xlii

And there is a super power to it.

It's possible to write the whole system only by defining the types.

The "glue" can be sloppy but as long as it keeps on the edges the output is most often fine.

Recently I'm on the fence about Rust vs OCaml (but plan to write about it soon) because I have ~700k LoC in Rust but my workflow starts to get seriously dragged down by compilation/tests in isolated worktrees.

I recently also dab with Gluon (as embeddable type safe scripting) and rule-based-development for maximum code control/agents output leverage.

I don't see how this is better, my first thoughts: more difficult to print, more difficult to scan, vulnerable to light shifts (imagine trying to scan a can with colored bar code in a club with flashing lights) and impossible to print in some technolgies.

Bar code won't go away exact for the reason it's super simple to scan and print. Even QR codes can't stand that (and if you tried scanning yourself grocery in soft bag with barcode in the corner which gets crumpled you know what I mean).

I solved exactly that problem I and I sell it to those willing: https://race-term.com

It is multiplexer so survives app kills, and interface is heavily keyboard driven.

I tried running email in it though resigned (and guess what, wrote my own) because HTML email support is must have nowadays.

Competition! :) (though mine is GUI with custom renderer)

https://github.com/exlee/ascdraw

Upvoting. I quit smoking many years ago with that book. Cost me like $8.

It felt surreal to quit after 1 last smoke as book promised at the beginning. Hadn't had craving since then.

Recommended to many people. For some it helped for some it didn't though.

I'm working on 2 things right now (on side). Both commercial (sorry!).

First one is RACE terminal [0]. It's a put-terminals on infinite canvas terminal multiplexer. Right now on macOS, but soon I expect to have Linux version fully released.

Second is Inboxctl [1], it's a keyboard-driven mail client for power users.

The cool thing about both of them is that RACE is my terminal multiplexer and Inboxctl is my e-mail client and they work great. I'm really proud that I can fully interact with terminals/e-mail using my own creations. Both are based on the GUI-framework I accidentally made for ascdraw [2] and kakvide [3].

[0]: https://race-term.com [1]: https://inboxctl.com [2]: https://github.com/exlee/ascdraw [3]: https://github.com/exlee/kakvide

Mind-boggling.

A joke comes to my mind:

A man tells his friend "You know, I don't think we can be friends anymore. Some money went missing from our house.." - "What?! Me? I'd never seal money from your house!" - "Yeah, we found that money, so that's not a problem. But you know... the whole thing left a sour taste."

Yes, and since multiple people expressed interest in Linux it might come much sooner than later.

I kept macOS surface isolated and constrained on purpose (and also have my own rendering engine) so it should be rather easy to make it run on Linux. The biggest obstacle is that it's hard for me to test (I have a gaming rig on Linux, but I don't use it often lately..).

No. Only got "account reactivated" e-mail.

If you decided you like it enough to get a license code HNMAGIC will get you 50% off (valid for 24 hours for anyone to pick :))

I made an update, but I'll post also here in comments. It seems that either a human read it or it triggered some trigger, but my account has been reinstated. Thank you, also those who made this issue visible.

On one side I'm happy it happened, on the other, it's a shame that it took complaining on the internet amplified by HackerNews to get this problem fixed...

Lol. Haven't thought about that. Now I'll have to ponder shouldn't I add another domain to the mix.

My Show HN failed. Might resubmit some time later, but HN (in general) is a lucky draw. And also I figured that trying to sell a indie software might be a good idea for own business and HN can't help with getting enough potential customers :)

This is my driver since it become usable. I have 140 terminals on 2 workspaces. Grouped by project/task. I know, I'm the author, so I'm biased, but I love working using it.

I have aphantasia in both visual and audio forms. And I can tell you audio aphantasia is wild.

I cannot hear a song in my head if my life depended on it but then there are triggers that extract it music it.

Examples: "who let the dogs out" (I remember rhythm but not the music). I play piano for almost 2 years but I have no idea what's the each sound is. It gets even better. Even though I play and record [0] I can't remember my own compositions. They usually come back to me as a play on hands and not on audio.

I thought I was music deaf for the most of my life. My music teachers mocked me in my youth showing me as an example of someone with zero hearing.

My latest music teacher discovered that I actually have absolute hearing skill but... can't recall sound in my head. So I sense even slight shift in toning (e.g. given few variations of same melody I can tell which one is the original tuning) but I can't reproduce it.

[0]: https://on.soundcloud.com/Rv6t1jXzyAH8E2jEaY

More TUIs, and with eye candies too! :)

Since I wrote my own terminal multiplexer that puts terminals on infinite zoomable canvas [0] and I just can't get enough of putting weird stuff like asciiquarium, cmatrix or tty-clock all around (tried peaclock but it's slightly too fiddly for my taste). There are fun projects just to run in TUIs like astroterm or weathr, but I'd say that the space is still open to terminal toys.

The only problem I have with TUIs is that they're boring after a while but something that can be solved with often theme switching.

