Opinion: Security depends heavily on UX. Crypto needs intuitive UX that allows features like 2FA and other simple, secure ways to verify and approve actions. One of the key foundations that enables this type of UX on Starknet is native AA. Good design choice.
FYI: Post-quantum security depends on crypto agility as much as it depends on cryptography. Explanation: Think of migrating to post-quantum security as upgrading the locks on your house. Having the strongest lock available is important. But what if a better lock-picking technique emerges tomorrow? Can you replace your locks quickly, or would you need to rebuild the entire house? What if you can replace the locks easily, but the other tenants have to agree first, and they prefer to wait until they actually see a thieve trying to break in? All these things that let you migrate quickly to safer cryptography - that's crypto agility. Getting close to post-quantum security matters. But so does being able to adapt quickly when cryptography needs to change. You need both.
If Satoshi were here today, would he prioritize keeping Bitcoin exactly as it is, or would he make sure Bitcoin survives the next 100 years? Without a doubt, he'd support OP_CAT, OP_STARK, preparing Bitcoin for Q-day, and anything else that helps secure Bitcoin's future.
My take: AI moving fast means crypto agility is no longer a nice-to-have.
Why AI threatens public key encryption (PKE), and how will that affect you? (1) What is PKE? You book your first vacation at a hotel in Hawaii. You enter your credit card number to that website. You may not be aware of it, but before your computer sends that data over the internte, it ENCRYTS it, so it won't get stolen. But here's a problem: you never met the hotel in Hawaii, so how does your computer encrypt your information so that (1) the hotel is able to *decrypt* it, and (2) no one else manages to decrypts it, even though you send it over the internet for all to see? PKE is how it's done: The hotel has two keys: - One key seals messages, and cannot open them. The hotel in Hawaii posted it and your computer used it to encrypt your credit card number. - The other key is a secret key, and it's kept with the hotel in Hawaii. Only this secret key can open messages that were sealed with the public key. (2) Enter super-AI: What's this got to do with super intelligence? All PKE schemes rely on some clean and beautiful mathematical assumption to work safely. For example: finding a divisor of a large number (say, finding out that 1,234 divides 28,495,528 with no remainder). Note: A PKE is safe only as long as the math problem remains hard. The problem is that AI got really really good at solving hard math problems. Much better than the smartest human mathematician, and it's already shocked mathematicians by showing that what they considered hard problems are not that hard to solve. (3) How will that affect all of us? What if AI shows us that each of the math problems underlying PKE is not that hard to solve? It'll be a huge earthquake. I'll repeat: IT'LL BE A HUGE EARTHQUAKE. We won't be able to rely on them to order our next vacation (in Bali), or to buy stuff from a new online store, or for a million things we do today. If PKE math is easy to solve, the Internet will need to rush pretty quickly to find new ways of doing business without PKE, without sending private information to counter parties. The good news: you can do that, using things like ZK STARKs and hash functions and digital signatures based on them. So it's a bit like a big meteor hitting earth. Dinosaurs go bust but some creatures survive and thrive.
Q-day is coming. Everybody needs to get ready. There's a lot to keep track of in the quantum discussion. We're putting significant effort into quantum research and developing quantum-safe solutions. For those interested, we've gathered a few projects we're involved in: 1. ECDSA/fail: Testing how much AI can reduce the resources needed for a quantum attack on elliptic-curve cryptography. 2. QSB: Quantum-safe Bitcoin txs, including the research paper and efforts to reduce transaction costs. 3. PQ Signatures Challenge: Working to reduce the cost to verify post-quantum secure signature. Go read, explore, and learn.
This whole conversation was fun and we touched on several interesting topics, but I want to spotlight one part: Would the technology powering Starknet today have been accepted and used in blockchain's early days? I say: absolutely. ZK tech back then wasn't nearly as developed as it is today, but based on what Satoshi said about ZK, I have no doubt he would have loved the ZK tech powering Starknet. Watch the whole conversation!
We're talking a lot about different types of cryptography and which types will survive quantum-computers and beyond that, superintelligence. Let me explain why Hash Based Cryptography is safest against superintelligence: All other forms of cryptography *imply* secure hash functions. - Factoring is hard? Great - you get secure hash functions! - Lattice problems hard? Yahoo! You have hash functions! - Decoding hard? Another hash function! AND EVEN IF ALL THESE MATHEMATICAL SYSTEMS GET BROKEN, YOU COULD STILL HAVE SAFE HASH FUNCTIONS. In case of Tsunami, 1,000 foot peak is safer than on a 10 foot hill.
Forecast: 20+ knots. Now sub 5 😩😩 https://t.co/9uCqJnN4qZ
Shadow Accounts and STRK20 are making privacy usable for DeFi. Another example below: (As always, not investment advice,DYOR)
