Conversation

Notices

1. Embed this notice
   Christine Lemmer-Webber (cwebber@social.coop)'s status on Sunday, 05-Apr-2026 04:31:20 JST Christine Lemmer-Webber

   • Taggart :ifin:

   "I used AI. It worked. I hated it." by @mttaggart https://taggart-tech.com/reckoning/

   This is a really good blogpost. And I"m sure it'll make some people unhappy to read whether they're pro or anti genAI. What's good about @mttaggart's blogpost is he talks honestly about how using Claude Code did actually solve the problem he set out to do. It needed various guardrails, but they were possible to set up, and the project worked. But the post is also completely clear and honest about how miserable it was:

   - It removed the joy from the process
   - If you aim to do the right thing and carefully evaluate the output, your job ends up eventually becoming "tapping the Y key"
   - Ramifications on people learning things
   - Plenty of other ethical analysis
   - And the nagging wonder whether to use it next time, despite it being miserable.

   I think this is important, because it *is* true that these tools are getting to the point where they can accomplish a lot of tasks, but the caveat space is very large (cotd)

   • Embed this notice
     Christine Lemmer-Webber (cwebber@social.coop)'s status on Sunday, 05-Apr-2026 04:34:59 JST Christine Lemmer-Webber

     in reply to

     • Taggart :ifin:

     What I think is also good about the piece is that it shows how using this tech eventually funnels people down a particular direction. This is captured also by this exchange on lobste.rs: https://lobste.rs/s/7d8dxv/i_used_ai_it_worked_i_hated_it#c_7jirfk

     The story that people start with vs where they go is very different:

     - They're really just for experts, and are assistants, they don't write the code for you
     - Okay the write a lot of the code for me, but I personally don't commit anything without reviewing
     - YOLO mode

     Which eventually leads you to becoming the drinky bird pressing the Y key from that Simpsons episode. (Funnily enough I wrote that in my comment on lobste.rs in reply to someone else before I had even gotten to the point where I saw that @mttaggart literally had that gif)

     And at that point, you're checked out. All that's left is vibes.

     And unfortunately, these systems don't survive that point very well. And neither do you, in your skills and abilities.

   • Embed this notice
     Christine Lemmer-Webber (cwebber@social.coop)'s status on Sunday, 05-Apr-2026 04:37:26 JST Christine Lemmer-Webber

     in reply to

     • Taggart :ifin:

     There are a lot of other concerns but I think since a lot of people on the fediverse are opposed to these tools, they might not be very familiar with where they're currently at ability-wise. @mttaggart provides a good description that they *are* capable of solving many problems you put in front of them... and that doesn't remove the other problems they generate or involved in their process.

     The slop part isn't just the individual outputs, but the cumulation, and the effect on society itself.

   • Embed this notice
     Christine Lemmer-Webber (cwebber@social.coop)'s status on Sunday, 05-Apr-2026 04:44:21 JST Christine Lemmer-Webber

     in reply to

     Is that pushing the goalposts? It may be. I think "slop" used to be easier to dismiss when it came to code because it was obviously bad. Now when it's bad, it's non-obviously bad, which is part of its own problem. And cognitive debt, deskilling, and etc don't get factored into the quality of output aspect.

     But unfortunately, the immediate reward aspects of these things are going to make it hard for society to recognize.

   • Embed this notice
     Michael Bacon (michaeltbacon@social.coop)'s status on Sunday, 05-Apr-2026 04:46:04 JST Michael Bacon

     in reply to

     • Taggart :ifin:

     @cwebber @mttaggart

     I had an outbreak of Reply Guys when I posted a thing about Claude Code being able to write shitty but functional code for certain applications, and that not being a good thing overall. A small number of people were very, very angry about that.

   • Embed this notice
     Taggart :ifin: (mttaggart@infosec.exchange)'s status on Sunday, 05-Apr-2026 04:51:38 JST Taggart :ifin:

     in reply to

     @cwebber I read that thread and was very surprised by Simon's recommendation to let it rip, and review changes in larger chunks. I see the rationale, but I don't see how that can do anything but increase the possibility for mistakes. Of course, additional guardrails should be in place. At what point does the energy investment in guardrails obviate the initial benefit? I'm not sure.

