Better Technology, Worse Motivation: GenAI’s Mediocrity Trap

While generative AI (GenAI) promises productive efficiency, it can paradoxically lead to lower-quality work. We conducted an experiment with professional illustrators and found that AI assistance flattens the quality curve—it accelerates initial gains but sharply diminishes the returns on sustained effort. Faced with this, a significant number of professionals made a strategic choice: they sacrificed the final quality to save time.

From http://www.jin-li.org/uploads/1/1/4/5/114595093/ai_and_motivation.pdf

I haven’t read this and can’t vouch for it; seems vaguely AI-boostery. Still, the conclusions are interesting. This seems to be the picture that is emerging about generative AI generally: most people don’t like it and find that degrades the quality of work. Coders seem to like it and think that it helps them, but in fact it makes the slower, less productive, and more bug prone.

By all measures it’s a bad technology. We should just be honest about it. There is no need to make excuses for multi-trillion-dollar corporations.

yakumo.dev is finally done for, November 26, 2025 in UTC time.

So long… I won’t be missing it though

FTR, I see one (two) issues with PyQt6, sadly:

1. The PyQt6 docs appear to be mostly auto-generated from the C++ docs. And they contain many errors or broken examples (due to the auto-conversion). I found this relatively unpleasent to work with.
   
2. (Until Python finally gets rid of the Global Interpreter Lock properly, it’s not really suited for GUI programs anyway – in my opinion. You can’t offload anything to a second thread, because the whole program is still single-threaded. This would have made my fractal rendering program impossible, for example.)

I should work on my client again and add some new features. Like adding a new feed directly in the client and not having to go to the config first. And showing a preview of a feed before actually adding it. Also, a search would be something to add. And finally combining my User-Agent analyzer with my subscription list to spot new feeds automatically.

@eldersnake@we.loveprivacy.club What version are you running btw? It’s probably time you upgraded and time I released a new version finally 😂 If you’re running a version that’s pre-SQLite-cache, then yeah I’m not surprised. The SQLite cache version is honestly much better 🤣

@prologic@twtxt.net Let’s go through it one by one. Here’s a wall of text that took me over 1.5 hours to write.

This section says AI should not be treated as an authority. This is actually just what I said, except the AI phrased/framed it like it was a counter-argument.

The AI also said that users must develop “AI literacy”, again phrasing/framing it like a counter-argument. Well, that is also just what I said. I said you should treat AI output like a random blog and you should verify the sources, yadda yadda. That is “AI literacy”, isn’t it?

My text went one step further, though: I said that when you take this requirement of “AI literacy” into account, you basically end up with a fancy search engine, with extra overhead that costs time. The AI missed/ignored this in its reply.

Okay, so, the AI also said that you should use AI tools just for drafting and brainstorming. Granted, a very rough draft of something will probably be doable. But then you have to diligently verify every little detail of this draft – okay, fine, a draft is a draft, it’s fine if it contains errors. The thing is, though, that you really must do this verification. And I claim that many people will not do it, because AI outputs look sooooo convincing, they don’t feel like a draft that needs editing.

Can you, as an expert, still use an AI draft as a basis/foundation? Yeah, probably. But here’s the kicker: You did not create that draft. You were not involved in the “thought process” behind it. When you, a human being, make a draft, you often think something like: “Okay, I want to draw a picture of a landscape and there’s going to be a little house, but for now, I’ll just put in a rough sketch of the house and add the details later.” You are aware of what you left out. When the AI did the draft, you are not aware of what’s missing – even more so when every AI output already looks like a final product. For me, personally, this makes it much harder and slower to verify such a draft, and I mentioned this in my text.

You, @prologic@twtxt.net, also mentioned this in your car tyre example.

In my text, I gave two analogies: The gym analogy and the Google Translate analogy. Your car tyre example falls in the same category, but Gemini’s calculator example is different (and, again, gaslight-y, see below).

What I meant in my text: A person wants to be a programmer. To me, a programmer is a person who writes code, understands code, maintains code, writes documentation, and so on. In your example, a person who changes a car tyre would be a mechanic. Now, if you use AI to write the code and documentation for you, are you still a programmer? If you have no understanding of said code, are you a programmer? A person who does not know how to change a car tyre, is that still a mechanic?

