2/ While it is easy to point fingers, the builders among us can't look away from the pressure points of web2 and its reliance on the backbone. DNS and BGP are the shadowy figures that move all the 1s and 0s by making the services discoverable.

3/ They do this by being the identity layer for services and reachability. An identity layer has a ludicrous amount of power and a target behind its back, so naturally, all orgs will invest in the tech and infrastructure to protect these crown jewels.

4/ Conventional wisdom on defending these assets consolidated on running permissioned systems in the last few decades or so. Need to make an authorized change and/or audit any config changes? Here is an IAM/RBAC policy that allows you to do that.

5/ Putting the economic and power reasons aside, centralized systems will eventually debase themselves into the permission sprawl that no sensible engineer can escape unharmed. Taking a brief pause and asking, "Why do these permissions even exist?" can reveal compelling insights.

6/ If a protocol doesn't include native security primitives, security gets latched on as they are not built-in. For the primitives that can't be latched on, trust reduces to hoping that the operators secure their end and not send incorrect data.

7/ Web3/decentralization flips this model as the protocols have native security parameters that are built-in, and to top it off, the economic incentives impartially maintain a nice and naughty list to reward or punish the behaviors.

8/ To quote from @alive_eth's talk, web3 enables us to go from "don't be evil" to "can't be evil." Why does this matter? https://t.co/Tg3kiJiCIx

9/ Web2 made the gatekeepers too big to fail. Not if, but when they fail they create cascading failures. Nonavailability is the first-order problem; unable to rollout fixes is the boss battle here. This forces us to take sledgehammers to our own data centers.

10/ What are some potential Marie-Kondo-ing look like for the protocols, we love, hate, and are most reliant on? All network protocols are abstractions on routable and physical addresses? At its core, these are the only two things we need to send data from point A to B.

11/ If we think of the network RPC as IP and wallet addresses as MAC, in theory, we can write smart contracts that make everything in the crypto-verse addressable and discoverable. Sure this calls for standardization across protocols, endpoints.

12/ There are dragons both known and unknown for sure. But I'm optimistic that one day we might have crypto addresses as first-class headers that go into the packets of the Web3 world, and gatekeepers are trustless and secure. FIN