l2beat/acc AI accelerates development, no doubt. And yet, even though we use the latest frontier models, there's still a lot of human research needed to make sure you can trust us. And if you don't trust - verify. All our claims are onchain verifiable

The will of the community to build truly trustless bridges is next to zero. Users instead are being told to trust Vendor B instead of Vendor A or build their own bridge. If we go with the latter advice we will end up with tens of thousands of individual bridges. Some maybe quite good, some horrible. No one will know or care to monitor that mess until the next hack Layer after layer of offchain infra is being proposed as a security improvement. The only problem with that is that this is unverifiable. Who still remembers Multichain that advertised the most advanced MPC setup only to be found that the majority of keys were on a founder's laptop ? Maintaining complex, secure setup is costly - who will guarantee that what's set up securely today will be secure in two years after cost cutting, team change, tech upgrades ? You cannot trust what you don't see and can verify yourself. But bridging turns to be one of few really good businesses in Web3 while building trustless bridges is essentially creating public goods infra.