> Ironically for Web3/etc. the "(no) trusted third party" is both the selling point and it's largest problem.

Here’s ethereum.org’s comparison of web 2 vs. web 3: https://ethereum.org/en/developers/docs/web2-vs-web3/

I don’t see any mention of trustlessness. Rather, the selling points seem to be around different network designs.

Looking at the same page I come to a different conclusion. The pros it lists are permissionlessness, censorship resistance, and avoidance of a single point of failure. Cons are scalability and cost.

While they don't mention the word "trustless", those 3 pros are all things that you lose when you rely on a trusted third party. And since you pretty much always have to add a trusted third party to make the dApp useful, you effectively lose the pros and keep the cons.

They don't mention smart contracts in that page either. However, if you look at https://ethereum.org/en/web3/ you will see this: "At its core, Web3 uses blockchains, cryptocurrencies, and NFTs to give power back to the users in the form of ownership". And this: "Web3 is trustless: it operates using incentives and economic mechanisms instead of relying on trusted third-parties."

So I guess, when they compare web2 with web3, they implicitly mean the things I quoted.

Funny that you use that example. That's a classic example of something that's achievable in a trustless manner. If the address has no outbound activity for x time, transfer assets to beneficiary. Something like who wins the sportsball event would have to rely on a 3rd party oracle, of which there are many, but for the right price any of them can be tampered with.

I wouldn't want to rely on a dead man's switch pretty much my whole live. But you are right: There are many more examples. The one you mentioned suits better.

Can you have a trusted third party in a decentralized world?