At the moment, I do find that the LLMs bring me more of what I want to see than the basic search engine used to a few years back… 15-20 years ago the basic search engine seemed to be much better then slowly deteriorated from there.

I have confidence, they’ll sell this version to the highest bidder faster than they did the last one, then it will be just as bad as listening to radio advertisements as a way to find a good local restaurant to eat at…

Agreed, and for the moment the “AI experience” is better than what Google Search was a couple of years ago… but I have faith, they’ll sell this one out too, after all: users are the product.

I mean, seriously, what advances of the past 10 years do desktop users need? What can’t they wait 2-5 years for it to percolate down into the XLTS kernel? Mostly security, which should be mostly invisible to the users.

There are varying opinions about the quality of the reports: https://www.theregister.com/software/2026/03/26/linux-kernel-czar-says-ai-bug-reports-arent-slop-anymore/5226256

I have recurring thoughts that the kernel needs to undergo a clear fork. One branch continues on as it is today. A new branch agressively restricts scope, drops support for sub 0.1% market (in use, not last quarter’s sales) share hardware - and software. Focuses intensively on making that core functionality as reliable and secure as possible. New features? No thanks, plenty of existing features already.

a real solution.

How about fixing the bugs so they’re not there to report? That’s the real “unpatched hole in the floor.”

Another “fight fire with fire” approach is to let agents do the screening for “duplicate report” and also pre-verify / test reports for reproducibility.

The problem with replacing e-mail is that e-mail works well enough.

Other hypothetically “superior” replacements come and go (Google Wave, Yammer, Jive) - some are sticking around in limited scopes (Slack / Teams) - but none have displaced e-mail completely. You always have to ask: “Are you on Teams/Facebook/X?” some people are, some people aren’t. Just about everyone at least has e-mail access, and uses it to some degree or another - if nothing else to verify identity for accounts on other services.

Cooperative agentic synthesis?

they all assume most people won’t bother

There’s a sort of built-in compensation for “response rates” or “perception rates” - in my industry we trend customer complaints and act according to the data we receive, but we also know that for every complaint we receive there are typically 30 similar events that go unreported. We also know that certain “responders” are outliers and will report every single instance they experience (and sometimes embellish and create additional instances for dramatic effect) but these are exceedingly rare and usually “adjusted” to normal responder levels once identified.

Now, when people create AI agents to file the complaints for them… that’s a new level of response rates. 25 years ago I came close to doing this for airport flyover noise complaints - our local (international) airport had an obscure portal for local residents to complain when they were bothered by jet flyovers - and our neighborhood would get dozens of events per month where the noise was so loud you couldn’t hear the other side of a phone call INSIDE your house with the windows shut. Thousands of homes were impacted by this, often 4 or 5 times in a row within an hour or two. But, the complaint channel was so obscure and the reporting process inconvenient enough that very few complaints were recorded, and they loved to point out that 40% of their complaints came from a single resident. Smart phones weren’t a widespread thing yet, if they were I would have “made an app for that” where anytime you were “impacted” by a jet flyover all you would have to do is pull out your phone and tap the app to file a report. (I considered developing it for Palm Pilot, but I doubt even 10 residents would have carried Palm Pilots for the purpose of filing reports…) If we got a couple hundred residents across the neighborhood reporting even 10% of the troublesome flyovers, we might have changed the conversation - as it was the airport used the lack of complaints to justify no change in flight patterns.

LLM powered mailing list filters.

Deep Seek and other locally hosted options should be up to this task…

The reports don’t have to be irrelevant slop to be overwhelming.

https://www.theregister.com/software/2026/03/26/linux-kernel-czar-says-ai-bug-reports-arent-slop-anymore/5226256

TL DR: they gave the agents a minimal initial prompt and zero additional feedback while they ran. Humorous / weird behavior ensued.

If you snatched a college student off a crosswalk, gave them the same prompt and stuck them in a booth with control of the station and zero feedback, even if they were willing and eager to take on the assignment I suspect similar psychotic behaviors would emerge.

Those who don’t learn from history, probably were homeschooled?

Edit to observe: when I visited Paris, in 1989, I was struck by the celebration of “baguettes! we have fresh baguettes!!!” apparently the presence of bread in the stores was (still?) and unpredictable / unreliable cycle for them. I also wandered into a crowded sandwich shop only to be told “no pain, NO PAIN!” meaning: we’re out of bread and therefore will not be serving you.

highly complex systems in situations of huge responsibility.

