jgb1984

You're missing the forest for the trees. I don't care if it's some kind of evil orchestration. That's the least of my worries. What I want is for it to be stopped. And that desire has an overwhelming democratic majority in most European countries AFAIK. Or is democratic decision making only important when it's an outcome you agree with?

It's baffling to me how we come up with economic justifications for forced mass immigration. Mass immigration of mainly non-western low skill / low education people is causing a catastrophic economic burden on western Europe. Unemployment rates of those are much higher than natives, even when you look at 2nd or 3rd generation. Not to mention that most European capitals have become poverty ghettos where the original inhabitants don't recognize their own country any longer. Non-EU foreign origins are overrepresented in crime and prison stats. There are not enough low skilled jobs. There is not enough housing. There is not enough infrastructure. Mass immigration is the most disruptive disaster that was ever forced upon western Europe (forced, because there has been a large democratic majority AGAINST this type of immigration since decades).

Very sad state of affairs. Good on Japan that they see the danger and try to do something about it before it's too late.

I've explained to Claude Opus and Fable how rg is recursive by default so they shouldn't run it with -r. It's in my CLAUDE.md file. It's in a memory too. And still the newest versions of those models make the mistake several times per week.

I couldn't care less what the filename is, I do care however that Claude still forgets about 14 times per week that -r with ripgrep means replace (and not recursive, like with grep). Even though it's literally spelled out in my CLAUDE.md. Claude code, even with the most advanced models, forgets or disregards very simple and clear instructions from my modestly sized MD file.

Following instructions. I've got a modest sized CLAUDE.md containing some simple rules to follow. Things to always do, things to never do. Not a day goes by where Claude Opus violates one or several of the instructions. He keeps making Django multi line template comment bugs. He keeps using -r with ripgrep thinking that means recursive, when actually that's a replacement instruction, he hits that problem several times each day. He sometimes just goes ahead and does a git commit without my approval. All of this is spelled out in CLAUDE.md but he forgets. He apologizes profusely when it happens. Tiring.

I tried opus V5 several times, but it's pretty awful compared to V4.8

Keeps going in circles, it complicates everything much more than it should. And like others have mentioned it just marches on, without questioning, and more often than not in the wrong direction. I'm sticking to opus V4.8

The title could've been shorter: don't use passkeys. Period.

Numbers from Denmark and the Netherlands (the only two European countries where it's allowed to gather such statistics) show that non-EU immigration is a net cost to the society (and economy). In the Netherlands a non-western asylumseeker comes to about 800.000 € to 1.300.000 € net cost to the state over the persons lifetime, depending on what you take into account. And that's purely the financial part, we're not even talking about the increase in crime and the ghettoisation of most western European cities. It's a tragedy, for everyone involved (because most 2nd and 3rd generation non-western immigrants still live a life of poverty in Belgium/Netherlands).

I like claude models, but crush and opencode are miles ahead of claude code. It's a pity anthropic forces us to use inferior tooling (I'm on a "team" plan from work). I can use an API key instead but then I'll blow past 25$ in an hour.

For me wayland offers only downsides, without any upsides. I feel the general idea behind it (pushing all complexity and work onto other layers) is broken. I'll stick to xorg and openbox for many years to come.

Based on my experiments with anthropic, openai and gemini, your codebase must be an unsalvageable nightmare by now.

I sure hope another winter is on the way, the LLM hype wagon is beyond exhausting.

Debian: The Universal Operating System, same as the past 25 years.

I did some digging in the issues and PR's of pre-commit, the guy seems to be a major douche. Too bad, because uv is amazing. Might look at an alternative to pre-commit in the future.

You're falling into the false dichotomy that always comes up with these topics: as if the choice is between the cloud and renting rack space while applying your own thermal paste on the CPUs. In reality, for most people, renting dedicated servers is the goldilocks solution (not colocation with your own hardware). You get an incredible amount of power for a very reasonable price, but you don't need to drive to a datacenter to swap out a faulty PSU, the on site engineers take care of that for you. I ordered an extra server today from Hetzner. It was available 90 seconds afterwards. Using their installer I had Ubuntu 24.04 LTS up and running, and with some Ansible playbooks to finish configuration, all in all from the moment of ordering to fully operational was about 10 minutes tops. If I no longer need the server I just cancel it, the billing is per hour these days.

