Conversation

Worth compiling a list of proof-of-Xs. For example: proof-of-who: digital signature proof-of-what: cryptographic hash proof-of-when: blockchain timestamp There's also proof-of-location, proof-of-space, etc. Each such proof expands the scope of the ledger of record. Got more?

304

This Tweet was deleted by the Tweet author. Learn more

Balaji Srinivasan

@balajis

Aug 5, 2020

Yeah, you could do that with an NFT signed by (say) Lambda School, quantum.country, or fast.ai where they assert that you had a particular skill at a particular time in their assessment. cc

A free introduction to quantum computing and quantum mechanics

Andy Matuschak

@andy_matuschak

Aug 5, 2020

I’m not sure about certification, but I’ve been thinking that “proof-of-memory” makes for an interesting social signal: if I see that Balaji’s diligently remembered N details from an article, that sure means a lot more than “Balaji liked this tweet.”

notes.andymatuschak.org

The mnemonic medium can surface “proof of memory” social signals

Balaji Srinivasan

@balajis

Aug 5, 2020

That's super cool. In some cases, N proofs-of-memory can be aggregated into a single proof-of-skill. Like a recent score on a history exam demonstrates working knowledge of a particular time period.

Ben Reinhardt

@Ben_Reinhardt

Aug 5, 2020

I wonder about the legibility barrier for converting proofs of memory into proofs of skill. If the different pieces of the skill are legible, memory is a good proof. Many skills are frustratingly illegible.

Replying to

Yes, I’m worried about Goodhart’s law here, particularly since there are significant advantages to self-grading in memory systems: notes.andymatuschak.org/z7gWUD4AnndX5C I’m more optimistic about its application a relative signal of interest/curiosity/value.

notes.andymatuschak.org

Self-grading vs. machine-grading in spaced repetition memory systems

4:23 AM · Aug 6, 2020TweetDeck

Likes