The dirty secret of crypto Twitter is that you can get like 1000 RTs for a "technical" explanation that is total bullshit
Just remember to end your first tweet with "buckle up for a deep dive 🤿🧵👇 1/"
Anyone who says that blockchain will protect us from naked shorting or complex derivatives or unsustainable leverage has not looked very closely at where DeFi is heading
Suppose you have a protocol that lets you split 1 USDC into two tokens, 1 of token A and 1 of token B.
Anyone can redeem 1 A + 1 B to get 1 USDC, at any time.
What price will these tokens trade at?
If you start incentivizing usage before you've found product-market fit, you hopelessly scramble the signals you need to find it
You might decide it's worth it, but know that you'll be flying blind
Most books should be papers
Most papers should be posts
Most posts should be threads
Most threads should be tweets
Most tweets should be replies
Most replies should be thoughts
Uniswap v3 flipped SushiSwap on volume in less than 4 days, but that isn’t the important part.
The important part is that it did it with less than 10% of the TVL.
Stages of Twitter:
<10k: earnest tweeting about your interests
10k-100k: optimal shitposting zone
100k-1m: can’t tweet anything other than vague philosophy
1m-10m: can’t even tweet the vague philosophy anymore. nobody at this level has ever tweeted
10m+: back to shitposting
Three hundred years from now, Ethereum haters will still be talking about the “premine” and “DAO fork,” with the meanings of those words having been lost to time
Conference tickets are a joke
In almost all cases the conference is just an excuse to be in the city and nobody goes to the talks
Just go to the city without a ticket
Today I'm publishing a new design for an off-chain decentralized exchange built on synthetics: the 🌈 Rainbow Network 🌈
- Supports any liquid asset (even non-crypto)
- Supports leverage and shorting
- Implementable on Ethereum, or on top of Lightning
Bar trivia question: “what did Satoshi Nakamoto invent in 2008?”
Answer: “Blockchain. We will not accept Bitcoin as an answer, he did not invent Bitcoin”
?!?!?!?!?!?
Man goes to doctor. Says he's depressed. Says crypto twitter seems harsh and cruel. Says he feels all alone in a threatening world.
Doctor says, "Treatment is simple. Follow
@0xtuba
. That should pick you up."
Man bursts into tears.
Says, "But doctor... I am
@0xtuba
."
We’ve had a great partnership with
@PacmanBlur
over the years and I’m excited to collaborate on research for Blast in the years to come
One interesting example is their design for rebasing tokens, which I initially pushed back against but they convinced me could work
It avoids…
Building a non-standard ERC-20 token (such as fee-on-transfer or inflationary/deflationary balances) is one of the nastiest things you can do to smart contract developers, wallet developers, and users
DEXes leak value to miners through three kinds of MEV:
1. Gas costs
2. Slippage/sandwiching
3. Loss-vs-rebalancing
Reduce any of these leaks, and you preserve more value for swappers and LPs.
So each of these categories corresponds to a promising line of DEX research.
To set a better precedent for future projects who may try to emulate them, we worked together with Blast to design a new timelocked upgrade system that they will upgrade to next week. The code is open sourced here and free for anyone to use.
Uniswap v3 is so flexible that it can simulate any static automated market maker.
My new post shows how this works, and how other AMMs like Curve, Balancer, and LMSR can be built on top of Uniswap v3:
Suppose it turns out that it's impossible to have a secure proof-of-work blockchain without a block reward. Which change to Bitcoin would you be more willing to accept?
Wrote up an idea for a new DeFi primitive: yTokens.
- Fixed-term secured lending
- Debt is fungible and can be traded on
@UniswapExchange
- Interest rates are implicit and determined by market
- Can be used to construct a yield curve
The Yield Protocol:
You'd think there would be a tradeoff between speed and quality
But the fastest developers I know are also the most obsessed with correctness, performance, and style
Maybe it's just quicker to do things the right way
Sometimes at conferences people ask me how they can become more crypto-native, and I just apologize as I take their phone and make them follow
@inversebrah
Bitcoin maximalism could be based on either of these beliefs:
1) Everything in crypto could be built on top of Bitcoin
2) Nothing in crypto should be built other than Bitcoin
The second is a normative claim and I don’t have much to say about it
But the first is a delusion
How do you construct a lending protocol that supports arbitrary collateral, has no oracles, and has no expirations?
Read the whitepaper to find out:
(hint: when in doubt, try Dutch auctions)
This is getting way too much attention in Bitcoin world
The protocol only works for two parties, so it can't be used in rollups or other multiparty applications
And Greg Maxwell proposed a much better protocol ("ZK contingent payments") to solve the same problem 12 years ago
The scariest thing about the Ledger announcement is not as much what they're doing, but the revelation that they can do it at all
Like if the government announced that anyone could now opt in to an over-the-air vaccine update
The ultimate testnet faucet
No more begging for testnet ETH one network at a time
No more tweeting and deleting your address
No more deploying your own mock tokens or NFTs
The era of Paradigm MultiFaucet has begun
We built a new faucet:
✅ Funds wallet across 4 testnets at once (Ropsten, Kovan, Görli,
@optimismPBC
Kovan)
✅ Drips 5 ETH, 5 wETH, 5K DAI, and 5 ERC721 NFTs
✅ No need to tweet your address, just sign in with Twitter instead
Squeeth is live.
This is one of the most elegant mechanisms I've seen come out of DeFi, and further proof that web3 enables innovation and iteration at an unprecedented pace
Now it's in production, thanks to a Herculean effort from the
@opyn_
team
1/
What if Clubhouse, but with a professional host asking questions instead of randos in the audience, and it's recorded so you don't have to listen to it live
Working for a modern DAO seems great
Who wouldn’t want every hiring, promotion, and budgeting decision to be hashed out in public by people you’ve never met
Why don’t Coinbase and Kraken have a payment channel set up so transfers between them don’t have to use the chain or pay tx fees? That seems like the low-hanging fruit on the way to full Lightning deployment
@hrdwrknvrstps
@lex_node
Right—so this is basically bonds of the staker, collateralized by their staked ETH. But there’s a default risk
You can complete this picture with slashing swaps (basically CDS) so the traders can buy insurance on the risk of someone being slashed