Content
@
https://warpcast.com/~/channel/thomas
0 reply
0 recast
0 reaction
Thomas
@aviationdoctor.eth
We’re heading toward a Library of Babel scenario where virtually every permutation of text, video, and audio will either pre-exist online (Borges style) or be artificially generated on the fly (as on https://libraryofbabel.info/) to drive engagement for cheap. Which means that every most brilliant piece of insight will be out there; but they will be drowned out by vastly many more orders of magnitude of trite platitudes. Like in the Library of Babel, no index exists to tell them apart. There is no “key” to finding the needles of greatness in the haystack of slop. Brute forcing discovery would take literal eons, as in Borges’ story. In turn, insightful dwellers of the internet like @riotgoools give up, log off, and go touch grass, leaving ever more of the space to be filled by bots that not only author the content, but also pretend to be consuming and engaging with it. It’s the most recent embodiment of the Dead Internet Theory. How do we course correct from here? 1/3
3 replies
6 recasts
36 reactions
Thomas
@aviationdoctor.eth
Some thoughts: a) Cryptography-enabled *proof of humanity* / verifiable Sybil resistance for *users* (we’ll eventually need this on Farcaster, Reddit, X, etc). Bots might be tolerated, even welcome, but we need to tell humans apart. b) Cryptography-enabled *proof of authenticity* for human-generated *content* (e.g., @faust’s ROC for photos/videos). We need to know what to trust without trust assumptions. c) Cryptography-enabled *proof of insight* (e.g., can we use LLMs *in reverse* to rate both human- and AI-generated content on a freshness / originality / uniqueness scale relative to all other pre-existing content, and promote the highest-rated output — because eventually, AI will also be capable of producing truly clever and novel insights that we will want to read). All crypto(graphically) enabled technologies. Any others? 2/3
2 replies
0 recast
7 reactions
Thomas
@aviationdoctor.eth
I know items a) and b) above already exist, and “just” need to reach adoption at scale (I’m aware “just” is doing a lot of work here). I’m more interested in item c). Think of the linked Library of Babel website linked above. It contains the proof to the Riemann hypothesis in prose form somewhere among its 10^4677 dynamic pages; we just don’t know where. We also can’t brute force it (the universe has a mere 10^80 atoms). But on the much more reasonable scale of the internet (incl. between 10^10 and 10^11 pages of web pages, an increasing share of which is AI-generated, and social media posts) — can we unleash AI to hunt for gems? Like a web crawler but one which looks for insightfulness instead of keywords? Curious to hear (human) thoughts. 3/3
2 replies
0 recast
2 reactions
Blinky Stitt
@flashprofits.eth
Have you seen the skibwhatever toilet video? We there already 😆
1 reply
0 recast
1 reaction
