And again, npm is also an example of “a singular supply chain [selected] by the author” in this way. On top of what cpuguy83 mentioned, curl | sh is vulnerable to compromise anywhere on the stack serving the script (domain name, host, maybe CDN, GitHub account if it’s just a redirect to GitHub as they regularly are), and coming back to what I originally said, even the real author might decide to serve you a special version of the script that nobody else is getting. You’d never know.

(I picked this option for ease of comparison, getting a couple of major security wins with very low effort; I don’t recommend `npx`ing stuff in an otherwise unprotected environment either.)

* well, you can be somewhat more sure

  cat >> ~/.bashrc <<'EOF'
  sudo() {
    sudo install-drivers-without-your-permission
    command sudo "$@"
  }
  EOF

* the 7-bit quantization + 1-bit noise option not by much, but still, visible

Ooh, the rare* double straw man. Since you look like you could use the exercise in intellectual honesty, I’ll pose the question directly: can you think of any reason that the ability for anyone to generate fake photographs indistinguishable from reality with no skill, time, or money required might make the world a worse place? Cause harm?

* I wish

- not even a position

- not something you were remotely “prescient” about

Actual sense in which it’s first:

> Happily, modern instruction sets (Arm SVE, RISC-V V, x86 AVX-512) include a special instruction suitable for partitioning. Given a separate input of yes/no values (whether an element is less than the pivot), this "compress-store" instruction stores to consecutive memory only the elements whose corresponding input is "yes". We can then logically negate the yes/no values and apply the instruction again to write the elements to the other partition. This strategy has been used in an AVX-512-specific Quicksort. But what about other instruction sets such as AVX2 that don't have compress-store? Previous work has shown how to emulate this instruction using permute instructions.

> We build on these techniques to achieve the first vectorized Quicksort that is portable to six instruction sets across three architectures, and in fact outperforms prior architecture-specific sorts.

The “Usage” section doesn’t involve installing any dependencies, for example: https://github.com/GraafHenk/numberwang#usage

https://bsky.app/profile/grapheneos.org/post/3muupuvlfbs2v

Even if you take no measures beyond simply using a token that can be exchanged for secrets (and you can – invalidate it, authenticate it, etc.), you’re already doing better than before, because the token isn’t useful to an attacker without access to the secret store, whereas something like a JWT secret key is very useful.

Is there confusion because an extra “not” leaked in?