What is Hackers' Pub?

Hackers' Pub is a place for software engineers to share their knowledge and experience with each other. It's also an ActivityPub-enabled social network, so you can follow your favorite hackers in the fediverse and get their latest posts in your feed.

0
0
1

photo of an eink device and handwritten text

day 16

I started writing a very compact little monitor rom for my ilo vm. This isn't quite done yet, but is about halfway finished.

Just for fun, and as a non-directly computer thing, I also sketched off a set of handwritten single stroke glyphs for the ilo instruction set, and started writing out some words using them.

Full logs at charles.childe.rs/DA2025/

a photo of an eink notepad with a handwritten alphabet for the ilo instrution set. This shows the instruction names and the glyphs below them, with a short definition for a forth word below that.
0
0
0
0
0

We've been struggling with a JSR publishing issue for nearly two months now—@fedify/cli and @fedify/testing packages hang indefinitely during the server-side processing stage, blocking our releases. Strangely, the problem doesn't reproduce on a local JSR server at all.

We've opened a GitHub issue to track this: https://github.com/jsr-io/jsr/issues/1238.

Fedify has been a Deno-first, JSR-first project from the start, and we really want to keep it that way. If you've experienced similar issues or have any insights, we'd appreciate your input on the issue.

1
0
0
0

“I think I know how to use passwords securely, so passkeys are annoying and nobody should use them.” is tech-speak for, “I don’t care about the online account security of other people.”, or more succinctly, “I don’t care about other people.”

People overestimate their competence and underestimate the cost of phishing and credential stuffing. Individuals having to learn to use password management software and be vigilant against phishing is an industry failure, not a personal success.

0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
1
0

I'm convinced that it's only a historical accident that we teach people classical logic before intuitionistic logic. CL is terrible.

Everyone is baffled by the CL claim that (A implies B) is equivalent to (A is false or B is true). And they should be, it makes no sense at all. Logic is supposed to be an abstraction of ordinary plausible reasoning, and the equivalence of A→B and ¬A∨B is neither ordinary nor plausible.

Instead IL asks: what do we want A→B to *do*? The answer is simple: If we want to prove B, then knowing A→B should mean that it suffices to prove A instead. That's simply modus ponens, most ordinary and plausible of all logical principles.

And that's exactly what A→B *does* mean in IL, nothing more. How do we prove A→B? By showing that A is false or B is true? No, ridiculous. We do it by demonstrating that we can take a proof of A and convert it into a proof of B—which is precisely what we wanted it for: if we want to prove B it now suffices to prove A instead and then convert the proof.

More generally, CL claims to be concerned with “truth”, whatever that is. Truth is super-complicated, something philosophy has struggled with for thousands of years, and is mixed up with the actual state of the world. But logic shouldn't be concerned with the actual state of the world. It's only trying to formalize ordinary plausible reasoning, not whether snow is white.

IL is much less ambitious. IL judgments are only about provability, which is a much more circumscribed notion than truth. Teaching CL to beginners is like trying to tackle proof theory and model theory at the same time.

0
0
0
0
0
0
1

To be clear, "it's there but you can turn it off" is not enough for me, I already stopped using SwiftKey, Microsoft Windows, Miro, etc on the basis of "we put AI in the software with an option to turn it off". If an org ships a "generative AI" product I want to boycott them and I want everyone to boycott them until they go out of business. I don't want a virus on my hard drive even if it's turned off. But Mozilla's ever-shifting checkbox scavenger hunt has never *resembled* "can easily turn off"

@mcc Once a project includes AI… no, once a project's leaders enthuse about AI, that's when I lose my trust in it. We know it's a monopolist eco disaster of a wrong answer machine. The only sensible move is not to touch it. If they can't see that, what other mistakes are they making and not telling me about?

0
0
1
0
1
0
0

We've been struggling with a JSR publishing issue for nearly two months now—@fedify/cli and @fedify/testing packages hang indefinitely during the server-side processing stage, blocking our releases. Strangely, the problem doesn't reproduce on a local JSR server at all.

We've opened a GitHub issue to track this: https://github.com/jsr-io/jsr/issues/1238.

Fedify has been a Deno-first, JSR-first project from the start, and we really want to keep it that way. If you've experienced similar issues or have any insights, we'd appreciate your input on the issue.

1
0
0
0
0
0
0

To be clear, "it's there but you can turn it off" is not enough for me, I already stopped using SwiftKey, Microsoft Windows, Miro, etc on the basis of "we put AI in the software with an option to turn it off". If an org ships a "generative AI" product I want to boycott them and I want everyone to boycott them until they go out of business. I don't want a virus on my hard drive even if it's turned off. But Mozilla's ever-shifting checkbox scavenger hunt has never *resembled* "can easily turn off"

0
0
0
0
0
0
0
0
0
0
0
0
1
1
0

Mozilla insist that people trust them. I assume they have insight and metrics that reinforce that view.

Perhaps I am getting old and this truly is the only path forward to maintain an independent browser.

I continue to feel that this is a terrible idea that has likely irrevocably jeopardized the future of firefox and, by extension, the open web.

The overhead of disabling every single AI-first change in Firefox is already starting to weigh on firefox-forks.

My view on this hasn't really shifted in the last few months: unless an existing organization, with strong principles, steps forwards and commits to a hard fork I don't really see a future for Firefox.

(I think there are probably only 1-2 orgs with the combination of experience / maturity to actually pull that off, and none of them seem to be even considering that kind of future)

0