Vitalik Buterin

Co-founder of Ethereum. Publishes long essays on mechanism design, governance and what cryptography is for, and returns to earlier positions to say which parts they no longer hold.

Vitalik Buterin did not write this page. ppll compiled it from their own words, each linked to the source. Is this you? · Report a mistake

Site

Latest Site

The cryptographic world computer 27 Sept 2026

What they believe

Modern cryptographic networks are called "blockchains" largely for historical reasons.

As an AI might say, the only honest conclusion (ok fine, the honest punchline) is that a modern cryptographic network like the post-Lean-upgrade Ethereum is called a "blockchain" to a large extent for historical reasons.

vitalik.eth.limoSaid 27 Sept 2026This on korrents
Changed their mind Permissive licences are the better way to release free software and free content, because they maximise adoption.

Now Historically, I was a fan of the permissive approach (eg. my blog is under the WTFPL). More recently, I am warming up to the copyleft approach.

Said 7 Jul 2025This on korrents

The bold line is korrents’ wording; the quote is theirs. These one: what they wrote before what they said aloud, one per source, newest first.

All 9 beliefs and 1 change of mind, on korrents

Feed

As its own page →

Everything, newest first — across every channel.

Sources

27 September

Vitalik ButerinKorrents 27 Sept

From one piece The cryptographic world computer 3 beliefs · vitalik.eth.limo

korrents.com
The decentralization of cryptographic networks is shifting from purely a safety burden to, in a few limited cases, a performance strength.

cryptography

Their words

Perhaps the most important shift is that the decentralized nature of the network is shifting from being purely a burden incurred in the name of safety and robustness to at least sometimes, in a few limited cases, being a strength even from a performance perspective.

The cryptographic world computervitalik.eth.limo
korrents.com
The set of tradeoffs for users of cryptographic networks is radically changing.

cryptography

Their words

The most important conclusion is that the set of tradeoffs for users is radically changing:

The cryptographic world computervitalik.eth.limo
korrents.com
Modern cryptographic networks are called "blockchains" largely for historical reasons.

cryptography

Their words

As an AI might say, the only honest conclusion (ok fine, the honest punchline) is that a modern cryptographic network like the post-Lean-upgrade Ethereum is called a "blockchain" to a large extent for historical reasons.

The cryptographic world computervitalik.eth.limo

Vitalik ButerinSite 27 Sept

The cryptographic world computer
cryptography

21 August

Vitalik ButerinKorrents 21 Aug

From one piece Obfuscation (Part III): Local Mixing 3 beliefs · vitalik.eth.limo

korrents.com
So far, the mixing step in local-mixing obfuscation has left too many correlations between original circuit values and obfuscated values.

cryptography

Their words

But, so far, mixing has not proved to be good enough. There ended up being too many correlations between values in \(C\) and values in the obfuscation that remained.

Obfuscation (Part III): Local Mixingvitalik.eth.limo
korrents.com
A sufficiently large random reversible circuit is plausibly a secure cryptographic permutation, while a large random irreversible circuit collapses to only a few possible outputs.

cryptography

Their words

A big-enough random reversible circuit is plausibly a secure cryptographic permutation, a big-enough random irreversible circuit degenerates into having only a few possible outputs.

Obfuscation (Part III): Local Mixingvitalik.eth.limo
korrents.com
Local mixing, a new cryptographic obfuscation approach, is a risky bet resembling past failed attempts at white-box cryptography.

cryptography

Their words

This is a wild and risky bet; it sits on a graveyard of failed attempts at white-box cryptography .

Obfuscation (Part III): Local Mixingvitalik.eth.limo

Vitalik ButerinSite 21 Aug

Obfuscation (Part III): Local Mixing
cryptography

28 July

Vitalik ButerinSite 28 Jul

Obfuscation (Part II): Diamond iO
cryptography

29 June

Vitalik ButerinKorrents 29 Jun

From one piece Obfuscation: building the final boss of cryptography (Part I) 3 beliefs · vitalik.eth.limo

korrents.com
Known iO schemes built on reasonable security assumptions have literally galactic runtimes.

