Profile picture
Vitalik Buterin @VitalikButerin
, 14 tweets, 3 min read Read on Twitter
My conclusion from yesterday's polls: people have very underdeveloped intuitions about how bad various kinds of 51% attacks on blockchains are and how easy or hard they are to recover from, and tradeoffs between security margin and cost.
Exhibit 1: people are apparently willing to accept basically exactly the same failure rate to avoid 1% issuance as they are to avoid 5% issuance.

I will make the usual behavioral economics caveat that what people say they want in surveys and what people can be deduced to actually want from their actions are very different, but that just means that the public discourse on this topic is warped and doesn't reflect reality.
To be clear: I define a blockchain failure as any situation that causes the blockchain to not fulfill its basic guarantees that is bad enough and unrecoverable in-protocol, and so requires out-of-band coordination among users to move on.
For example, if a 51% attack on PoW happens, then the attacker likely has enough hardware to keep doing it forever ("spawn camp attack"), so the community has to change the PoW algorithm to "delete" everyone's ASICs.
In PoS, you can recover from 51% attacks by coordinating a minority UASF, and the community can do this an unlimited number of times, but out-of-band coordination is still required.
And yes, in either PoW or PoS, this WILL happen. The idea that if a 51% spawn camp attack happens, all $190b of bitcoin's users will just pack up their bags and leave is absurd; way too much incentive to coordinate and try to continue the ledger.
Though an important question is, how easy or hard is this? IMO it's much harder in PoW than PoS, and so PoS can achieve the same level of reliability in practice with a higher frequency of attacks, and hence can survive with a quite low cost of consensus
Now the second question. Now that we know 51% attacks are survivable, can we try to estimate the cost of one? I see two main factors:

1. Cost of loss of service during the attack itself
2. Cost of giving the social layer too much power by over-actively using it
After all, minimizing use of the social layer *is* what blockchains are about. But reducing it to zero has infinite cost, and so there are real tradeoffs between minimizing the social layer and minimizing cost.
There are opportunities to improve things with better technology, for example things like… can do 80% of the work of social coordination automatically, making it easier to use against attacks and harder to abuse for other ends.
Reminder: if a blockchain fails and recovers, you still have all your assets, unless they were in channels and loss of liveness during the attack enabled an attack on the channel. It's not 100% truly fully yours unless it's on-chain; channels are already a security/cost tradeoff.
Though with long withdrawal times and well-designed gadgets (bonded service providers, insurance markets, etc) they can be quite a good security/cost tradeoff.
So accepting a 1% issuance rate to avoid a 1% chance of attack per year is actually really not that smart a choice to make, at least if your use of a blockchain is for a cryptocurrency to store your value.
Missing some Tweet in this thread?
You can try to force a refresh.

Like this thread? Get email updates or save it to PDF!

Subscribe to Vitalik Buterin
Profile picture

Get real-time email alerts when new unrolls are available from this author!

This content may be removed anytime!

Twitter may remove this content at anytime, convert it as a PDF, save and print for later use!

Try unrolling a thread yourself!

how to unroll video

1) Follow Thread Reader App on Twitter so you can easily mention us!

2) Go to a Twitter thread (series of Tweets by the same owner) and mention us with a keyword "unroll" @threadreaderapp unroll

You can practice here first or read more on our help page!

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just three indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member and get exclusive features!

Premium member ($3.00/month or $30.00/year)

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!