mirzap

Read the article first:

> Although engineers use calculators when carrying out arithmetic, they still need to know mathematics. Likewise, developers will have to understand software. We will still need someone to assess and review the implementations produced by an agent. And since the cost of producing code is decreasing, a greater proportion of our work will then be devoted to deciding if it is correct.

I wonder why they didn't try to discover if the user already has those binaries and use them before downloading.

Omarchy ISO is downloaded 20000 times A DAY. It's in the top 3 distros now. Only very sad and delusional people hate its success. I pity them, but I don’t want them anywhere near me.

Ubuntu wouldn't be Ubuntu we know today if Mark hadn't put his own money into it ($10m initially, oh, the irony); the same goes for Omarchy.

Yup, that's why I think they will fork it.

Hyperland being funded for 3 years benefits all distros and users that rely on it. The same goes for Quickshell. And my guess is that this is only the beginning. They will support any project that Omarchy is built on. They even reached out to AUR maintainers to support them financially and build better, more stable infrastructure for any distro that depends on AUR, but they refused.

You realize Omarchy just started a couple of months ago, right? The founder himself put his money and time where his mouth is.

Most of the thousands of distros are just a collection of config changes in GNOME, etc. There is no fundamental change in 99% of Linux distros. It's just different configurations of different tools, with the main differences being the package manager and window manager they use, plus their configs. Everything else is exactly the same.

They just founded the Foundation, and the money has not been allocated yet. But Hyperland, Quickshell, and some others already got 3 year sponsorship deals.

The AUR maintainers refused the cooperation and the money. So they will start working on new package management for Omarchy.

My guess, though, is that most of the money will go to DHH's obsession with Macs and anti-Apple stance (porting Linux to work on M architecture), either by directly funding Asahi Linux or by forking it to speed up development.

Why do people keep insisting that subscription plans are subsidized? They are not. API prices are outrageous, designed to milk enterprise users. They probably have an 80-90% profit margin on subscription plans.

Inference is cheap; Dax once said on a podcast that Opencode has a close to 90% profit margin on the openweight models it provides inference for (at 10x cheaper pricing). Those models are close in size and spec to models from large labs.

Not really. MFA still protects against the much more common case where the password itself is compromised, either through a breach, reuse, phishing, interception, bad storage, etc. An MFA code is short-lived and can’t simply be reused later, unlike a password.

Keeping the password and MFA secret in the same password manager reduces separation (if someone fully compromises your vault, they will gain access to both factors). But that doesn’t make MFA pointless; it just means it doesn’t protect you against that particular failure mode.

And if someone has full access to your password manager, you already have a much bigger problem.

Awesome write up! This is now one of my favorite articles on HN.

They are not taking any losses; they just charge the regular price with a very tiny profit margin. US healthcare costs are heavily inflated. Insurance is a scam that lets the rich get richer. All insurance should be non-profit.

We all contribute to the pool at a time when we don't need it so we can use the money when we do, not to make CEOs or stockholders rich. That is how an insurance pool should work. Any excess money at the end of the year should be moved to the pool for the next year, not as a bonus on CEOs' 100m+ salary.

And it's not just on OpenRouter; Vercel offers the same:

https://vercel.com/changelog/gpt-5-6-sol-is-50-off-on-ai-gat...

It's discounted if used via OpenRouter, not the official API.

Nope. You are charged for the reasoning tokens as well. Actually, most of the expense is the reasoning part.

On an industrial scale, yes, of course.

Yup, I agree. I’m just clarifying what I meant by my original comment here: https://news.ycombinator.com/item?id=49206654

I agree. We all make these kinds of trade-offs sometimes. We trade some amount of money for peace of mind, more interesting work, better colleagues, a healthier environment, and so on.

But I never buy the "mission" story, or the "we’re a family" line. We’re not a family, and I’m not on a mission.

I’m fine with the word "vision." I’ll be part of your vision if I find it interesting or if I believe in it too. But I’m not on a mission to achieve it.

To me, "mission" implies sacrifice and renunciation - whether that means sacrificing your lifestyle, your time, or your family. I’m not doing that.

Anyone who thinks people work for anything other than money is delusional. Anyone who works for "the mission" is an idiot - no company or billionaire running a company cares about you. They will let you go in a heart beat.

Just proving my point. "Hey, look, AC current can kill you. We must stop this nonsense"

https://www.wired.com/2012/01/0104edison-uses-ac-to-kill-ele...

I used the softer word "foolish" - but in reality, they are idiots. Yes, they are, and they will look like idiots in the future when China wins the AI race. In the same way, nuclear opponents are idiots and have contributed to the decline in nuclear research in the West over the past 50 years (looking at you, Europe, especially Germany), while China keeps advancing its nuclear research (fission and fusion as well) and building nuclear plants to ensure its future energy independence and dominance.

Search engines are free.

The majority of people opposed AC current, cars, steam engines, rail tracks, airplanes, computers, the internet, and so on. There was always irrational fear present with new tech advances. People opposing data centers will not look any less foolish than those who opposed AC current.

Even if they can solve problems like this every day, you still have a very limited number of grad students who can solve them. With model capabilities like this, you can have the equivalent of millions of grad students who can solve problems like this.

Ah, ok, thanks for the explanation. I didn't want to create a duplicate; I checked to see if anyone posted it, but didn't see it at the time.

Why move to a thread that is clearly a copy of my post, created 2 hours after mine?

Exactly. Git is an amazing piece of software. If a team can’t use Git, which is still the simplest and most reliable way to track project history, I wouldn’t have much confidence in their ability to produce quality software.

The Flash model costs more than the Frontier models. Didn't see that coming.

I’m more interested in the topic myself. Do you have any recommended reads (besides the usual ones - tldp, kernel docs, linux foundation, etc)?

I’m curious how it’s legal to size a ship in international waters under any circumstances? We have a word for that - piracy.