Certainly! Your point raises a crucial, multifaceted issue that merits deeper examination. It's not just about detection accuracy, it's about the evolving interplay between human intention and algorithmic assistance — a dynamic that resists simple binary classification.

When we delve into the mechanics of modern AI tools, it becomes evident that the boundary between "correction" and "generation" is increasingly porous. Consider the following nuances:

Gradual authorship spectra: Text exists on a continuum — from fully human-crafted prose, to human writing with AI-suggested punctuation, to passages where the AI has restructured entire paragraphs. Detection tools, however, often operate on categorical assumptions that fail to capture this complexity.

The feedback loop phenomenon: Furthermore, as writers internalize AI-suggested patterns — Oxford commas, em dashes, and certain syntactic rhythms — their "human" baseline shifts. This creates a fascinating, almost paradoxical scenario: the more one learns from AI feedback, the more one's unassisted writing may trigger detection algorithms.

Contextual markers vs. stylistic choices: Notably, certain signals — like the increased use of "delve," "testament," or negative parallelism — may reflect genuine stylistic evolution rather than automated generation. It's a testament to how AI influences language culture broadly, not just individual documents.

In conclusion, the challenge you identify is not merely technical but philosophical — forcing us to reconsider what we mean by "authentic" voice in an era of human-AI collaboration. The question, ultimately, is not whether a tool assisted in the writing, but whether the ideas, arguments, and intellectual labor remain genuinely human.

I hope proves helpful! Please let me know if you would like to delve deeper into any specific aspect of these findings, or if you would like me to explore the broader implications for content strategy.

"do a little research" - you do realize that your post would have had just the same educational value but not made you come across as a jerk if you had just left off that first sentence?

https://www.cs.wisc.edu/2025/0...

And DARPA has a whole program for it:

https://www.darpa.mil/research...

I think the obvious counter to an influx in bug reports / pull requests is to have the first layer of review of bug reports / pull requests to be AI as well.

Final remains humans.

I dunno, if that's how people vote, that's how it should be.

Though there is a big question of how they plan to run a vote with multiple options. I hope it's a Condorcet system to avoid the risk of watering down one side with similar competing options...

Oh, but surely they'll win, they indicted [strike]a ham sandwich[/strike] this man! And the Trump administration has such a spectacular record with high-profile prosecutions thusfar...

That is not what they found. What they found is that if you spend 60 minutes doing the homework with AI, you learn approximately as much as someone who spends 60 minutes doing it without AI. If you spend 20 minutes doing it with AI, you learn as much as someone who spends 20 minutes doing it without AI. Etc. The only real difference is that the AI people get way better homework scores, and as a result, tend to stop spending as much time, and thus tend to learn less.

That's what this preprint study found.

100% this. Even a short 5-minute quiz with just a couple questions at the start of each day will do the trick. But you simply can't rely on homework scores to be a proxy of learning or time spent studying anymore. Brief daily quizzes on the previous day's material will force them to.

Also, It's not like students have the excuse of, "I couldn't study, I didn't have anyone to explain X to me" anymore. The very AI that they might otherwise use to cheat on their homework can also explain to them the things they missed in last year's algebra class that are blocking them from learning this year's calculus, or whatnot. Congrats - you now have an infinitely patient personal tutor who will spend hours dumbing things down for you until you get it. Now you have no excuse - learn the material. It becomes all about actually putting the time in.

Yeah. There's been widespread misrepresentation of this (preprint!) in the media. What they actually found: test scores correspond to how much time you spend doing homework, regardless of whether you use AI or not. Which should not be a surprising result! Of course, the AI people, esp. those who had been using AI for more than a few months, stopped spending much time on their homework, and then their test scores fell to match the non-AI users who didn't spend much time in their homework either.

The TL/DR is, in today's age, you cannot use test scores as a proxy of how much time a user is studying or knows the material.

There are of course some caveats, beyond the fact that it's just a preprint - notably that there's nothing to separate why the user started using AI to do their homework, so there's potential confounders in there. For example, if a student is in a poor family, maybe they had to get a job to help pay the bills. But working a job itself can lower test scores - for example, having to work late the night before an exam. So there are some confounder problems with the study. That said, I'm sure the general conclusion is correct: time spent doing the thing = amount of learning the thing (relatively independent of what tools you're using in the process).

