Micali, Silvio
▶ Watch on YouTubeCore ideas
Micali frames a blockchain as fundamentally a mechanism for creating verifiable "common knowledge": a shared, append-only ledger where everyone sees the same page at the same time, with a guarantee that no one can be prevented from writing, and no one can erase or alter what's written. He argues this is historically unprecedented — even instantaneous digital communication doesn't normally guarantee that what you received is identical to what everyone else received — and illustrates the power of this with an auction example, where public visibility of all bids lets participants trust a "fair price" was reached without relying on a central authority's word. He extends this into a broader philosophical claim about money itself: money, in his view, is fundamentally a social construct and shared belief system — gold, livestock, fiat currency, and cryptocurrency are all just different vehicles for the same underlying trust that others will accept what you receive in exchange — with scarcity (whether physical, as in gold or Bitcoin's fixed supply, or trust-based, as in a disciplined central bank) being a practically important but not strictly necessary feature that makes that shared belief easier to sustain.
The technical core of the conversation addresses what Micali calls the "blockchain trilemma" — the claim that a system can only achieve two of scalability, security, and decentralization at once — which Algorand was designed to refute. He identifies the central unsolved problem of any blockchain as: who gets the power to decide what goes on the next page of the ledger, since whoever controls that has more effective power than any historical monarch. He walks through existing approaches: Bitcoin's proof-of-work (a deliberately hard cryptographic puzzle whose difficulty scales up over time, which he argues has driven a de facto re-centralization into a small number of specialized, energy-intensive mining pools) and delegated proof-of-stake (entrusting a small fixed committee, e.g. 21 parties, which he compares to a small number of nobles replacing a single monarch — less centralized, but still centralized). Algorand's alternative, which he explains in detail, uses a cryptographic "self-selection" lottery: each of billions of tokens has an equal, privately-computed chance of being randomly selected onto a small verification committee for each block, with selection taking about a microsecond regardless of the number of tokens, and no one — including the selected members themselves — knowing who else was chosen until they've already broadcast their signed opinion. He argues this makes the system's security depend on the honesty of a majority of the entire token supply (rather than a majority of some small privileged subgroup), and makes targeted corruption practically impossible, since an attacker cannot identify committee members before they've already propagated their votes across the network — analogous, he says, to the impossibility of un-leaking a message already spread virally like a WikiLeaks disclosure.
On Bitcoin and Ethereum specifically, Micali credits Bitcoin with recognizing, ahead of anyone else, that the world needed a cryptocurrency and successfully making a ledger extremely hard to tamper with — calling this timing itself a major achievement — but argues it fails as a practical medium of exchange because of low transaction throughput, functioning well mainly as a store of value. He credits Ethereum's insight that programmable smart contracts (removing the need for costly, hard-to-find trusted intermediaries — he cites financial "friction" costs equal to roughly 6% of world GDP) represent an important next layer beyond simple payments, but argues it likewise suffers from scalability limits that constrain how much real demand for smart contracts it can serve. Asked why he didn't try to patch Bitcoin's or Ethereum's architecture rather than starting over, he says he judged that a from-scratch design was the better path given the depth of the problems he wanted to solve. On NFTs, he sees genuine value in giving digital creators fast, secure, verifiable ownership and trading rights over unique digital assets — extending the same scalability/security/decentralization requirements that apply to currency — while noting (in dialogue with Lex) that markets built around tokenizing individual creators or ideas represent a still-unexplored and potentially significant extension of this concept.