Bang for the buck is unmatched, and none of the endless layers of cloud abstraction getting in the way. A fixed price, predictable, unlimited bandwidth, blazing fast performance. Just you and the server, as it's meant to be. I find it a blissful way to work.

I don't use AI. Problem solved. The software I make will be better off in the long term.

There is nothing python-2 about my python-3 dynamically typed code. I'm pretty confident a majority of new python code is still being written without type hints.

Hell, python type annotations were only introduced in python 3.5, the language was 24 years old by then! So no, the way I write python is the way it was meant to be, type hints are the gadget that was bolted on when the language was already fully matured, it's pretty ridiculous painting code without type hints as unpythonic, that's the world upside down.

If I wanted to write very verbose typed code I would switch to Go or Rust. My python stays nimble, clean and extremely readable, without type hints.

Couldn't agree more! I've been using Python for almost 20 years, my whole career is built on it, and I never missed typing. Code with type hints is so verbose and unpythonic, making it much harder to read. Quite an annoying evolution.

I was using autojump for years (on debian) until I lost my jump history several times in the past few months. Turns out it's a known race condition bug fixed in a newer version:

https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1110899

Migrated to zoxide instead, seems to work fine! Only need to get used to using z instead of j, muscle memory hard to adjust, might set an alias :)

I've been using Nvidia gpus exclusively on debian linux for the past 20 years, using the binary Nvidia drivers. Rock solid stability and excellent performance. I don't care for Wayland as I plan to stay on Xorg + Openbox for as long as I can.

Specifically: buy a good desktop computer. I couldn't imagine working on a laptop several hours per day (even with an external screen + keyboard + mouse you're still stuck with subpar performance).

No way an AI built tool is getting anywhere near my database.

I'm not using AI for anything. I read and write my own emails, make my own slides, write my own python code using vim, debian, openbox, bash and tmux, just as I have been for almost 20 years. I don't even use an LSP or autocompletion! Hell, I even read actual books, on paper!

And yes, I did test ChatGPT, claude, cursor, aider... They produce subpar code, riddled with subtle and not so subtle bugs, each of my attempts turned out to be a massive waste of time.

LLM is a plague and I wish it had never showed up, the negative effects on so many aspects of the world are numerous and saddening.

I don't understand how you can have the patience to deal with an LLM like that. As if you're dealing with a low skilled but highly stubborn intern. Sounds like an awful waste of time to me.

What a load of crap.

Today we vibecode our software, so the rot is built in from day one!

I'm tired of LLM, period.

Any code that chatgpt or claude ever gave me was overly verbose and highly inefficient in the best case, and riddled with subtle or not so subtle bugs in the worst case. Before the average junior dev will even discover those bugs his grasp on the codebase will have fallen so far behind that he won't stand a chance to redeem the project. Software quality is taking a nosedive. Our industry is vibecoding itself off the cliff.

I'm long on no-AI becoming a label of quality, a badge of honor, as people are more exposed to the slop LLMs forces upon them. I want prose, music, photography and movies made by humans, for humans. I want to use software built by developers. The last part in particular because I've seen my share of bug riddled code spewed by the LLM du jour, I'm convinced our industry is vibecoding itself into the abyss.

What worries me more is the steep decline in code quality. The python and javascript output I've seen the supposed best LLM's generate is inefficient, overly verbose and needlessly commented at best, and simply full of bugs at worst. In the best case they're glaringly obvious bugs, in the worst case they're subtle ones that will wreak havoc for a long time before they're eventually discovered, but by then the grasp of the developers on the codebase will have slipped away far enough to prevent them from being compete t enough to solve the bugs.

There is no doubt in my mind that software quality has taken a nosedive everywhere AI has been introduced. Our entire industry is hallucinating its way into a bottomless pit.