markjenkinswpg

This is probably not a convincing argument, but I'll share it anyway.

Think of bootstrapable builds as a hedge against tyranny.

Is it likely that a future regime could come into being that would be backdooring FLOSS builds and undermining confidence in the entire ecosystem? Not at all likely, as this requires a purge of all the install/boot media already out there.

Though after the passage of decades, the chain of custody starts to weaken. Maybe I'll still believe in the magic of the CD in my retro closet with the three humans holding hands in a circle, but in an environment where children are told to report the thoughtcrimes of their elders, how can my nephews and nieces in turn trust me and my testimony that my copy is the "good stuff" from "the before times".

This is all very far fetched sci-fi material, but think in terms of tail risk, low probability multiplied by a high cost if comes to pass.

Perhaps a regime horrible enough to break FLOSS would be horrible enough to purge all legacy hardware and software sitting in closets, but I do think it is easier to mess up the internet and the social fabric than it is to do house to house cleanouts.

Maybe messing around with untainted FLOSS will be too dangerous in such an environment anyway, but if we can assume a state of liberty eventually returns, having a great bootstrap ready to go can help rebuild society, cyberspace, and trust on an accelerated basis.

All of this hinges on people actually auditing source code, though once you've read enough bootstrapped source to have confidence in the generation of checksums you can at least start to fall back on a social consensus as to the safety of additional source archives that have known checksums and from there also digital signatures once you've built the tools for validating those.

When it comes to "trust, but verify" of other people's source code, there is a new tool in the toolbox: large language models have demonstrated some considerable audit value. Of course almost nobody is training their own models and it's only a slightly larger circle that does inference on their own hardware or attested remote hardware, but all of this is at least something that can increase confidence when working from other folks source code.

-------

As far fetched as full on cyber tyranny sounds, its worth remembering that in North Korea this is actually reality. Folks do not have access to the internet and are handed the state built binary of Red Star OS.

The ballooning of flash drives containing western media is still a thing despite attempts to stop it.

It is unlikely that anyone will ever send in a software bootstrap, that it will get in the hands of someone who can understand it, have the time to read it and build it.

But the scenario does make us think about the value of bootstrapable FLOSS as a tyranny hedge.

Without real internet access, what good does having some freedom on a computer even do? I suppose one can keep a diary encrypted, though possessing an encrypted file is probably more of a liability than any value it provides. Someone may also find encryption helpful for friend to friend networks by sneakernet, though the downsides are considerable.

Someway, somehow the regime may fall. After decades of not trusting the outside world and being governed by a hostile state, perhaps there will be at least one deeply computer literate person to emerge from that place who will appreciate being able to bootstrap their own personal cyber citadel.

To describe "Predictable firmware randomness can expose wallet seeds and ephemeral private keys" as severity "high" and not "critical" is a severe understatement.

If I machine model generated this, it shows a lack of understanding this has on people.

One reason for waves is that end-users had the option to add additional entropy via dice rolls mixed in with the bad internals or an arbitrary BIP39 passphrase mixed in after the initial 12-24 word seed phrase is output.

With the level of paranoia culture around the Cold Card, it wouldn't have been too uncommon for folks to take one or more of these additional measures.

How much effective entropy someone adds these ways would vary. When you're told the initial entropy is solid, what's the point of adding too much of your own?

Long passphrases on top of seeds are also seen as a good way to cause a loss by forgetting and also a usability PITA that takes away some of the convenience of a hardware wallet. Some folks no doubt opted for short BIP39 passphrases just as a means of protecting against a situation where someone else in their household stumbles upon a paper or metal backup of a seed but doesn't know anything about the cracking required to try BIP39 passphrases.

Main reason these extra sources of entropy would be grabbed in waves is different bad actors coming along and making the extra effort after the disclosure.

The website cktripwire.com has a nice dashboard of honeypot bitcoin put out there with the extra entropy. So far only low amounts of additional entropy like 1-5 dice rolls or one extra passphrase word have been swept, though this honeypot was published after additional waves. More will certainly swept in the end.

From my perspective, the bug became shallow because the machine world extended the number of effective eyes. The aphorism applied, but just not fast enough as it took too long to get the right number of eyes.

More people would have been harmed had the bug been in production longer.

One unfortunate aspect is that responsible disclosure would have only saved some people, a responsible disclosure would have lead to a notice to migrate seeds and bad actors would have immediately realized seed generation was the root cause and would have swept many folks before they could even see the notice for themselves and get around to migrating.

One comedy that did a good job of depicting programmers with no sense of hope circa 1999 was Office Space.

Interesting summary. The relationship with some Arab speaking states has had some recent relevance.

Consider this. There are circumstances in Venezuela that some would consider worthy of a civil war. This award winner has chosen peaceful resistance, acts that may have prevented war.

In my experience, the 10-15 year old salvaged computer that still works okay with GNU/Linux is increasingly a 64 bit machine.

Case in point, I'm writing on a x86_64 laptop that was a free give away to me about a year ago with a CPU release year that is 2012.

I have personally given away a x86_64 desktop unit years ago that was even older, might have had DDR1 memory.

Circa 2013 my old company was gifted a x86_64 motherboard with DDR2 memory that ended up serving as our in-office server for many years. We maxed the RAM (8GB) and at some point bought a CPU upgrade on ebay that gave us hardware virtualization extensions.

This is correct. But it can talk in their ear and be a good sycophant while they attend.

For a Star Wars anology, remember that the most important thing that happened to Anikin at the opera in EP III was what was being said to him while he was there.

I'm curious if Wine and ReactOS can run this, sounds like a good stress test.

Not only does the government have the right to destroy it, but so does anyone else.

There's no reasonable expectation of recovery when you launch a balloon calibrated to go globe-trotting.

It's basically no different than the person who abandoned a mattress at my apartment block's dumpster. It's littering. They're giving it away, they're not expecting to be able to come back for it later.

Furthermore, nobody has a right to unregistered, unmanned, long duration balloon flight. Such a right would only exist if a state constructed it for its citizens. (and it would end at their boundaries)

I think it's cool that in practice people have been able to do so, but it seems the cool times are coming to an end.

Switching.