pikachu

I doxed this guy. @Eljaboom

The timeline for a quantum computer capable of breaking blockchain encryption is accelerating. A few years ago it seemed we wouldn't see such a computer in our lifetimes. 2025 showed that one could exist within a decade. With AI, the timeline accelerated faster than expected. There's a world where quantum computing follows the same trajectory, and blockchains are forced to respond to this existential threat. This is an extremely unlikely scenario, but one that would have existential consequences for the blockchain industry. For @Sei_Labs, the logical path forward is to make sure that we have a practical, actionable solution for making @SeiNetwork post-quantum, well before a cryptographically relevant quantum computer can exist. Right now the priority of every blockchain is scaling: increasing transaction throughput and reducing latency, without compromising security or decentralization. The next-generation of blockchains need to feel like the internet. That is the purpose of forthcoming Sei Giga upgrade, which will scale the chain to 5 gigagas/s throughput and 400ms latency. Taking today's iteration of Sei Network post-quantum would be relatively straightforward: upgrading the network's signature schemes. On the current network, user signatures are currently ECDSA, and validator signatures are currently Ed25519. @NISTcyber has finalized two post-quantum signature schemes: ML-DSA (module lattice, derived from CRYSTALS-Dilithium) and SLH-DSA (stateless hash-based, derived from SPHINCS+). Simply upgrading to these would make Sei Network post-quantum with relatively low-lift. But our plans to scale the chain complicate this massively. The current chain requires roughly 1,100 bytes/sec in bandwidth. Based on initial estimates, with the current signature scheme Sei Giga will require 13 megabytes/sec. The credible post-quantum signature schemes are huge in comparison. With these signature schemes, Giga's bandwidth requirement would increase to, at minimum, 1.57 gigabytes/second. Possible alternatives to NIST would be isogenies such as SQISign and smaller lattice schemes like Falcon. Assuming SQISign is secure (there are reasons it might not be), it takes 50ms per signature to verify. As @muursb puts it in the blog: “At 200k TPS that’s 10,000 core seconds of work per second for just verification. On a 192 core AWS Hpc7a that’s 52 seconds of signature verification every second, assuming you do nothing else. This feels at least somewhat problematic.” The takeaway from this is that right now there is no perfect solution. It's possible to make blockchains quantum secure with today's technology, but it makes scaling much harder. This will be largely true of all blockchains, not just Sei Network. The advent of quantum computing will have massive effects on the blockchain sector, potentially disrupting many of the competitive dynamics that define the winners and losers between today's chains. For such a threat, there's no such thing as being too prepared. Only too complacent.