No, you’re something else. You should not be hired as a programmer or a mechanic.

Yes, that is “skill evolution” – which is pretty much my point! But the AI framed it like a counter-argument. It didn’t understand my text.

(But what if that’s our future? What if all programming will look like that in some years? I claim: It’s not possible. If you don’t know how to program, then you don’t know how to read/understand code written by an AI. You are something else, but you’re not a programmer. It might be valid to be something else – but that wasn’t my point, my point was that you’re not a bloody programmer.)

Gemini’s calculator example is garbage, I think. Crunching numbers and doing mathematics (i.e., “complex problem-solving”) are two different things. Just because you now have a calculator, doesn’t mean it’ll free you up to do mathematical proofs or whatever.

What would have worked is this: Let’s say you’re an accountant and you sum up spendings. Without a calculator, this takes a lot of time and is error prone. But when you have one, you can work faster. But once again, there’s a little gaslight-y detail: A calculator is correct. Yes, it could have “bugs” (hello Intel FDIV), but its design actually properly calculates numbers. AI, on the other hand, does not understand a thing (our current AI, that is), it’s just a statistical model. So, this modified example (“accountant with a calculator”) would actually have to be phrased like this: Suppose there’s an accountant and you give her a magic box that spits out the correct result in, what, I don’t know, 70-90% of the time. The accountant couldn’t rely on this box now, could she? She’d either have to double-check everything or accept possibly wrong results. And that is how I feel like when I work with AI tools.

Gemini has no idea that its calculator example doesn’t make sense. It just spits out some generic “argument” that it picked up on some website.

The AI makes two points here. The first one, I might actually agree with (“bad bot behavior is not the fault of AI itself”).

The second point is, once again, gaslighting, because it is phrased/framed like a counter-argument. It implies that I said something which I didn’t. Like the AI, I said that you would have to adjust the copyright law! At the same time, the AI answer didn’t even question whether it’s okay to break the current law or not. It just said “lol yeah, change the laws”. (I wonder in what way the laws would have to be changed in the AI’s “opinion”, because some of these changes could kill some business opportunities – or the laws would have to have special AI clauses that only benefit the AI techbros. But I digress, that wasn’t part of Gemini’s answer.)

Except for one point, I don’t accept any of Gemini’s “criticism”. It didn’t pick up on lots of details, ignored arguments, and I can just instinctively tell that this thing does not understand anything it wrote (which is correct, it’s just a statistical model).

And it framed everything like a counter-argument, while actually repeating what I said. That’s gaslighting: When Alice says “the sky is blue” and Bob replies with “why do you say the sky is purple?!”

But it sure looks convincing, doesn’t it?

This took so much of my time. I won’t do this again. 😂

@bender@twtxt.net Not sure, if we actually have a law like that. But I wish it was the case. The clamp doesn’t say anything like that, just that it is now cactus.

The glue takes three days to reach its final strength. Let’s see. I’m sceptical.

@lyse@lyse.isobeef.org now that I could finally read you RSVP message (or, should I say, essay, LOL), I had to see the video, and well, https://netbros.com/1761846417/. Hahahahaha!

@bender@twtxt.net @prologic@twtxt.net Let’s see on which day we’ll finally settle.

I reckon the white-space: nowrap is a bit evil on the gatherly notes, though.

@prologic@twtxt.net That sounds horrible. 😅 I wouldn’t want to own such a car. (My plan is not to buy a new car after my current one finally broke down entirely.)

@lyse@lyse.isobeef.org First time I heard about eCall. I don’t think I like this. 🫤 Feels like another attempt at going for complete surveillance. Yes, yes, it’s about “security”/“safety” … it always is.

Finally, new books arrived. Let’s see if Dead Silence is as good as it sounds. 😃

Looks like this yarn is finally back again after like a week

Hello again everyone! A little update on my twtxt client.

I think it’s finally shaping a bit better now, but… ☝️

As I’m trying to put all the parts together, I decided to build multiple parallel UIs, to ensure I don’t accidentally create a structure that is more rigid than planned.

I already decided on a UI that I would want to use for myself, it would be inspired by moshidon, misskey and some other “social feeds” mock-ups I found on dribbble.