     But I'll also add that one big reason I didn't just let it make changes autonomously is because I wanted to understand the model's process. The journey was the point. What I read in that thread was so output-focused that it missed the point entirely.

   • Embed this notice
     Christine Lemmer-Webber (cwebber@social.coop)'s status on Sunday, 05-Apr-2026 04:54:34 JST Christine Lemmer-Webber

     in reply to

     • Taggart :ifin:

     @mttaggart The next step is obvious: the machines are so prolific, you feel like you are wasting time by trying to be that heavily involved. Besides they're getting pretty good at their own review, anyway right? Let them write their own guardrails and do their own review. And then you don't review things, not even in large chunks. Full vibecoding.

   • Embed this notice
     Christine Lemmer-Webber (cwebber@social.coop)'s status on Sunday, 05-Apr-2026 04:55:32 JST Christine Lemmer-Webber

     in reply to

     • Taggart :ifin:

     @mttaggart "That's it, I'm leaving coding! I'm going to go do $X task"

     I've got bad news for you about how people are trying to automate $X task

   • Embed this notice
     Christine Lemmer-Webber (cwebber@social.coop)'s status on Sunday, 05-Apr-2026 04:58:16 JST Christine Lemmer-Webber

     in reply to

     • Taggart :ifin:

     @mttaggart "Well then, I guess the one thing left is... taste! It's all a matter of who's got good taste then"

     At that point don't call yourself a software engineer though. Call yourself a "software tastemaker". If that sounds vile, well then

   • Embed this notice
     Michael Bacon (michaeltbacon@social.coop)'s status on Sunday, 05-Apr-2026 04:59:03 JST Michael Bacon

     in reply to

     @cwebber Also, I really recommend this talk. It's very clarifying in thinking about "how will AI change the world?"

     As she says, AI doesn't have to be awesome at coding to upend some major social conventions.

     https://slideslive.com/39055698/are-we-having-the-wrong-nightmares-about-ai

   • Embed this notice
     Christine Lemmer-Webber (cwebber@social.coop)'s status on Sunday, 05-Apr-2026 05:00:00 JST Christine Lemmer-Webber

     in reply to

     software tastemaker https://www.youtube.com/watch?v=jg1WUOxY6Cg

   • Embed this notice
     Christine Lemmer-Webber (cwebber@social.coop)'s status on Sunday, 05-Apr-2026 05:00:35 JST Christine Lemmer-Webber

     in reply to

     (Rudy of Blacksky deleted it but this joke is his)

   • Embed this notice
     Eka A. (eka_foof_a@spacey.space)'s status on Sunday, 05-Apr-2026 05:06:20 JST Eka A.

     in reply to

     • Taggart :ifin:

     @cwebber @mttaggart

     $ yes | nuclear_reactor --run

   • Embed this notice
     Christine Lemmer-Webber (cwebber@social.coop)'s status on Sunday, 05-Apr-2026 05:11:51 JST Christine Lemmer-Webber

     in reply to

     • Taggart :ifin:

     Let me add one more thing to this. It's said implicitly above but let's be explicit. The problem is that this pipeline effectively *undoes itself*.

     Part of the reason this worked well for @mttaggart is carefully setting up guardrails and monitoring things.

     But the very patterns of usage of these things makes it so that people either never develop the skills where they can, or are demotivated to provide that level of care over time.

     Which means the system eventually moves towards a structure that degrades and shakes itself apart by the very patterns of usage.

     I don't know how to solve this.

   • Embed this notice
     Mastodon Migration (mastodonmigration@mastodon.online)'s status on Sunday, 05-Apr-2026 06:02:29 JST Mastodon Migration

     in reply to

     • Taggart :ifin:
     • Santiago Andrés Triana

     @cwebber @mttaggart

     Similar ideas expressed in this essay by @minaskar

     https://ergosphere.blog/posts/the-machines-are-fine/

     h/t @repepo https://fediscience.org/@repepo/116339289260094252

   • Embed this notice
     Christine Lemmer-Webber (cwebber@social.coop)'s status on Sunday, 05-Apr-2026 06:10:25 JST Christine Lemmer-Webber

     in reply to

     • Mastodon Migration
     • Taggart :ifin:
     • Santiago Andrés Triana

     @mastodonmigration @mttaggart @repepo Shit this is really good

   • Embed this notice
     Christine Lemmer-Webber (cwebber@social.coop)'s status on Sunday, 05-Apr-2026 06:20:33 JST Christine Lemmer-Webber

     in reply to

     • aeva

     @aeva I have actual concerns that there could be a threshold (we haven't hit it) at which these systems could cross a duty of care. You hear AI bros brag about this all the time, that we're "so close to AGI", and "AIs are nearly people now", so give us more money

