Conversation

Replying to
I’m genuinely surprised by the number of people whose basic posture is “give me one real use case and convince me.” I can understand hostile rejection or just diving in to either explore or grift. But standing around like it’s somebody’s job to win you over? That is odd.
5
61
Kinda ironic that Web3 thing presents as 100% transactional and money driven, and you have to buy in with cash, but it’s nobody’s job to actually sell it to you. Yes, there are self-appointed evangelists and people shilling coins, but this isn’t *actually* a product.
2
27
It’s a cousin of Paul Ford’s “why wasn’t I consulted?” (wwic) phenomenon. “Why wasn’t I marketed to?” (wwmt) Why didn’t somebody carefully analyze your tastes and values and figure out a positioning and value proposition to appeal to you? Why indeed. ftrain.com/wwic
2
51
I suspect this is how people miss booms, waiting to be invited. Same thing happened with PCs, Web1, Web2. This is a crash-only party. Nobody is invited. Everybody is a party crasher. Starting with Satoshi crashing central banker party. Crash in or sit it out. wwmt is silly.
1
88
I think Web3 is waiting for its cgi moment. Web1 went from memoryless to stateful over ~5 years with cookies+cgi. Web3 is coming from the other extreme, from only immutable memory towards stateful.
Quote Tweet
Replying to @vgr
@ricmac’s Web History Project is the best thing like this that I am aware of webdevelopmenthistory.com/1993-cgi-scrip
1
35
It’s not just gas fees. The whole persistence model is like eeprom. Write operations are like flashing in embedded programming. Need a throwaway state on top that gives up some trustless verifiability for convenience for the less critical stuff.
1
24
Like if you wanted to use NFTs as tokens in a decentralized wallet-to-wallet game, where do hold game state? In a Web2 style db layer? That seems wrong. But it also seems silly to store game state on chain.
2
19
Okay fine not necessarily state, but *some* controlled transient persistence model that’s decoupled from browser session but is less expensive/permanent than on-chain, while still being Web3 native paradigm-wise.
Quote Tweet
State is usually bad. What you really want in most situations are values and references (see infoq.com/presentations/ and infoq.com/presentations/).
Show this thread
3
17
Replying to
doesn't that end up violating the laws of "physics" that exist here? I mean, as soon as it's off-chain, isn't it essentially web2-equivalent?
1
Replying to
Yeah but there’s still a lot of stylistic design room to be Web3ish. Like use ipfs to store rather than a regular db, make state consensus a wallet-to-wallet negotiation rather than a queue at a server etc.
Replying to
Sure, that’s fair. I do think we’re going to see “web3”-washing for a long time, though, and I wonder about the analogy to the development of “AI” / what the right way to think about the ultimate applicability will be.