- Thanks for the information.
This sounds like something I would use HKDF for. But, to your point, it's nice to be able to build the design with a fewer number of primitives, and likely more performant, too.
- This particular piece has been shared near me several times, in the context of this recent AWS outage, the previous big AWS outage, non-AWS outages, and others. Every time, I feel like I'm in vague agreement with the author, and at the same time, none of it is the least bit actionable. Even if Cook is correct, so what? There's no concrete change I can make in my working.
- The edit window passed, so let me add: where and why one would use an extensible-output function in particular.
- The nice thing about Blum-Blum-Shub or Blum-Micali is that they come with a proof of security. Even then, they tend to be impractical, due to performance and side channels.
This one is missing the most important part, the proof. Indeed, a sibling comment notes that empirical results look pretty flawed.
- I freely admit, this is "smell" based speculation.
- Can anyone share real-world examples of where and why one would use these, please?
- This smells like self-dealing on the part of A16z.
- Is there a parser ambiguity/confusion vector here?
- I thought the same. Although, perhaps we have too few hatchways, and too much surface area inside each.
- After a few dumb accidents involving header pins, I've come to the conclusion that exposed male header pins on my desk are a hazard.
- As this is a tech forum, I expect that the name that stands out is Radio Shack, and I do agree that that story is tragic. For me, as a long time New Yorker, the other name that stands out is Modell's. They were, for quite a long time, a fixture in New York City and beyond. Even for us not sports inclined, we still knew Modell's as a reliable retailer of sneakers throughout the years. Arguably, their Covid era bankruptcy was already the end, and this was just private equity puppeting the corpse. Nevertheless, it's a sad and nostalgic story.
- Kinda tricky. Last time this came up, the consensus was that approximately nothing commercially available supported RVA23 at the time.
- The unemployment rate statistic is a bit misleading, because people who drop out of work permanently don't factor into it.
- That's surprising. Why is there such a comparatively large number using 32-bit Firefox on 64-bit Windows?
- In most enterprises, the choice isn't Github vs Buildkite, it's Github vs Github plus Buildkite. That's what makes it so hard to pay for a separate CI vendor that costs more, when your source code hosting vendor already bundles one, as good or as bad as it might be.
- I evaluated Buildkite at a previous job, and I came to these conclusions.
1. Buildkite is probably the best commercial, off-the-shelf CI system right now, in terms of providing all the correct building blocks at the correct level of abstraction. 2. The impact of your CI system itself being good or bad is tiny in comparison to everything else in your end-to-end CI workflow. Far more important are your own CI scripts and what they run. A distant second is the observability tooling around your CI. 3. It's hard to justify the per-seat pricing of Buildkite, as a separate line item, when whatever CI offering your source control host bundles in will suffice.
- If you do it correctly, you've reinvented Fisher-Yates[1]. If you do it wrong, you've reinvented this unnamed, broken algorithm[2], instead.
But the issue in the article isn't application of pseudorandom numbers. It's seeding the generator.
[1]: https://en.wikipedia.org/wiki/Fisher%E2%80%93Yates_shuffle [2]: https://en.wikipedia.org/wiki/Fisher%E2%80%93Yates_shuffle#N...
- Juggling all the fragments inside the database, garbage collecting all the unused ones, and maintaining consistency are all quite challenging in this use case.
- One of these is not like the other. Go is also garbage collected. Embedding a garbage collected language inside another means you have two garbage collectors fighting each other.
[1]: https://xkcd.com/1172/ [2]: https://xkcd.com/1053/