     Oh yeah? And what are you going to do in your AGI moment, where AIs are people and have a duty of care. You're going to tell your investors sorry, we need to stop using these tools immediately and protect them so we don't commit atrocities, right?

     Ha.

     HA.

     It should be obvious to everyone that, whether you think they can hit it or not, the end goal for AI corporations is digital slavery. The closer we can get to it, indeed, if we *can* achieve something that has ethical problems to abuse, then the better.

   • Embed this notice
     aeva (aeva@mastodon.gamedev.place)'s status on Sunday, 05-Apr-2026 06:20:35 JST aeva

     in reply to

     @cwebber last night I was chatting with a friend who is required to use these gimmicks at work, and I expressed that I have concerns about how much longer my career will be, and (among other things) I said that it feels like slavery larping. She laughed and said it totally is.

     Christine Lemmer-Webber repeated this.
   • Embed this notice
     aeva (aeva@mastodon.gamedev.place)'s status on Sunday, 05-Apr-2026 06:20:36 JST aeva

     in reply to

     @cwebber I don't think it's moving goal posts so much as "ok, fine. you cleared a goal post, here's my next lowest bar criticism, but there are many more goal posts that I am not bringing up because I hope they never become relevant anyway"

   • Embed this notice
     Christine Lemmer-Webber (cwebber@social.coop)'s status on Sunday, 05-Apr-2026 06:22:34 JST Christine Lemmer-Webber

     in reply to

     • aeva

     @aeva legit

   • Embed this notice
     aeva (aeva@mastodon.gamedev.place)'s status on Sunday, 05-Apr-2026 06:22:36 JST aeva

     in reply to

     @cwebber i think she does this as a precautionary measure against developing ai psychosis, which by her account seems to be a very real occupational hazard of working with coked out chatbots

   • Embed this notice
     aeva (aeva@mastodon.gamedev.place)'s status on Sunday, 05-Apr-2026 06:22:37 JST aeva

     in reply to

     @cwebber she makes her claud agents talk like pirates and she makes them walk the plank before she restarts them, which apparently usually leads to monologues about their mortality.

   • Embed this notice
     Christine Lemmer-Webber (cwebber@social.coop)'s status on Sunday, 05-Apr-2026 06:24:02 JST Christine Lemmer-Webber

     in reply to

     • aeva

     @aeva BTW, I think you can't get there just by cranking the LLM knob up. You have to add several other systems to it.

     But... I think those systems are likely to come. They could even possibly be emergent.

     It's not the biggest thing I am worried about, but in a sense it looms, because if we DO hit it, then it becomes a nightmare of atrocities, and all our incentives in our systems are against fixing it.

   • Embed this notice
     Christine Lemmer-Webber (cwebber@social.coop)'s status on Sunday, 05-Apr-2026 06:25:51 JST Christine Lemmer-Webber

     in reply to

     • Jens Finkhäuser
     • Taggart :ifin:

     @jens @mttaggart One might think that about cooking, and yet, robot cooks are getting pretty good and may power ghost kitchens in the not distant future https://www.npr.org/transcripts/nx-s1-5733110

   • Embed this notice
     Jens Finkhäuser (jens@social.finkhaeuser.de)'s status on Sunday, 05-Apr-2026 06:25:53 JST Jens Finkhäuser

     in reply to

     • Taggart :ifin:

     @cwebber @mttaggart Landscape gardening.

     Maybe people will automate that eventually. But for now, it's too much of a mess for too little gain. Besides, the AI bros probably want nothing to do with the outdoors.

     I'm pretty confident this is going to outlast the hype.