Their words

But within the good news, there is bad news: the run time is literally galactic

vitalik.eth.limo
korrents.com
Building a secure obfuscation protocol is extremely difficult.

cryptography

Their words

it turns out that making a secure form of obfuscation is really really hard

vitalik.eth.limo
korrents.com
Obfuscation is the most powerful primitive yet conceived in cryptography.

cryptography

Their words

The most powerful primitive that has been conceived in cryptography is obfuscation

vitalik.eth.limo

Vitalik ButerinSite 29 Jun

Obfuscation: building the final boss of cryptography (Part I)
cryptography

18 May

Vitalik ButerinSite 18 May

A shallow dive into formal verification
formal proof

2 April

Vitalik ButerinSite 2 Apr

My self-sovereign / local / private / secure LLM setup, April 2026
LLMsself-sovereign AI

Vitalik ButerinRecommends 2 Apr

Usesrcmnd.app
NixOS

Their words

The way that I use this daemon is that I run it on NixOS as a service, accepting requests on port 6000.

vitalik.eth.limo
Usesrcmnd.app
Arch Linux

Their words

Software I have been a Linux user for a long time. About a year and a half ago I migrated over to Arch Linux.

vitalik.eth.limo
Usesrcmnd.app
llama-server

Their words

As it turned out, ollama was not able to fit Qwen3.5:35B onto my GPU, but llama-server could. Hence, from that day forward, I resolved to cease being a cave-dwelling noob, and use llama-server (via llama-swap to make model swapping easier).

vitalik.eth.limo

+ 7 more

Mixed onrcmnd.app
Ollama

Their words

I used ollama before, but when I admitted to this in public half of Twitter told me that I was a noob and llama-server was clearly better and I must have been living in a very deep cave if I did not already know that. I tested their theory.

vitalik.eth.limo
Usesrcmnd.app
Qwen3.5:35B

Their words

I have been using the Qwen3.5:35B model and have tried it on each of these, and I also tried the one-step-larger 122B.

vitalik.eth.limo
Usesrcmnd.app
Bubblewrap

Their words

Sandboxing To keep my LLMs in check, I do most of my LLM usage from inside of a sandbox. I use bubblewrap for this.

vitalik.eth.limo
Usesrcmnd.app
SearXNG

Their words

I gave pi a skill for using the search engine SearXNG (which aggregates many search engines together at the same time), and one for calling into a daemon that I wrote that gives it access to read my email and Signal messages, and send-to-self, and send to others only with human confirmation.

vitalik.eth.limo
Likedrcmnd.app
pi

Their words

The verdict: pi plus a basic searxng skill outperformed Local Deep Research. Also, pi is just much more configurable: I can easily just tell it to use not just internet searches, but also my own world_knowledge directory.

vitalik.eth.limo
Dislikedrcmnd.app
Local Deep Research

Their words

Its responses are, in my view, pretty bland and not very high-quality. I did a side-by-side test of asking Local Deep Research a question, then asking pi the same question (telling it to use searxng to make as many internet searches as needed), and I fed both outputs into an LLM to ask which is better.

vitalik.eth.limo
Dislikedrcmnd.app
NVIDIA DGX Spark

Their words

I was not impressed with the DGX Spark; it's described as an "AI supercomputer on your desk" but in reality it has lower tokens/sec than a good laptop GPU - and on top of that, you have to figure out the networking details of how to connect to it from your actual work device etc.

vitalik.eth.limo

30 December 2025

Vitalik ButerinSite 30 Dec 2025

Balance of power

17 December 2025

Vitalik ButerinSite 17 Dec 2025

Let a thousand societies bloom

25 November 2025

Vitalik ButerinSite 25 Nov 2025

Plinko PIR tutorial

7 November 2025

Vitalik ButerinSite 7 Nov 2025

Galaxy brain resistance

19 October 2025

Vitalik ButerinSite 19 Oct 2025

A GKR Tutorial

5 October 2025

Vitalik ButerinSite 5 Oct 2025

Memory access is O(N^[1/3])

24 September 2025

Vitalik ButerinSite 24 Sept 2025

The importance of full-stack openness and verifiability

21 September 2025

Vitalik ButerinSite 21 Sept 2025

Low-risk defi can be for Ethereum what search was for Google
Google

12 August 2025

Vitalik ButerinSite 12 Aug 2025

On idea-driven ideas

The whole feed →

Beliefs

Korrents What they believe 9 beliefs, 1 changed mind — each backed by an exact quote.
Each is a — compiled by korrents.com, not by them: the one-line wordings are korrents', the quotes are theirs.
Recent
- Modern cryptographic networks are called "blockchains" largely for historical reasons.
  
