Continuing my thought from the previous post: there are as you say a few rays of hope, around 'Web3' though I don't think they are what the people doing VC right now (Venture Capital, not Vinyl Choride, which is also in the news) think they are.
First, it has been clear since the era of Goedel and Turing that 'cryptography' (and number theory) is frightfully important as a driver of 'algorithms'. Algorithms like Satoshi's (a pseudonymous group like Nicholas Bourbaki) give us a *heuristic* but workable solution to the Byzantine Generals problem and Sybil attacks, which have created a huge opportunity. More recently, NFTs, despite the stupidity and grift of the Sam Bankman-Fried sort, are complete game changers.
I invite you to think about NFTs only don't think of them as jpgs of apes or the next tulip mania High-Yield Investment Programme (HYIP) game to hop into, but rather as the natural representation of METADATA. As you know, metadata is 'data describing data' -- database schemas, classes [metadata objects] vs object literals -- that sort of thing. Think of anything that Microsoft might consider representing in a MOF file ('metaobject framework'), or registering at schema-dot-org, as what schemas are in a data base -- a bunch of rows in a table with a names like mysql.tables or mysql.users each of which describes the SQL statement in the DDL (not the DQL etc). In short, databases have data and metadata (their schema) and it is key that the metadata, being after all *data*, is capable of defining the data tables themselves, as to their FORM, but not their MATTER, using the same mechanism as is used to store the data. We call such a schema a 'model' and then there is a sort of 'metamodel' which is the data structure for the 'system tables' in any database -- which you could also make self-describing but basically is hard coded in the RDBMS.
Now here is my assertion: every ROW in the system database (the thing the MOF describes)[1], that stores the 'configuration objects' or 'properties' (object literals) of which any software system ever designed has ZILLIONS... is a natural NFT. You want your schema under source control? Done, on a blockchain. Each such release or build or deployment configuration or possible 'registry entry' is registered, in a permanent, distributed, and irrevocable sense, by minting an NFT to represent it.
[1]: I know it's been renamed a half dozen times on different timelines. It's worse than the Berenstein Bears killing Mandela.
With this vision (meta data queries should search and return NFTs) we can now re-envision 'The Search Engine' properly -- What 'classic Google' does is not return 'all the worlds data' but a metadata representation, itself based on things such as hard coded RDF, representing the original dmoz index, or the 'Google Knowledge Graph' which build on a project freebase that slurped dbpedia which is a tabular representation of the greatest crowd-sourced collection of HUMAN GENERATED and ANNOTATED METADATA, wikipedia.
Curated data cannot 'beat Google at its own game' but curated data *on the blockchain* can.
In short, every invention of the Semantic Web -- every vCard, every Atom RSS feed, both the 'Resource Description Framework' and 'Resource Description Framework Semantics' rdfs-standard[2] and their graph sequellae 'RDF Triples' in Notation-3, along with 'Description Logics' such as OWL and LISPy schema languages like KIF -- indeed the whole of 'Knowledge Engineering' as a field ... are so many NFTs to be registered forever, immutably.
[2]: Dayum, that was a nice piece of work that.
The Semantic Web was premature -- NFTs are not. So what does a search engine look like? Well, it's a searchable store of NFTs (each of which encapsulates an URL in the conventional representation, along with a snippet, and whatever 'features' are needed to build the index)... that uses a human-machine text interface to query in 'natural language'.
Let's say I want to build a search engine to search Phora threads -- well each thread has some metadata (and an RSS like Title-Link-Description triple that can be described as a Knowledge Card or a Graph or an item in a newsfeed -- old style)... well we mint one NFT with the appropriate properties for each thread. The query searches the METADATA (the NFTs, which can all be at opensea.io[3] or the latest hotness) and gives you a QR code and text snippet to your link. Do you have the right to SEE the thread? Well, that's part of the smart contract and if so you get to retrieve the data from the IPFS using the CID in the NFT.
[3]: would you like to buy a Salo Archive thread? How about a Unicode emoji?
Very Hairy?
Now you have an uncensored, global, distributed search engine that is resistant to DDOS and Sybil attacks and lots of others. To stop the search, you have to take out the Ethereum Virtual Machine and destroy the entire global economy, Sam Bankman-Fried and friends included.