Again, another piece in which Larry Fink predicts mass tokenisation. Whilst I relish this continuous push towards RWAs and the on-chain onboarding of everything, I cannot help but wonder why those involved with hyping all of this up never ask THE most pertinent question of all: How do they plan to verify all of these underlying computations that will make it all possible? Once everything is tokenised, how are the computations that feed the ledgers verified, and by whom? Fink and Goldstein say that tokenisation is the next SWIFT-level upgrade to the financial world in which equities, bonds, real estate, etc., are stored on-chain so they can be moved around faster and with less middlemen (and back-office shenanigans I might add). The assumption here is that when you implement it all and add a sprinkle of regulation, somehow the trust problem is solved. Ethereum and other chains are good at telling you wallet x has balance of y, or Bob sent X to Alice, but how will it tell you it ran through an off-hain compliance workflow, and most importantly, will it give a potential auditor or regulator the ability to replay the calculation after it has happened and for as long as needed? People tend to assume that an on-chain asset's proof of state automatically means proof of process. That is not the case. A blockchain will eagerly record whatever number the off-chain engine gives it, and if the engine (a euphemism for black box in my world) misprices risk, incorrectly records a NAV, or mis-checks compliance, the destination ledger makes the error a permanent one. I urge everyone to read truebit.io/the-trust-problem… and truebit.io/why-compliance-is… over and over again until the penny drops. @JasonTeutsch talks a lot about black boxes, and black boxes don't jus relate to AI. In this kind of tokenisation it'll be anything that outputs information to a ledger and is then treated as truth. The article states "tokenisation can replace paper with code," which I agree with, but the codes need to be verified and free of blindly trusting some New York-hosted black box that says "here is your yield and risk score, just trust us." My prediction is that within the next few years, you will not be able to launch a serious tokenised product without declaring your 'Task Stack' - see here for how to write tasks on Truebit Verify - devs.truebit.io/developing-t…. I guarantee you that as tokenisation spreads, every single business involved will discover internal black box processes that will need to be proven. This is where the tasks evolve and grow (devs will be paid to wrap these in reusbale tasks), to the point where different industries will have their own task libraries, so for example: funds/banks, insurers, AI agent firms, DePIN networks, and so on. There will also be crossover in these task stacks, naturally. Once the tasks are there, the requesters can get their certified transcripts, and everyone will be happy (including the node operators getting paid to perform the verification in between). At the moment I imagine there is a small amount of generic tasks already developed (I know Forte are busy writing them), which is why we're so early on this technology, and in the future we are likely to see many (thousands?) domain-specific verification primitives. Business models built on verifiable compute do not exist yet, but they will (my crystal ball says so :D). #AI #VerifiableComputation #Truebit #ProgrammableTrust #OnChainTrust #TrustlessProofs #Web3Infra #ERC3643 #RWAs #JustVerifyIt #TruebitVerify #TRU #AgeOfVerification #VerificationLayer