Encryption Is Not a Crime

As an example, the civil rights act necessarily curtails the freedom of association.

https://www.nysenate.gov/legislation/bills/2025/A2228

a) This seems like a decent introduction to the subject of cryptographic regulation in the last 30 years. It's far from exhaustive, however. I do appreciate the collected references from diverse points in the last several decades.

b) I would have mentioned "Sink Clipper" and the ACLU "dotRights" campaigns. Neither are especially easy to find in the increasingly enshittified google cache, but Le Monde Diplomatique has this article, complete with a link to Sink Clipper poster (I think from the mind of Kurt Stammberger) that no collection of CypherPunk oriented ephemera from the era can be without: https://mondediplo.com/openpage/selling-your-secrets

The ACLU dotRights.org site seems to have receded into history, but some of it's content is still available at the archive. For example: https://web.archive.org/web/20100126102126/http://dotrights....

c) Herb Lin presented a very nice paper back in the day comparing PROPOSED encryption regulation with ACTUAL encryption regulation. I think the thesis was through the 90s, proposed regulation was increasingly draconian (clipper, etc.) but actual regulation was liberalizing (effective deregulation of open-source tools.) I found Herb's page at Stanford and heartily recommend it if for no other reason than it's sheer volume of written material: https://herblin.stanford.edu/recent-publications/recent-publ...

d) I was a little surprised the wired article linked to at the beginning of the piece didn't have that issue's front cover, which was sort of a cultural touchstone at the time. But you can see it here: https://pluralistic.net/2022/03/27/the-best-defense-against-... - and this one: https://www.reddit.com/r/Bitcoin/comments/1cgpktp/31_years_a... (dang, look at those non-receding hairlines!)

e) Making the web "secure" or "private" is like putting lipstick on a pig. Modern web technology is designed to de-anonymize and collect identifying information to enable targeted ad delivery. Thought I generally respect Moxie Marlinspike and have no great beef with Signal, there has been a concerted effort to exploit its device sharing protocol and your carrier and national governments can easily extract traffic analysis info from people using it. Were I to add one sentence to this guide, it would be "While these tools are better than nothing, they are far from perfect."

f) The guide seems to conflate encryption with privacy. Encryption technology can enable privacy, but you're not going to get privacy from encryption technology unless you pair it with well reasoned policy (for organizations) and operational guidelines (for both organizations and individuals.)

The extreme example is to say "nothing stops a participant in an encrypted communication from sharing the un-encrypted plaintext after it's recovered." People earnestly trying to maintain message security probably know not to do that, but when talking about exchanging keys and figuring out which keys or organizations you should trust, it's easy for even the well-informed to make privacy-eroding decisions.

So... I think this article is a good jumping off point, covering material I would call "required, but not sufficient." I would just view it as the beginning of a deep-dive instead of the end.

More importantly, the thing you're asking for (law enforcement retroactively snooping without there existing a master key) is always impossible.

For other forms of snooping (like a warrant to tap communications for a single device for a period of time), you have related issues. Suppose you magically make such a thing flawless -- the client can't detect intrusion, a single key is actually time-bound, etc. There still exists a group of people with the power to hand out such keys, and that power, however it's implemented, is still a master key to all future communications over that protocol.

You can partially mitigate the risk in various ways, but you can't eliminate it. Every proposal for weakening cryptography in that way has had glaring flaws, and many known attempts at actually weakening it have later been cracked by nefarious actors. Spying, but only for the "good guys," should be met with extreme skepticism as far as cryptographic protocols are concerned.

For all of these schemes, what happens when the people holding keys and power are physically forced out (DOGE et al)? Even if we assume the thing is implemented flawlessly, the people involved never leak anything, the master keys stay secret, ..., you still have the human problem of transitions in power. Do you want the current US administration, one currently arguing that it can "deport" actual citizens to torture prisons with no recourse or court case, to know that six years ago your daughter confided to her best friend that she got an abortion once? That she doesn't believe Israel should be committing genocide? Or, suppose you approve of the current administration, what about the next one that takes the reins with this new set of powers? It's bad enough without decades of chat history to let 70%-accurate AI comb through and make deportation decisions.