I don’t like the trend towards public revelation of private communication.
I think it is increasing.
I think the abuses will become more dangerous than most yet comprehend.
And I think we need a variety of cultural and technological countermeasures towards such abuses.
Conversation
It’s everything.
It’s China’s digital yuan.
It’s Snowden’s turnkey tyranny.
It’s media corporations that profit from invading your privacy.
It’s tech companies merging with surveillance states.
It’s the rising foreign & domestic conflicts that spur the erosion of civil liberties.
9
48
385
Another idea is to lean into the fact that AI content creation will soon flood the zone with fake versions of everything.
Anything that’s not digitally signed is deemed fake till proven real.
Important thread on this by below.
Quote Tweet
I cannot decide whether this would be ethical or not 
I think the ends would be net-positive, but would the action erode even-more-net-positive societal norms? Thoughts, y'all? twitter.com/zooko/status/1…
I think the ends would be net-positive, but would the action erode even-more-net-positive societal norms? Thoughts, y'all? twitter.com/zooko/status/1…Show this thread
11
39
263
A third idea is the pseudonymous economy.
Use zero knowledge, homomorphic encryption, secure multiparty computation, cryptocredentials, AI avatars, and the like to bring your minimum necessary self to work.
Thereby make both cancellation and discrimination impossible.
Quote Tweet
Excellent talk by @balajis on The Pseudonymous Economy. High ratio of ideas per minute. youtu.be/urtXRg9Nl3k
18
40
251
A fourth idea in combination with pseudonyms are screenshot resistant fonts.
In combination with disappearing messages and end-to-end encryption, you can make it hard to share screenshots without traceback to the screenshotter.
Quote Tweet
7
13
103
A fifth idea — the obvious one — is to start consciously building high trust communities with strong social conventions in favor of privacy.
See this talk on how to robustify your social network supply chain.
Replying to
A sixth idea is *provable* end-to-end encryption.
You no longer need to take Signal or WhatsApp’s word for it. Instead, by connecting privacy to private keys, we can enable provable E2E messaging. twitter.com/farcaster_xyz/
This Tweet is unavailable.
15
18
156
A seventh concept is AI-based stylometric *re*anonymization.
It runs on your phone. You hit a button a few times to rephrase your text. Preview, then send.
Coupled with AI voice & face changers, pseudonyms, screenshot resistant fonts. Makes you hard to track.
A digital ghost.
Quote Tweet
Style transfer, but for text. Will probably soon be Google Docs plug-ins that will change your sentences into the style of Hemingway, Austen...or Satoshi. Run locally, could help defeat stylometric deanonymization. May be possible to use as a kind of ML-based linter as well. twitter.com/stanfordnlp/st
9
15
156
Replying to
Fiction does work here too: a shared language of strife, struggle, and values builds trust.
Replying to
This is akin to trustful vs zero trust networks. Currently, the software engineering trend is towards zero trust.
High trust communities might not work due to faster dynamic shifts between bith physical and digital locations