I also plan on building a raw HTML version (for anyone wanting to do a full DIY client).

I would love to get any suggestions of what you would like to see (and possibly use) as a client, by sharing a link, app/website name or even a sketch made by you on paper.

I think I’ll pick a third and maybe a fourth design to build together with the two already mentioned.

For reference, the screens I think of providing are (some might be optional or conditionally/manually hidable):

• Global / personal timeline screen
  
• Profile screen (with timeline)
  
• Thread screen
  
• Notifications screen or popup (both valid)
  
• DM list & chat screens (still planning, might come later)
  
• Settings screen (it’ll probably be a hard coded form, but better mention it)
  
• Publish / edit post screen or popup (still analysing some use cases, as some “engines” might not have direct publishing support)
  

I also plan on adding two optional metadata fields:

• display_name: To show a human readable alternative for a nick, it fallback to nick if not defined
  
• banner: Using the same format as avatar but the image expected is wider, inspired by other socials around
  

I also plan on supporting any metadata provided, including a dynamically parsable regex rule format for those extra fields, this should allow anyone to build new clients that don’t limit themselves to just the social aspect of twtxt, hoping to see unique ways of using twtxt! 🤞

@lyse@lyse.isobeef.org Finally! The end is near! Rejoice! \o/

@alexonit@twtxt.alessandrocutolo.it prologic has me sold on the idea of hashv2 being served alongside a text fragment, eg. (#abcdefghijkl https://example.com/tw.txt#:~:text=2025-10-01T10:28:00Z), because it can be simply hacked in to clients currently on hashv1 and provides an off-ramp to location-based addressing (though i still think the format should be changed to smth like #<abc... http://example.com/...> so it’s cleaner once we finally drop hashes)

I finally solved the loading issue in my WIP reader, TwtStrm (and apologies again to anyone that got spammed while I was diagnosing the issue).

After another round of coding this weekend, I’m happy to report that it now renders all the twts (with markdown parsing), complete with localstorage and server-based file caching.

I finally resolved my issues with hashing twts… with REGEX!

Dates in JavaScript are truly strange creatures.

holy shit the future is here gopher is finally for the modern mar

The big QR code canine, has been one of my favourites - because even after a few months, I still find the pose really cute. Always thought a chibi version is a necessary addition and now I finally drew it.

This is the best hack I’ve seen in a while: https://github.com/mmulet/term.everything Finally a wayland compositor for the rest of us ;-)

And I am finally able to post from CLI again!

finally ordered a motherboard to replace the one i fried in my brand new server T__T it hurt but i had to do it, i put it off for literal months

Hmm, gnu.org is slow as heck. Shorter HTML pages load in about ten seconds. This complete AWK manual all in one large HTML page took a full minute: https://www.gnu.org/software/gawk/manual/gawk.html Is there maybe some anti AI shenanigans going on?

In any case, I find the user guide super interesting. My AWK skills are basically non-existent, so I finally decided to change that. This document is incredibly well written and makes it really fun to keep reading and learning. I’m very impressed. So far, I made it to section 1.6, happy to continue.

replies and following implemented! next step is further parsing of post contents, rendering threads, and then maybe i can finally start adding remote feeds…! though i kinda wanna redo the whole ui ^^’

I finally have my new (top-secret) twtxt client in a working state. Next comes the deployment, which I hope to finish tonight. Release date: TBD. Stay tuned!

I finally have my new (top-secret) twtxt client in a working state. Next comes the deployment, which I hope to finish tonight. Release date: TBD. Stay tuned!

website is finally (almost) mobile friendly!

Joining the Clippy profile picture club, now that I finally finished my custom one.

We went out an hour earlier to escape the rain. But we got drizzled on, nevertheless. As expected, the rain jacket was finally wet on both the out- and inside. Sweating under that plasic layer is unavoidable with these temperatures. It was still a very nice walk and great vibe with the gray soup. Just right. :-)

Didn’t bring my camera, I guarantee that it would have gotten soaking wet.

twtxting from Linux Mint uiiiiiiii! finally ~completely kiked windows #linux

Today I finally got rid of my /29 IPv4 subnet with my ISP used to power my ingress. No longer.

