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 had to add minimum paragraph and sentence counts to the prompts it uses to keep it from trying to write a few terse sentences each time.

I noticed this in Claude (from DDG). It invariably tries to shorten sentences or make them more compact and concise. Sometimes that can be good, but most times long-winded sentences are what are needed.

and there is no way I can find to make it stop popping up giant intrusive ads for their other services every single time I open it.

Just like Firefox. There is no easy way to turn off being harassed by an intrusive pop-up telling me about a new update.

There used to be. It was a simple checkbox. But somewhere along the way someone had to justify their existence and removed that simple checkbox in favor of harassment.

2. Every phone mounting solution is janky shit and looks like janky shit. If I was still 16 and it was a beater from Craigslist, sure - but neither are the case.

I use the Quadlock system in my car same as on my motorcycle....with magnetic charging.

Works great, secure quick mounting and detaching the phone when leaving either vehicle.

On the bike it's on the handlebars obviously, but for there car I got a windshield mount that holds it and it works great!!!!

Me? I like my big iPhone, but I only replace them ever 6 years or so....so, with the cost spread out over that long, it isn't really $$.

My 2017 car doesn't have CarPlay...but no big deal, I just use Bluetooth to stream Spotify to the cars OEM head unit....plays great.

If I need to use maps, My phone is right there in easy vision and can still be playing tunes while giving me directions.

...the newer ones that also tend to have better functioning voice driven commands.

I do NOT want to talk to my fucking car.

Hell I can't stand phone assistance that tries to make me talk to the computer.

The only time I want to be talking to my car is when I'm cursing it fucking up, or the damned traffic situation at the time if it isn't moving at speed....

Given some of the places I drive, a pure EV is not an option for me ...

Amish communities?

Southeast in the US.....

Hell last time I used the app to look for public charging stations in New Orleans.....there were precious few.

Most of the ones on the app were private ones....

I"d really not want to be stuck with an EV just before or after a Hurricane when power can be out for a month or more easily.....

You can drive gas in....but it's not so easy to drive batteries in to re-fuel your car, even if you had a widespread public EV charging infrastructure down here.

I am thankful I have a car that is till majority populated with physical switches, buttons and dials I can operate 98% from muscle memory without having to take my eyes off the road....of if I do it is usually a second or less.

I prefer 2 seater sports cars, that's all I've had my life to date.....I like analog gauges, manual transmission and all.

Frankly if they could come out with a simple car like in my youth....a '78 280Z type car...simple transmissions, I was more than able to deal with crank windows....that thing ran forever, simple and fun to drive while I owned it.

No worrying about telemetry, no computer needs....I mean I'll be happy with it to have Bluetooth so I can stream Spotify to it, but other than that, just keep analog controls, and as little computer as possible.

Cars like that, can last multiple decades with a little care.....and I pretty much only want my distractions to driving to be when the stereo starts playing "Highway to Hell" or the like....and I start head banging while driving down the highway...a little faster.

Would like to add one tidbit because I've had this same conversation at work. The professor stated they used AI to help edit only after hundred of hours of writing. In other words, they did the heavy lifting and only used AI at the end to help refine what they had already done.

Is this any different than a human editor? If I write a book, do what I can to edit it down, then go to an editor who makes suggests, removes or reorders parts, and so on, is that any different than having AI do the same thing? In both cases you wrote the words. In both cases you went somewhere else to help edit.

If people claim that using AI to edit a work is "cheating" because it's no longer your work, then the same should apply to a human editor.

No, but our aerospace industry learned a lot of useful things about how to operate in space. We can't be risk free if we want to make progress. In order to learn you have to fail sometimes; If we, the taxpayers, insisted on a refund that would likely put this company out of business and a lot of the learning would be lost.

So..what exactly are we learning for $30M?

How to fuck up a launch/deployment of a low orbing space craft?

Haven't we been in space long enough to learn those lessons?

Yes I'm a bit sarcastic....but seriously, what learning are we getting from our $30M.

If this company sucks....should they NOT go out of business and open bids up to more qualified companies for next shots?

We seem to be in a parallel to Rural Electrification now.

If someone lives five miles from the nearest connection, currently on 100A AC and POTS, is the limited budget better spent on running five miles of fiber or dropping off a satellite terminal?

They'll be going from either 3Mbps DSL to 300Mbps satellite or 800Mbps fiber.

If we had unlimited money I'd say fiber almost everywhere. But that isn't realistic.

The flip side is some of these rulings which ought to be applied to edge cases may well be applied to areas where FTTH is entirely reasonable and achievable.

Government almost always sucks at applying broad rules without subtlety at both the median and the far sigmas.

I live in an area where welfare broadband is offered for about $15/mo for about 60Mbps. Frankly I could live with that almost all the time. I do download some massive datasets every few months but my smoothed average is probably 2% of my max.

But those welfare programs exist because society has decided that some Internet is required for participation so there is some responsibility to provide.

The bigger risk is probably dual-use dangers. Ukraine has been using Starlink for terrorist attacks and while Musk has said he cut them off the Russian FM has warned that Russia will take out the Starlink constellation if it continues because apparently it is.

So we don't run fiber and then Starlink goes away, and then what?

There are also ice storms, hurricanes, tornados, wildfires, etc. that can take out landines for months so having a few satellite terminals in an area would be more antifragile.

FCC is seemingly treating this as simple when it's complex.

I recently downgraded from Cable 800MB to starlink. Simply because for some reason at 2pm EVERY SINGLE DAY my ping times with charter would go from around 30/ms to 100-1000/ms and it lasted for a few hours. I work from home and that was unusable. The only other "broadband" available was starlink. So now I'm much slower in the 100-200MB range, but overall we haven't noticed much difference.

So I'd say, yeah, you don't have to go all the way up to gigabit to be considered broadband.

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)