I have to disagree, for several reasons. First off, file sizes just keep increasing. Instead of people thinking their televisions at home look great in "full 1080 HD resolution"? Now, everyone's moving to 4K TV, with 8K being the high-end offering. That means streaming television supporting those higher resolutions is going to use 4x to 8x the bandwidth it used to.

Second, any avid gamer knows how big the downloads have gotten for modern games. The regular updates they push out, alone, are typically 35GB to 350GB+ in size. Now, imagine you've got a family at home with at least 2 or 3 people playing online games on their machines. They don't want to have to wait 20-25 minutes to start playing every time it says an update is required first.

Beyond all of that, though? I've found fiber connections used in an office environment are handled a little differently than what people have at home. EG. I can start uploading 4 or 5 large archives to, say, Microsoft Azure cloud storage, right in the middle of the workday -- and nobody else using the office network even notices. That's largely because we have multi-thousand dollar routers in the office and they're properly configured to prioritize certain types of traffic like VoIP phone calls, and to throttle connections intelligently so everyone gets a fairly equal piece of the total connection "pie" as they work. It's not that reasonable to expect residential users to have similar setups, or to even know how to configure them.

My cable provider now offers 2GB download speed service, but it's problematic in the sense most people still have devices with wired Ethernet ports that only handle up to 1 gig. speeds. I have a pretty expensive Linksys Atlas mesh router at home (cost $850 or so for the 3 piece set, a few years back), and it doesn't have a port on it faster than 1 gig. for the cable modem to attach to.

Obviously, the provider is happy to supply their own combo modem/router to use but with a monthly cost attached.

But there's just not a good reason in 2026 to not make a least gigabit connections a standard offering. Nobody's "unfairly discriminating" against services like satellite with any of this. Those technologies have always been considered wireless alternatives to use when you're in too rural of an area to expect anything else. Invest in building out fiber networks and those techs will just become what you use when you're mobile.

I feel like I'm in a constant race to flee enshittification : Used to use Cursor (it's still my only option at work), but then Musk bought it, and they instantly went in on hard-sell tactics. Switched to vscode + copilot + OpenRouter, but was getting concerned about API costs. Just yesterday I set up a Tencent TokenHub account with their Max package, which required that I switch off copilot if I wanted to be able to use some models like Kimi K3 (had a temperature=1.0 requirement, copilot doesn't let you set temperatures and hardcodes them to very low), so now it's vscode + Pi + TokenHub.

Sigh. Please don't make me move again...

(Not that Tencent is in any way "the good guys"... but they're also providing a very generic product)

Honestly? I think your last statement there is a key point that should receive a little more attention. It seems possible --- heck, even prudent -- for a person to question possible negative outcomes from mRNA vaccines, while still fully recognizing that it makes good sense to use one on an existing disease. Even in a nearly worst-case scenario? You're talking about cancer, which surely kills a person if not treated, vs a theoretical risk of some long-term problems from an mRNA vaccine (some of which even anti-vax types admit aren't showing up except in a certain percentage of vaccinated people).

Ah, the good 'ol "No True Communism" argument :)

That has much more to do with people being creatures of habit than being stupid.

I'm old enough to have spent a long time driving cars before voice recognition was a thing. Probably dating myself here, but I remember how excited my friend was when he got a new Nissan Z car and it actually spoke to him, saying things like "Your fuel is low!" and "The door is ajar!"

I spent years trying to utilize voice recognition systems when they came out, and was consistently frustrated. Some required I memorize their specific set of words they understood (cough, Chrysler ...) and didn't do anything else. Others just couldn't understand what I said half the time. By the time any of these systems in cars worked well, I'd given up on them as more hassle or gimmick than they were worth.

At least w/Infotainment touchscreens, you get visual confirmation of what you picked/selected. Voice commands get super annoying as soon as it misunderstands you once and starts the wrong process.