We finally got a caliper donated for this year’s scout flea market. We didn’t sell it, but kept it ourselves. It will come in very handy every now and then in our material store. For example, I missed having a caliper in the past when sorting our random assortment of screws or measuring the depth of a hole. It’s a wee bit banged up (probably happened during transport) and didn’t come with a box, but the latter is now solved.

The lid and bottom came from a wardrobe back panel I got from a mate, the sides were rocket sticks in their former lives. I found some scrap of felt in our material store and some hinges laying around in the drawers of my own workshop.

Unfortunately, the table saw teared up the plywood veneer fibres badly, even though I put tape around to prevent that. This is the first time it didn’t work. At. All. To cover that up, I painted the box with some decades old tinting paint (price tag says Deutsche Mark, not Euro!) from my paint cabinet. It’s awesome, works absolutely perfectly and doesn’t smell the slightest bit. I reckon, this caliper box is plenty good enough for occasional use at our scout material store.

I have a Python script that transforms the original YouTube channel Atom feed into a more useful Atom feed by removing the spam description and replacing it with the video duration, filtering out videos by title, duration, etc. I just updated it to exclude the damn Shorts garbage more efficiently. Finally, YouTube updated their Atom feed generation, so that the video URL contains /short/ if it’s of this useless kind. Never thought that they ever actually will improve their Atom feeds. Thank you, much appreciated!

After many weeks and probably at least a hundred hours of research, discussions and in-person viewing, I think I’ve finally come up with my Final Choices (shortlist) of a Hybrid Camper / Caravan that I think will suit my family and that I’ll enjoy (far less work for me to setup and teardown). The one at the top of the list I’m leaning towards os the SWAG SCT16 Family 4B #Camping #Campers

@bender@twtxt.net Finally! Let’s wait and see how it turns out. :-D

@prologic@twtxt.net Yeah, it’s not a strong sandbox in jenny’s case, it could still read my SSH private key (in case of an exploit of some sort). But I still like it.

I think my main takeaway is this: Knowing that technologies like Landlock/pledge/unveil exist and knowing that they are very easy to use, will probably nudge me into writing software differently in the future.

jenny was never meant to be sandboxed, so it can’t make great use of it. Future software might be different.

(And this is finally a strong argument for static linking.)

Sweat and tears but I finally tamed this beauty on Fedora/CentOS https://www.wireguard.com/netns/#the-new-namespace-solution

❤️ 🎶: I think I finally know love by J-Cera

I was wondering: What the heck is the light on my boot!? Turns out between sock and shoe tongue was a firefly, unbelievable! ;-D I’ve no idea how that happened. After untying, it took me five attempts to finally get it off. How crazy!

Watching several hundred glowworms tonight did not get boring. It’s just so damn cool. :-)

Finally, the two drawers are mounted on the workbench. Some kind of a lid board on top to keep the dust out is still missing. I also gotta build the drawer inserts for the saws.

I upcycled decades old table football aluminium pipes to become my handles. The spacers are made from the inner tube. Two minutes of handsanding with 400 grit sandpaper polished it up nicely.

https://lyse.isobeef.org/tmp/hobelbankschubladen/

Achievement Unlocked: I have finally gotten to sit my ass down and watch Back to the future from start to the end.

i ordered some fun colorful new minidiscs so i can finally get back to recording my mixes :D looking forward to it

🏀 OKC vs IND 🔥 #NBA #Finals

@lyse@lyse.isobeef.org Nice! The final desk looks like it’s right out of Skyrim. 😃

🏀 IND vs OKC ��� #NBA #Finals 🔥🔥🔥🔥

🏀 IND WIN 🔥 🔥 #NBA #East #Finals

@thecanine@twtxt.net with this you meant Conversations, not XMPP, right?

“Also, finally getting full screen view for avatars in XMPP - a better integrated one, after 25 years. Y@ay!”

QR codes, already posted about them in the last two posts, but I want to hear your hot takes: Should they only be black and white, are they even worth doing in 2025, incorporating them into things,..?
Also, finally getting full screen view for avatars in XMPP - a better integrated one, after 25 years. Y@ay!

🔥 MIN vs OKC 🏀 #NBA #west #finals

🔥 IND vs NYK 🏀 Game 2 https://0x0.st/8xin.png #NBA #East #Finals