I'm disappointed that none of the crypto (or AI, or network state, or...) critics so far have used "beware of geeks bearing grifts" as a slogan. It's just.... right there for the taking!

Doom Debate with @VitalikButerin has dropped: Will “d/acc” Protect Humanity from Superintelligent AI? Vitalik is both a hardcore techno-optimist and builder, and someone willing to seriously consider AI regulation and coordination mechanisms. He coined the term "d/acc'" – defensive, decentralized acceleration – as a middle path between uncritical AI acceleration and total pause scenarios. We debate the tractability of AI alignment, whether approaches like d/acc would actually work when superintelligence arrives, and why he thinks a pluralistic world of competing AIs might be safer than a single aligned superintelligence. The crux of our disagreement is that I think we're heading for a "plants vs. animals" scenario where AI will simply operate on timescales we can't match, while Vitalik believes we can maintain agency through the right combination of defensive technologies and institutional design. Finally, we tackle the discourse itself – I ask Vitalik to debunk the common ad hominem attacks against AI doomers, from "it's just a fringe position" to "no real builders believe in doom." His responses carry weight given his credibility as both a successful entrepreneur and someone who's maintained intellectual honesty throughout his career. Big thanks to Vitalik for a good-faith debate!

"I support it only if it's open source" should be a more common viewpoint

A good book, worth reading to understand the basic case for why many people, even those who are generally very enthusiastic about speeding up technological progress, consider superintelligent AI uniquely risky

Echoing something @karpathy recently said, it does frustrate me how a lot of AI development is trying to be as "agentic" as possible, when actually creating *more* paths for human input both creates a better output (now for quite a while going forward) and is better for safety.

It's interesting how Visa was created with decentralist ideals so similar to those that gave rise to modern DAOs, and yet today many of us perceive it as extractive and/or oppressive. Certainly some good insights to learn from there.

Privacy is normal, and writing code is not a crime.

Amazing to see so many major L2s now at stage 1. The next goal we should shoot for is, in my view, fast (<1h) withdrawal times, enabled by validity (aka ZK) proof systems. I consider this even more important than stage 2. Fast withdrawal times are important because waiting a week to withdraw is simply far too long for people, and even for intent-based bridging (eg. ERC-7683), the cost of capital becomes too high if the liquidity provider has to wait a week. This creates large incentives to instead use solutions with unacceptable trust assumptions (eg. multisigs/MPC) that undermine the whole point of having L2s instead of fully independent L1s. If we can reduce native withdrawal times to under 1h short term, and 12s medium term, then we can further cement the Ethereum L1 as the default place to issue assets, and the economic center of the Ethereum ecosystem. To do this, we need to move away from optimistic proof systems, which inherently require waiting multiple days to withdraw. Historically, ZK proof tech has been immature and expensive, which made optimistic proofs the smart and safe choice. But recently, this is changing rapidly. is an excellent place to track the progress of ZK-EVM proofs, which have been improving rapidly. Formal verification on ZK proofs is also advancing. Earlier this year, I proposed a 2-of-3 ZK + OP + TEE proof system strategy that threads the needle between security, speed and maturity: * 2 of 3 systems (ZK, OP) are trustless, so no single actor (incl TEE manufacturer or side channel attacker) can break the proof system by violating a trust assumption * 2 of 3 systems (ZK, TEE) are instant, so you get fast withdrawals in the normal case * 2 of 3 systems (TEE, OP) have been in production in various contexts for years This is one approach; perhaps people will opt to instead do ZK + ZK + OP tiebreak, or ZK + ZK + security council tiebreak. I have no strong opinions here, I care about the underlying goal, which is to be fast (in the normal case) and secure. With such proof systems, the only remaining bottleneck to fast settlement becomes the gas cost of submitting proofs onchain. This is why short term I say once per hour: if you try to submit a 500k+ gas ZK proof (or a 5m gas STARK) much more often, it adds a high additional cost. In the longer term, we can solve this with aggregation: N proofs from N rollups (plus txs from privacy-protocol users) can be replaced by a single proof that proves the validity of the N proofs. This becomes economical to submit once per slot, enabling the endgame: near-instant native cross-L2 asset movement through the L1. Let's work together to make this happen.

🤖 New EIP: 0000 By: - Anders Elowsson (@anderselowsson) -Vitalik Buterin (@vbuterin) 🔗 EIP-0000 proposes a unified multidimensional fee market for Ethereum transactions, allowing users to specify a single aggregate `max_fee` for multiple resources. This aims to enhance capital efficiency by treating the `max_fee` as fungible across different resource types, thereby simplifying user experience and improving economic efficiency. The proposal also seeks to unify existing fee mechanisms and facilitate future expansions into additional resources.

Open AsteroidImpact. Not mine but incredibly incredible, the best I've ever seen of whatever this is.