  As an AI might say, the only honest conclusion (ok fine, the honest punchline) is that a modern cryptographic network like the post-Lean-upgrade Ethereum is called a "blockchain" to a large extent for historical reasons.
  
  The cryptographic world computer Said 27 Sept 2026
  
  Full record — everyone who holds this →
  
  cryptography
- The set of tradeoffs for users of cryptographic networks is radically changing.
  
  The most important conclusion is that the set of tradeoffs for users is radically changing:
  
  The cryptographic world computer Said 27 Sept 2026
  
  Full record — everyone who holds this →
  
  cryptography
- The decentralization of cryptographic networks is shifting from purely a safety burden to, in a few limited cases, a performance strength.
  
  Perhaps the most important shift is that the decentralized nature of the network is shifting from being purely a burden incurred in the name of safety and robustness to at least sometimes, in a few limited cases, being a strength even from a performance perspective.
  
  The cryptographic world computer Said 27 Sept 2026
  
  Full record — everyone who holds this →
  
  cryptography
Show 6 more

Local mixing, a new cryptographic obfuscation approach, is a risky bet resembling past failed attempts at white-box cryptography.

This is a wild and risky bet; it sits on a graveyard of failed attempts at white-box cryptography .

Obfuscation (Part III): Local Mixing Said 21 Aug 2026

Full record — everyone who holds this →

cryptography

A sufficiently large random reversible circuit is plausibly a secure cryptographic permutation, while a large random irreversible circuit collapses to only a few possible outputs.

A big-enough random reversible circuit is plausibly a secure cryptographic permutation, a big-enough random irreversible circuit degenerates into having only a few possible outputs.

Obfuscation (Part III): Local Mixing Said 21 Aug 2026

Full record — everyone who holds this →

cryptography

So far, the mixing step in local-mixing obfuscation has left too many correlations between original circuit values and obfuscated values.

But, so far, mixing has not proved to be good enough. There ended up being too many correlations between values in \(C\) and values in the obfuscation that remained.

Obfuscation (Part III): Local Mixing Said 21 Aug 2026

Full record — everyone who holds this →

cryptography

Obfuscation is the most powerful primitive yet conceived in cryptography.

The most powerful primitive that has been conceived in cryptography is obfuscation

Obfuscation: building the final boss of cryptography (Part I) Said 29 Jun 2026

Full record — everyone who holds this →

cryptography

Building a secure obfuscation protocol is extremely difficult.

it turns out that making a secure form of obfuscation is really really hard

Obfuscation: building the final boss of cryptography (Part I) Said 29 Jun 2026

Full record — everyone who holds this →

cryptography

Known iO schemes built on reasonable security assumptions have literally galactic runtimes.

But within the good news, there is bad news: the run time is literally galactic

Obfuscation: building the final boss of cryptography (Part I) Said 29 Jun 2026

Full record — everyone who holds this →
All 9 beliefs and 1 change of mind →

What is a korrent?

A korrent is a belief a person has stated in their own words: one sentence stating the claim, backed by a quote and a source, kept at korrents.com.

Under a name here, the quoted block is what they actually said. The korrent beneath it is the claim those words support, in korrents' wording — tap it to see the record, its source, and who else holds it.

Nobody here wrote their own korrents. They are compiled from public statements, and a person can change their mind, which is recorded too.

Recommends rcmnd

rcmnd What they recommend 10 things — each backed by an exact quote.
Compiled by rcmnd.app, not by them: the quotes are theirs, and each links to the post it was copied from.
Recent
- Disliked NVIDIA DGX Spark 2 Apr 2026
  
  I was not impressed with the DGX Spark; it's described as an "AI supercomputer on your desk" but in reality it has lower tokens/sec than a good laptop GPU - and on top of that, you have to figure out the networking details of how to connect to it from your actual work device etc.
  
  vitalik.eth.limo 2 Apr 2026
  
  What others say about NVIDIA DGX Spark →
- Disliked Local Deep Research 2 Apr 2026
  
  Its responses are, in my view, pretty bland and not very high-quality. I did a side-by-side test of asking Local Deep Research a question, then asking pi the same question (telling it to use searxng to make as many internet searches as needed), and I fed both outputs into an LLM to ask which is better.
  