Micali argues decentralization of power is not merely an ideological preference but often a more resilient design, drawing a biological analogy (complex living organisms are themselves radically distributed systems of cells and bacteria, even though humans habitually design centralized, pyramid-shaped power structures) and citing a Machiavellian observation that decentralized, republican polities (Florence, Venice) proved more durable over time than the more easily-seized but more brittle centralized Ottoman model. He frames historical centralization partly as a consequence of limited communication technology rather than a permanent human necessity, and argues that blockchain-enabled common knowledge finally removes that technological excuse. He explicitly designed Algorand's governance to be adaptable rather than fixed ("code is law" permanently frozen in place, in his view, is "a recipe for disaster"), with protocol changes requiring roughly 90% supermajority stake approval — arguing that systems which cannot evolve eventually become "fossilized" and die. On leadership, he holds up George Washington's voluntary relinquishment of power after two terms as the model of successful leadership (a leader should "ignite and disappear"), contrasting this with leaders who cling to renewed mandates indefinitely, and argues charismatic founders (Satoshi Nakamoto, Steve Jobs) matter mainly insofar as they successfully embody and launch an idea that ultimately outgrows and outlasts them — arguing Satoshi's specific identity is essentially irrelevant compared to the coherence of the Bitcoin idea itself.
On privacy, Micali (a career cryptographer) affirms that privacy-preserving techniques are technically achievable on blockchains, but argues public understanding needs to mature before more advanced privacy features are widely deployed, cautioning against a world where people trust cryptographers' privacy claims on faith rather than understanding the tools themselves. He also discusses foundational ideas from his own field: the one-way function (easy to compute forward, computationally infeasible to invert — e.g., breaking versus reconstructing glass) as, in his view, one of the most elegant ideas in cryptography and computation, underlying pseudorandom number generation (expanding a small secret random seed into arbitrarily long unpredictable sequences) and digital signatures; interactive and zero-knowledge proofs, which he explains as reframing mathematical truth as a statistically verifiable "game" that can be played and checked far faster than reading an exhaustive formal proof, and which established that knowledge and verification are logically separable (you can be convinced something is true without learning why); and his own work in mechanism design, aiming to design "games" (incentive structures) where self-interested behavior — even just avoiding clearly foolish choices, rather than solving for one unique optimal strategy — reliably produces the desired social outcome, which he argues is the design principle underlying Algorand's incentive-free consensus mechanism (since generating a block is computationally trivial and therefore doesn't need to be compensated the way energy-intensive Bitcoin mining does).
On free will and quantum computing, discussed briefly and speculatively, Micali says he believes it's practically important for people to believe free will exists regardless of its ultimate metaphysical status, drawing a parallel to how pseudorandomness is functionally indistinguishable from "true" randomness over any observer's lifetime even if a philosophical distinction technically exists. On quantum computing, he argues that if quantum computers become practical, the classical notion of a "one-way function" would need to be redefined in terms of quantum-hard problems rather than becoming obsolete, and expresses hope that such functions continue to exist, since he views the existence of computationally hard, non-reversible processes as essential to making life and problem-solving meaningful. He closes with reflections on cultural identity (contrasting Italian cuisine's minimalism with French/Chinese culinary complexity, and Dante's invention of literary Italian from a limited regional dialect as an example of how self-imposed limitation can be a creative "superpower"), and on meaning and mortality, framing death as life's "ultimate motivator" and describing his own hoped-for legacy as both biological (his children) and intellectual — his scientific work on privacy and security as, in his framing, fundamentally about enabling human freedom to interact and take risks without fear.
Excerpts
On the core problem every blockchain must solve:
“Who the hell chooses or publishes the next page of the ledger? ... If it is me, I would have more power than any absolute monarch in history, because I would have a tremendous power to say these are the transactions that the entire world should see, and whatever I don't write will never see the light of day.”
On money as a shared belief rather than a physical necessity:
“Money is a social construct... whether it is gold, whether it is livestock, or whatever it is, money is really a shared belief. You must believe that somebody else will accept them in return for something else.”
On why systems should be designed to evolve rather than be fixed:
“Nothing fixed lasts very long... a lot of crypto projects are actually very proud to say 'code is law,' this will never change — I'm saying this is a recipe for disaster... without evolving, any system essentially becomes fossilized and is going to shrivel and die sooner or later.”