_jackdk_

This argument applies to learning to code in general. What is your estimate of the ratio between "just shipped whatever the clanker spat out" vs "interrogated the LLM about its own output†" vs "deliberately interrogates LLMs to learn"?

†Untrustworthy, because it will either double down on its own output in its context, or will flip-flop to whatever the user said .

As in the Steve Jackson one where you have conspiracies and front groups and fronts for those fronts? I wanted to like it but never could get into it; too many fiddly bits placing megabucks all over the power structure and adding up modifiers. Did you play house rules with friends, or does it get better with more plays, or was I just doing something wrong?

Unfortunately that's accurate. From the SNB whitepaper linked at https://www.taler-systems.com/en/digital-currency.html you get to https://www.snb.ch/de/publications/research/working-papers/2... and then to https://www.snb.ch/public/asset/de/www-snb-ch/publications/r... :

> In addition, charging an exchange fee would allow the central bank to implement negative interest rates, if deemed necessary. The central bank could also impose a conversion limit per customer for AML/CFT (“cash” limits) or financial stability reasons (to prevent hoarding or bank runs), if desired.

Well goody.

The idea of blind signatures is cool, as is the fact that the merchant doesn't need to identify the customer, as is the fact that you could deploy neat local currencies with something like this, but AFAIK you still can't give 20€ to your friend without the exchange knowing and approving, or it becoming a race to see who spends the tokens corresponding to the 20€. And that's not very privacy-preserving.

Speaking of page 3, Google's been asking me to sign in if I try to go beyond page 2 of search results. What a wonderful modern WWW we live in.

I agree with your concerns but also think that a new viable browser engine will be such a benefit to the health of the open web that I'm willing to overlook such irregularities, at least for now. sqlite doesn't accept contributions either and it's done well for itself, and if Ladybird ends up going sour there's always the possibility of a fork. This is similar to my stance towards Firefox: despite the actions of Mozilla Corp, having at least one non-Blink engine is too important, and for friends and family that I care about, working ad-blockers are essential so there's no practical alternative.

Firefox is great, and I use and recommend it as daily driver, but the actions of Mozilla leadership are really hard to differentiate from deliberately managed decline. It will be interesting to see what happens as Ladybird matures.

I did, yes. A but ironic indeed.

I wonder if at any point during writing, the heroic author of that Super Metroid guide learned that "missiles" are not spent "missles", and decided to own the mistake rather than reword the entire guide.

Also, the font used for the monospace examples in the post is a great kick to the nostalgia. Which one is it? It reminds me of https://int10h.org/oldschool-pc-fonts/fontlist/font?ibm_vga_...

The blog link I mentioned avoids this cost with literals, by providing using a required type argument to check the string length at compile time without TH. It requires a relatively recent GHC:

    make :: forall symbol -> (IsNonEmptySymbol symbol) => NonEmptyText

    type family IsNonEmptySymbol symbol :: Constraint where
      IsNonEmptySymbol "" = Unsatisfiable (Text "Expected a non-empty string")
      IsNonEmptySymbol _ = (()::Constraint) -- empty constraint is always satisfied

Liquid Haskell might require you to prove that the divisor is nonzero, but even in standard Haskell there's common idioms for ensuring that a list is non-empty (data NonEmpty a = a :| [a]) or that text is non-empty (newtype NonEmptyText = NonEmptyText Text, with non-exported constructor, helpers like make :: Text -> NonEmptyText, or more advanced tricks like https://exploring-better-ways.bellroy.com/haskell-koan-type-... ).

The big problem preventing this approach from working for numbers is that it's just so cumbersome there. Most of this is because all the arithmetic operators are bundled into a single Num typeclass, and `fromInteger :: Num a => Integer -> a` has a type that's impossible for a "non-zero number" wrapper to satisfy.

Given the passing of nitter, xcancel, etc., thank you for quoting the relevant text.

It helped that Apple had a page for their statutory warranty in AU, so I could point to the sticker price of both machines and say "look, Apple acknowledges the consumer guarantee for at least 24 months after purchase and possibly longer, and I have paid a similar amount of money for a Framework".

https://www.apple.com/legal/warranty/docs/statutory_warranty... (Page 2):

> Without limiting consumers rights, Apple will provide its own remedies equivalent to those remedies in the consumer guarantee provisions of the Australian Consumer Law at any time within 24 months of the date of purchase. For the avoidance of doubt, Apple acknowledges that the Australian Consumer Law may provide for remedies beyond 24 months for a number of its products.

