Happy new year!
Today, a mini-tweetstorm of some of the things I've said and written over the past decade, and what I think about those subjects today.
Conversation
1. In 2013, I wrote this piece on "how Bitcoin can actually help Iranians and Argentinians". The core point: Bitcoin's key benefit is internationality and censorship resistance, NOT "the 21m limit". I predicted stablecoins will prosper.
bitcoinmagazine.com/culture/how-bi
103
340
3,193
Last week, I actually went to Argentina! My verdict: generally correct. Cryptocurrency adoption is high but stablecoin adoption is really high too; lots of businesses operate in USDT. Though of course, if USD itself starts showing more problems this could change.
88
321
3,730
2. This piece from 2013 on the consequences of Bitcoin services becoming more "regulated":
bitcoinmagazine.com/culture/bitcoi
Core argument: Bitcoin is resisting the government not by being clever about what *legal category* it's in, but rather by being technologically censorship-proof.
41
112
1,750
My views today: sure, Bitcoin's decentralization would let it still *survive* under a super-hostile regulatory climate, but it could not *thrive*. Successful censorship resistance strategy requires a combination of technological robustness and public legitimacy.
50
189
2,494
3. My projections from 2015 of when we will get PoS and sharding. Honestly, these were very wrong and worth laughing at; I'll share a screenshot of one of my presentations from 2015 so everyone can laugh more easily.
58
213
2,779
But what was my core underlying mistake? IMO it's that I deeply underestimated the complexity of software development, and the diff between a python PoC and a proper production impl. 2014-era ideas were waaay too complex, eg. "12-dimensional hypercubes":
blog.ethereum.org/2014/10/21/sca
54
191
2,115
Today the Ethereum research team values simplicity much more - both simplicity of the final design *and* simplicity of the path to getting there. More appreciation of pragmatic compromises.
Dankrad's new sharding design is very much in this spirit.
Quote Tweet
I'm thinking about a new sharding design, where instead of having independent proposers for each shard, all shard blocks in one slot are proposed together with the beacon block. This leads to a major simplification of the sharding design 1/n
notes.ethereum.org/@dankrad/new_s
Show this thread
4. Though I 100% stand by my comment that "the internet of money should not cost more than 5 cents per transaction". That was the goal in 2017, and it's still the goal now. It's precisely why we're spending so much time working on scalability.
268
487
3,492
5. I should also add that the core *idea* of sharding has survived unscathed.
Blockchain 1.0: each node downloads everything, have consensus
BitTorrent: each node downloads only a few things, but no consensus
Ideal: BitTorrent-like efficiency but with blockchain-like consensus
120
217
2,360
The core enabling technologies are committees, ZK-SNARKs and data availability sampling:
54
174
1,855
6. In 2012, I was briefly an apologist for PoW energy waste. Though fortunately, by 2013 I got excited about proof of stake as a promising alternative. By 2014, I was sold.
bitcoinmagazine.com/business/the-w
bitcoinmagazine.com/culture/what-p
blog.ethereum.org/2014/11/25/pro
71
97
1,511
This reflects a broader intellectual evolution I've had: from "X is what I must defend, so whatever is favorable to X must be correct" to "I like X, but X has flaws and it seems like Y fixes them, so I support X+Y now"
Soldier mindset --> scout mindset.
140
636
4,804
7. This 2014 piece on self-enforcing contracts:
blog.ethereum.org/2014/02/24/dao
Basically, it tries to argue that making all of society more like a formal system is good and we should be excited about it.
87
211
1,372
I got to see the limits of this style of thinking through my work on collusion:
vitalik.ca/general/2019/0
vitalik.ca/general/2020/0
Basically, the problem with pushing everything into a formal system is that almost any formal system with more than 2 participants is attackable.
78
208
1,285
8. I liked altcoins before altcoins were cool. See this from Sep 2013 (so, 2 months before Ethereum):
bitcoinmagazine.com/business/defen
Three core arguments:
(i) diff chains optimize for diff goals
(ii) costs of having many chains are low
(iii) need outlet in case core dev team is wrong
161
238
1,407
Do I still agree?
The above args are less strong today, because (i) chains are more universal, (ii) apps more complex so bridging more risky, (iii) experimentation doable at L2.
But even still, I think there are things that you can't do at L2, and there's room for different L1s
106
270
1,491
9. I was optimistic about Bitcoin Cash specifically, because I agreed with the big-blocker arguments in the scaling war more than the small-blocker arguments.
Quote Tweet
Replying to @BitcoinErrorLog @TokenHash and 13 others
1. I consider BCH a legitimate contender for the bitcoin name. I consider bitcoin's *failure* to raise block sizes to keep fees reasonable to be a large (non-consensual) change to the "original plan", morally tantamount to a hard fork.
2. Theymos's censorship.
153
207
1,155
Today, I would call BCH mostly a failure. My main takeaway: communities formed around a rebellion, even if they have a good cause, often have a hard time long term, because they value bravery over competence and are united around resistance rather than a coherent way forward.
520
1,150
5,845
10. These posts from 2016-17 basically advocating for someone to build Uniswap.
vitalik.ca/general/2017/0
reddit.com/r/ethereum/com
146
181
1,422
Obviously feel proud of this one😀
Though it's interesting how I got the "do something really simple and dumb even if it's suboptimal" concept so right here, but it took me so long to get there in the proof of stake and sharding designs.
112
177
1,528
11. Applications envisioned in the Ethereum whitepaper:
ethereum.org/en/whitepaper/
* ERC20-style tokens
* Algorithmic stablecoins
* Domain name systems (like ENS)
* Decentralized file storage and computing
* DAOs
* Wallets with withdrawal limits
* Oracles
* Prediction markets
228
410
2,315
A lot correct (basically predicted "defi"), though incentivized file storage + compute hasn't taken off that much (yet?), and of course I completely missed NFTs.
I would say the biggest thing I missed in the details is collusion issues in DAO governance:
164
293
1,681
12. A more detailed opinion on stablecoins from 2014: blog.ethereum.org/2014/11/11/sea
A big part of this post tried to tackle the question of whether or not you could have a stablecoin without oracles, by using blockchain data (eg. PoW diff) as a pseudo-price-oracle.
178
200
1,207
I'm more pessimistic about this now especially because of the PoS switch. We will need oracles. And if we want to make stablecoins robust to USD collapse (by switching to their own native CPI if that happens) we'll need more active governance.
284
552
2,359
Conclusions:
* My thinking about politics and large-scale human organization was more naive then. Too focused on simple and complete formal models; I did not appreciate challenges of culture and vitalik.ca/general/2021/0 then; I do now.
192
322
1,990
* I did have good instincts early on for avoiding the craziest parts of bitcoin maximalist thinking. A couple early mistakes, but I corrected quickly
* But X being wrong does not imply that any specific rebellion against X will go well! Yet another way in which politics is hard
352
328
2,680
* On tech, I was more often right on abstract ideas than on production software dev issues. Had to learn to understand the latter over time
* I have a deeper appreciation now of the need for even more simplicity than I thought we needed
707
429
4,336