What’s complex about “there’s a salt mine under this lake you’re drilling in?” Or “you’re putting a gas tank in the most common impact crumple zone on the vehicle?” or “We’ve seen this problem before, many times, but we’re just going to continue to let it happen again and again?”

I wouldn’t say “if humans weren’t capable of learning we wouldn’t be here in the first place” - I would say “by random evolution of circumstances, humans are where we are today - some capable of learning, some apparently not.”

the (human) hive mind learning works worse than in hiving insects.

That’s an entirely opinion driven statement. What is better, or worse? From whose perspective? Do you know what you don’t know? If you think you do, you’re wrong.

From my perspective, people are a squishy mess. AI/LLM are also somewhat of a squishy mess, but I find them to be a lot more consistent and predictable in their behavior than randomly selected people. And, as far as the hiring process to find “the right” people for a particular job, that’s a long complicated unpredictable usually costly and error prone process, even before you get to the point that the people you have engaged for a certain task might start to learn and improve in their role. I can hire AI agents for the equivalent of pennies per hour of equivalent human output, and while they have their issues, I can get as many of them as I want with that same predictable behavior / capability for just a few dollars more. They haven’t started suing for slip and fall (yet), their performance doesn’t degrade based on time of day, day of the week, phase of the moon. They don’t call out sick, or pregnant. They don’t want healthcare insurance… whatever they can do, they would seem to be the ideal employees to do it.

I wish they would spin CoPilot into its own APPLICATION. Same for Edge. Don’t deeply integrate them to the point that the whole system falls apart when they are missing. They should be optional. Allow them to inter-operate with other apps, sure, but don’t make everything depend on everything else.

The problem is: momentum. Starting over is relatively easy, it also ignores the “value” in all of those legacy systems which you can continue to milk for income if you stay at least a little compatible with them.

Do bees learn? Like how to deal with mites? Or do they just die off every 45 days and only get replaced by bees who accidentally happen to be a little better at dealing with mites?

People with no experience are going to fuck shit up completely.

As they always have.

which is not the case for people because humans can spot the “obviously stupid” or “obviously dangerous”

No AI was used in the creation of these clusterfucks:

1. The Lake Peigneur Maelstrom - In November 1980, Texaco was conducting exploratory oil drilling directly on top of a shallow, 10-foot-deep freshwater lake. Operating directly underneath that same lake was a massive, active multi-level salt mine

2. The Banqiao “Iron Dam” Collapse (China, 1975) Built in the early 1950s for flood control, the Banqiao earthen dam was heavily reinforced with Soviet engineering assistance and proudly nicknamed the “Iron Dam” by the government, which declared it completely unbreakable

3. The Capsizing of the Vasa Warship (Sweden, 1628) In 1628, King Gustavus Adolphus built the Vasa, an opulent warship meant to serve as the crown jewel of the Swedish Navy. It was designed to intimidate enemies with unprecedented firepower.

4. The gas tank in the back of the Ford Pinto.

5. The Tesla Cybertruck (well, maybe some AI got in there, but the core bad ideas were well established before ChatGPT was “a thing”.)

6. Lead in gasoline

7. The Triangle Shirtwaist Factory Fire (New York, 1911) The Triangle Shirtwaist Factory occupied the top floors of a Manhattan building, employing hundreds of young immigrant women. Management routinely ignored basic industrial safety measures to maximize profits and prevent employee theft. The “Obviously Dangerous” Reality: Locking workers inside a high-rise room filled with flammable textiles and scraps creates a lethal death trap in an emergency.

8. Bhopal 1984

9. Chernobyl 1986

10. The Hillsborough Stadium Disaster (Sheffield, UK, 1989) During an FA Cup semifinal match between Liverpool and Nottingham Forest, thousands of fans arrived outside the Leppings Lane end of the stadium just before kickoff, creating a massive, chaotic bottleneck at the turnstiles. The “Obviously Dangerous” Reality: Opening a massive exit gate to let thousands of frantic people rush blindly down a narrow tunnel into an already overcrowded, fenced-in terrace creates a lethal human crush.

11. The Who - December 3, 1979 at the Riverfront Coliseum in Cincinnati, Ohio.

12. School shootings…