What fun. War stories like this are what keep me coming back to HN.

Yeah, the ACL is fantastic. I have an AMD Framework 13 that had a fan bearing fail, which caused the fan to scream and spin up/down erratically. I had a long back-and-forth with Framwework support where they told me what I already knew (the fan bearing failed and the fan needed replacing) and that it was out of warranty (yeah nah, not in Australia). As soon as I mentioned the consumer guarantees under the ACL, I was invited to remove the SSD and ship the unit to their service centre for a fix.

The machine has been fine since and I still really like it.

The 8088 MPH demo, for example, pushed the PC Speaker pretty far: https://www.youtube.com/watch?v=yHXx3orN35Y

I haven't seen a https://wiki.c2.com/?SmugLispWeenie in the wild for a very long time. I thought they'd all gone extinct!

Because Fastmail is one of the relatively rare non-American companies that's managed to get a good position in a challenging market. The UK's "Snooper's Charter" also has a mechanism to send Technical Capability Notices that require tech companies to build things to spy on their users.

And how, exactly, will one acquire this "full EU Digital Identity Wallet"? Will I be able to compile it from source and run it on a computing device of my own choosing?

Thanks for replying, that's an interesting architecture. The idea of using a high-up parallel state seems particularly neat as a way to ensure every part of the page render stays independent, and to burst up the compute serving the response.

That is not a use cases I would've expected Step Functions to handle; I'm intrigued. You're using express ones, I assume, but how are you wiring them up and why?

Single Instruction, Multiple Data.

So did consumer advocacy sites; this one is slop.

Perhaps coincidentally, it was also around the same brief beautiful moment in time where one could talk with all of one's friends using a single messaging client, possibly even over federated XMPP.

How much do we trust those guarantees? Apple currently alleges that OpenAI employees have been stealing Apple's trade secrets: https://9to5mac.com/2026/07/10/apple-sues-openai-trade-secre...

> [Mr. Tan] has directed job candidates still working for Apple to bring “Actual parts” from Apple to their interviews for “show and tell” sessions in which he and his team at OpenAI can elicit still more Apple confidential information.

> As part of its investigation, Apple found a “pattern by employees who depart for OpenAI of taking steps to evade the security processes intended to protect Apple’s confidential information.”

> Apple also claims former engineer Liu exploited a security bug to download confidential engineering files after leaving the company. Rather than report the exploit, Liu allegedly joked about it in messages (“LOL,” “so funny”). Liu also failed to return an Apple-issued laptop after his departure.

This seems pretty close to "they trust me, dumb fucks" behaviour.

USB-C maximalism is great for travel. I personally recommend travelling with a USB-C desktop charger, but use one that accepts a IEC C7 ("figure-8") cable. Then, travel with only the cable corresponding to your destination (or buy one when you get there). That avoids having a wall wart that might not fit in narrow spaces, or whose weight might make it fall out of older sockets.

I wrote about it a few years ago; the post does not have affiliate links: http://jackkelly.name/blog/archives/2024/10/06/travel_tip_us...

This is also something to be aware of when teaching people, too. I've seen advice for designing Anki-style flashcard decks that reminds people to create flashcards for both A->B and B->A.

That'd be https://www.costarastrology.com/why-haskell, if I remember the job ad correctly.

It's such a shame because back then even Windows was motivated by actual human user research and had thorough guidelines. https://ics.uci.edu/~kobsa/courses/ICS104/course-notes/Micro... (HN discussion from a few years ago: https://news.ycombinator.com/item?id=22475521 .)

https://movq.de/blog/postings/2026-06-16/0/POSTING-en.html was an interesting look at Win2K's UI controls and how much clearer they are compared to modern UIs.

Qantas gave its final 747s a nice send-off too, back in July 2020, flying its final commercial 747 flight as QF7474.

https://www.qantas747.com/

They also had QF7474 trace the Qantas logo in its track along the way, which is so cool and I am surprised I couldn't find it in the official site:

https://www.escape.com.au/destinations/australia/how-qantas-...

The problem is that Wayland just isn't a compelling alternative for many people, so they don't move. For me, I see no benefit because I got used to avoiding HiDPI and don't have a mixed-DPI workspace. For some bizarre reason they made each compositor implement input-handling separately, so for example artists might have to switch compositor just to use their tablet of choice. And worse yet, some people with input accessibility needs are just going to get straight up locked out of libre computing. See https://nocoffei.com/?p=451 for one example.