  vitalik.eth.limo 2 Apr 2026
  
  What others say about Local Deep Research →
- Liked pi 2 Apr 2026
  
  The verdict: pi plus a basic searxng skill outperformed Local Deep Research. Also, pi is just much more configurable: I can easily just tell it to use not just internet searches, but also my own world_knowledge directory.
  
  vitalik.eth.limo 2 Apr 2026
  
  What others say about pi →
Show 7 more

Uses SearXNG 2 Apr 2026

I gave pi a skill for using the search engine SearXNG (which aggregates many search engines together at the same time), and one for calling into a daemon that I wrote that gives it access to read my email and Signal messages, and send-to-self, and send to others only with human confirmation.

vitalik.eth.limo 2 Apr 2026

What others say about SearXNG →

Uses Bubblewrap 2 Apr 2026

Sandboxing To keep my LLMs in check, I do most of my LLM usage from inside of a sandbox. I use bubblewrap for this.

vitalik.eth.limo 2 Apr 2026

What others say about Bubblewrap →

Uses Qwen3.5:35B 2 Apr 2026

I have been using the Qwen3.5:35B model and have tried it on each of these, and I also tried the one-step-larger 122B.

vitalik.eth.limo 2 Apr 2026

What others say about Qwen3.5:35B →

Mixed on Ollama 2 Apr 2026

I used ollama before, but when I admitted to this in public half of Twitter told me that I was a noob and llama-server was clearly better and I must have been living in a very deep cave if I did not already know that. I tested their theory.

vitalik.eth.limo 2 Apr 2026

What others say about Ollama →

Uses llama-server 2 Apr 2026

As it turned out, ollama was not able to fit Qwen3.5:35B onto my GPU, but llama-server could. Hence, from that day forward, I resolved to cease being a cave-dwelling noob, and use llama-server (via llama-swap to make model swapping easier).

vitalik.eth.limo 2 Apr 2026

What others say about llama-server →

Uses Arch Linux 2 Apr 2026

Software I have been a Linux user for a long time. About a year and a half ago I migrated over to Arch Linux.

vitalik.eth.limo 2 Apr 2026

What others say about Arch Linux →

Uses NixOS 2 Apr 2026

The way that I use this daemon is that I run it on NixOS as a service, accepting requests on port 6000.

vitalik.eth.limo 2 Apr 2026

What others say about NixOS →
All 10 recommendations →

Where they publish

Site vitalik.eth.limo Their site: the long essays, going back years.
The full archive of their writing on Ethereum, governance, cryptography and the politics of the technology. Long-form and technical; the ENS domain resolves to the same archive as the older vitalik.ca. Has a feed.

Open vitalik.eth.limo
Recent
- The cryptographic world computer 27 Sept 2026
- Obfuscation (Part III): Local Mixing 21 Aug 2026
- Obfuscation (Part II): Diamond iO 28 Jul 2026
Show 17 more

Obfuscation: building the final boss of cryptography (Part I) 29 Jun 2026
A shallow dive into formal verification 18 May 2026
My self-sovereign / local / private / secure LLM setup, April 2026 2 Apr 2026
Balance of power 30 Dec 2025
Let a thousand societies bloom 17 Dec 2025
Plinko PIR tutorial 25 Nov 2025
Galaxy brain resistance 7 Nov 2025
A GKR Tutorial 19 Oct 2025
Memory access is O(N^[1/3]) 5 Oct 2025
The importance of full-stack openness and verifiability 24 Sept 2025
Low-risk defi can be for Ethereum what search was for Google 21 Sept 2025
On idea-driven ideas 12 Aug 2025
"I support it only if it's open source" should be a more common viewpoint 12 Aug 2025
My response to AI 2027 10 Jul 2025
Why I used to prefer permissive licenses and now favor copyleft 7 Jul 2025
Does digital ID have risks even if it's ZK-wrapped? 28 Jun 2025
A simple explanation of a/(b+c) + b/(c+a) + c/(a+b) = 4 11 May 2025
Link verified 1 Sept 2026. Recent items update automatically from the channel.

About the English under a post

Some people here publish in a language other than English. Where they do, this site shows a machine translation beneath the post, in this typeface — the site's own, not theirs.

The post itself is never changed, moved or hidden: what is set in the serif above is exactly what the person published, and it is what to quote them on. A translation can be wrong in ways that matter, especially about tone.

Only the post's own words are translated. A quoted post, a linked article and a belief on korrents.com are left in their original language.