Notices by mhoye (mhoye@cosocial.ca)

"There are colors that I want to show you, but I can’t. They exist in the real world. You probably saw some of them today, but I can’t show them to you on a screen. A digital photograph can’t capture them, and your screen can’t display them. No game you’ve ever played has contained them. Unless you have specialized equipment, they are entirely absent from the digital world.

Most of them are cyans."

https://moultano.wordpress.com/2026/06/19/where-to-find-the-colors-your-screen-cant-show-you/

ouch

Because the only thing I bring into this world that I don't care about looking at or thinking about or doing better is shit. My own literal, physical feces. I've even had to care about other people's shit, I've changed diapers, picked up after dogs and shovelled fertilizer.

You're don't even look at your code now? You're proud of it, because productivity, velocity?

When was the last time you talked about your happiness or self-worth without using the word "productive"?

Answer the question.

The one bright spot of LLMs is that we've learned exactly how morally and creatively bankrupt this industry is, and how fast people are willing to throw away everything they said they cared about for decades - craft, understanding, efficiency, detail, all of it - just to cosplay competence and bandwagon their way to groupthink targets they don't even realize were made up to manipulate them.

Christ it's embarassing.

The fact that we're willing to rely on LLMs to generate code knowing that it only mostly works because that code has been generated before thousands of times is not an indication that stochastic models are good, it's a massive, punishing indictment of computing as a field.

Using an insanely huge expensive model to quasi-reproduce work that's been created thousands of times already isn't productive or efficient. It's a symptom of profound failures of language, practice, process and imagination.

One of the most important things about solar energy is that you don't need to buy it again next year.

https://mastodon.social/@anirvan/116772353347087105

@dalias Respectfully, the possibility of incomplete, adversarial or malicious telemetry is one that has occurred to other people, some of whom work on these problems full time.

@dalias @Osmose

You can call them obedient fools if you like, but tarring the massively overwhelming part of the population - and I mean, massively overwhelming, well north of 99% - of the software population - with that brush is you talking about you, not about anything else.

I doubt we're going to find any common ground about this, but I don't believe it's possible to responsibly build load-bearing technology without understanding how it fails in real world use. I think that's negligent.

@dalias @Osmose

A significant improvement over the old way, that amounted to product leadership by whoever got yelled at the loudest and the mostest on the orange website and never knowing how or why the product failed in real-world use.

An exceedingly uncharitable belief that I've long held is that the majority of the anti-telemetry hackernewser crowd aren't opposed to telemetry because of privacy, but because telemetry generates evidence of how little their opinions actually reflect reality. Further and even more uncharitably I believe that most of the world's AI enthusiasm supports this belief: the hackernewser commentariat doesn't care about reality or craft nearly as much as being seen as holding a smart person's opinions.

Nobody wants to believe this is true, but the number of people who ever change any of the default settings is a rounding error in the overall user population of any piece software.

https://gts.nyquil.org/@nyquildotorg/statuses/01KTZ64GBJERY2YXR730WZ0GF3

@dalias Those tools exist, and they get used, and there are even circumstances where statistically significant anomalies emerge spontaneously and unintentionally from marginal outlier circumstances.

Bruce Banner canonically has 7 PhDs and honestly the fact that he periodically turns into a giant green ragemonster after spending that much time in academia without tenure is the most believable thing in the entire Marvel canon.

RE: https://mastodon.social/@nixCraft/116752805722997877

In its infancy, the idea of "DevOps" was that if you wrote the software, you carried the pager for that software. It wasn't "Operations was handled in a developer-style environment", it was "the developers share the operations burden" and aligned incentives are magic.

Today DevOps just means "managing service contracts in a terminal window", and it is not magic.

@timbray @jwz None of these guys have ever had so many people cheering them on, let them have this.

I know things are difficult and complicated and muddy sometimes, but I still think that if your argument amounts to "I know about all the externalities and the consequences but those happen to other people and it works great for me so I think it's great and I'm going to keep doing it", you've told me as much as I need to know about you.

Whatever they say about "loving the Web", the Chrome team will choose Google's business model over everything else, web standards, user agency, security, privacy, interoperability, compatibility, literally anything and everything else every single time.

What anyone on that team says they love or care about or want does not change anything they're going to do.

Yes, they're thrashing. Yes, the AI stuff is bullshit nobody wants.

Still: get Firefox, add uBlock Origin.

https://www.pcworld.com/article/2595287/ublock-origin-is-officially-dead-for-chrome-but-ad-blockers-live-on.html

@mos_8502 Oh, yes of course. It cuts off both kids from getting parents help and parents from getting help from other parents, all for the low low price of total, pervasive surveillance of everyone's online activities.

What's not to love, really.

Here's a true fact about age lockouts and social media bans: security is in direct tension with usability, and 100% of tech companies will choose malicious compliance to the letter of the law over losing a single sale to the terrible ergonomics of effective security.

e.g.: Agelocks on TVs all show the code on screen in text large enough that poorly-sighted seniors can use it. It doesn't matter that a kid can see the code. What matters is the person who might return the TV isn't inconvenienced.

Age restriction tech won't work because one thing market realities and industrious kids both want like crazy is to make sure they can never work.

But it will mean that internet bullying gets immensely worse, because kids can't admit they need help if that means admitting they Crossed A Line, and that means they'll need to go somewhere else, to somebody else.

If the goal here is harm reduction it will be a disaster. But hey, it will definitely let shitty parents pretend nothing is their fault.