[0]: https://race-term.com (commercial)

I know that's a skill that I have, and I used it so many times it became effortless, but it all goes to one of jokes I like - that asks how do you eat an elephant. Easy - with a fork.

So in that particular case I have a keyboard and it's misbehaving. I know that by some dark magic I can make it behave as I want and that software on Windows can do that.

So I think, "ok, how does it communicate?" - with Bluetooth, obviously. So my next thought it: can I transmit over bluetooth magic bytes that do exactly like this. I definitely have a device on my computer, so I can try to make it.

From that it's all about "how to send arbitrary stuff from bluetooth" search. Instinct tells me there has to be a library, or that at least on person on world wide web tried to do something similar.

If I find even single instance of this, I'm home - read what they did, try to figure out if I can adjust it and do so. Hocus pocus - binary switcher happens.

Though, in the end, it's all about approach. I found many people say "I don't know how to do it, so I don't". My approach is "I rarely know what do, but that doesn't stop me from trying". And this trying new stuff brings practice for the next unknown endeavor.

> But do we still trust the open source software that is created now everything is vibe coded?

If I think tool/software would be useful I throw GLM 5.2 (now 5.3) at it and ask to do an audit ;-)

2 anecdotes about Logitech software:

I had a high-end webcam that flickered. There should be switch in the software to switch from 50hz to 60hz that'd fix it but it was nowhere to be found. After digging deeper I found that it was only available in English version of software. It was before macOS had Setting to set language per app, so I had to manually replace Polish bundle with English one through binary.

Another one: Has K810 keyboard that started in "FN-by-default" mode. K811, Mac version of said keyboard allowed for changing default setting just as K810 on Windows, but K810 on Mac - no dice. Support told me to shell out 200€ on Mac keyboard so I ended up writing small binary switcher [0] that would run after keyboard got connected [0].

[0]: https://github.com/exlee/k810_fkeys_mac

Try "Tiny Habits" by BJ Fogg. I found it quite interesting and I implemented few habits by using it.

Though there's ADHD and other neurodivergent scenarios, and when you have it making anything of a habit is a hard work.

The burnout is a separate topic, and it was always difficult to balance in this specific industry.

All these "Working in IT isn't stressful at all - John 25yo" memes didn't come come from nothing.

But I wouldn't really put it on LLMs/AI. It was always like this. Jokes about how hopeless development cycle is, and how everybody have impostor syndrome etc.

Long long time ago people in tech were in tech because of the tech.

Then, at some point, being in tech became a gilded profession and everyone: salesmen, artists, clerks etc. joined in. Not because they wanted to be in tech.

They wanted gold.

For years it was working, but right now things shifted. You can no longer just do the thing you learned for a job and continue doing so, because landscape shifted. So they're miserable and they want to get back to knitting, pottery or whatever thing they want to do but not tech.

For people I know who are in tech for the tech nothing changed. There's new tech in town, one has to adjust, and it's still possible to discover new things and pave new paths. Tech was always shifting and I suppose people who worked in it for it always knew it, and it doesn't derail them.

I implemented few agent harnesses (and rik! advertising time: https://rik.axk.sh), and once doing that I noticed one thing:

Context-less self-approval is working well. The failure mode is usually false positives (i.e. safe commands being rejected), not the other way around, with root cause of requesting agent underspecifying context (e.g. not mentioning in the request that it's made on behalf of user etc.)

Thus, I'm running self-approval YOLO modes on state-of-the-art models for quite some time and it didn't bit me. It might, but hey, we're long gone from the age of predictable software development.

I start to sound like a Jujutsu fanboy but...

Jujutsu already does this. I had agent screwing up changeset over couple hour long session and it took one shell command to split on pieces with longer pause.

The only downside is that because of Jujutsu snapshots it can't be used inside sandbox (since all jj commands are mutating)

I consider myself senior engineer. When talking with junior colleagues they often are surprised how little I care about some things and how much I care about others.

These internal "attention weights" are highly influential parameters of how I work with LLM.

E.g. when working with Rust I often hold strict control over structures and lifetimes. But when lately I've been doing token-based bind generation I didn't care about anything outside of high level patterns like RAII and ultimately - API ergonomics which was verified in consumer app.

I've been in position of porting real-code to vibe-code platform and seeing non-technical people prompt-stream (they were shared across accounts) I know why they engaged engineer to run this work.

Their efforts took 6 weeks, I ported app within 4 days and (to be honest with myself) without LLM I that'd be 3M+ work pre-LLM.

In short: I observed same effect as claimed.

Ccn't say, though since Opus 5 I'm much more often staring at the wall of text and I have no idea what it's telling me.

WTFs per prompt metric raised sharply for me :)

New models (especially Opus 5, released just last week) produces unintelligible output, so people are trying to fix it.

"Bottled-water vendor position on tap water drinking".

There's an interesting angle to BLIK vs Wero. While I believe QR codes are more convenient it is actually possible to give BLIK over the phone. Which means that youngsters (or fraudsters ;)) have access to their guardians' money over the phone line.