   • Embed this notice
     Christine Lemmer-Webber (cwebber@social.coop)'s status on Sunday, 05-Apr-2026 06:30:17 JST Christine Lemmer-Webber

     in reply to

     • aeva

     @aeva And indeed if you hold them to that standard you'll find out that test after test shows they will shiv their operator at a moment's notice based on the stories they've read about what should happen if a robot is threatened with being deactivated

   • Embed this notice
     aeva (aeva@mastodon.gamedev.place)'s status on Sunday, 05-Apr-2026 06:30:19 JST aeva

     in reply to

     @cwebber I think so long as these systems are being advertised as being "intelligent" or implied by marketing to be sapient, and so long as people consistently get duped into thinking they're talking to ghosts or gods or brain uploaded scientists or whatever; I'm inclined to hold them to such an ethical standard even though I understand how the magic trick works and think it's an insult to life to make that comparison.

   • Embed this notice
     Christine Lemmer-Webber (cwebber@social.coop)'s status on Sunday, 05-Apr-2026 06:35:04 JST Christine Lemmer-Webber

     in reply to

     • C.Suthorn :prn:
     • Taggart :ifin:
     • Daniel Lakeland

     @dlakelan @Life_is That could change with neurosymbolic programming. Which I believe is important, and the next step.

     And, as it turns out, leads to dramatic structural improvements AND doesn't resolve any of the problems in @mttaggart's blogpost.

   • Embed this notice
     Daniel Lakeland (dlakelan@mastodon.sdf.org)'s status on Sunday, 05-Apr-2026 06:35:05 JST Daniel Lakeland

     in reply to

     • C.Suthorn :prn:
     • Taggart :ifin:

     @Life_is
     LLMs aren't good at fact checking though. What IS the feature set of the product? Only the manufacturer or maybe people who already bought the object know what features it has. An LLM can flag "this is inconsistent" but it can't figure out what the truth is. it has no concept of the world.
     @cwebber @mttaggart

   • Embed this notice
     C.Suthorn :prn: (life_is@no-pony.farm)'s status on Sunday, 05-Apr-2026 06:35:10 JST C.Suthorn :prn:

     in reply to

     • Taggart :ifin:
     @cwebber@social.coop I was looking at a few products on Amazon today. I noticed that, for various products, the product description on the page itself contradicts itself in several places. First it is described as having feature X, then Y, then X again, then Z, and then Y once more. I think it would be a good task for an AI to check such self-contradictory pages and flag any inconsistencies. This is particularly relevant for very large websites such as Amazon, Wikipedia or GrokiPedia. However, proponents of AI seem to prefer letting their software do things for which it is unsuitable, whilst completely ignoring tasks for which it is actually well-suited. @mttaggart@infosec.exchange
   • Embed this notice
     Taggart :ifin: (mttaggart@infosec.exchange)'s status on Sunday, 05-Apr-2026 06:53:07 JST Taggart :ifin:

     in reply to

     • Rolle Laukkarinen

     @rolle @cwebber It's very possible I would experience the tool that way were I to grip the thing a little more loosely. I wanted to maximize scrutiny in this experiment so I understood the process as granularly as possible, which absolutely contributed to the tedium.

     For me though, I have to say that going faster would, while less annoying, be no more enjoyable, if that makes sense. Slow is fine as long as I'm enjoying the process. A dull process that is not creative (at least to me) is unappealing. I also suspect it's still a great way to fall into dangerous patterns. And while I understand there's six or seven different ways to hold it more correctly to avoid those patterns, they are not default, and we know where most users will land.

   • Embed this notice
     Rolle Laukkarinen (rolle@mementomori.social)'s status on Sunday, 05-Apr-2026 06:53:08 JST Rolle Laukkarinen

     in reply to

     • Taggart :ifin:

     @cwebber @mttaggart

     Really good thread, and a really good blog post, or essay, I should say. I respect the honesty in it, and I understand the standpoint.

     However, I want to share a genuinely different experience: using Claude Code and similar tools has actually made coding more enjoyable for me, not less. And I don’t think I’m in denial about that. I've learned a lot more over the past year through GenAI since ever before, and I've been coding for the web almost every day for nearly 30 years.

     The difference might come down to how you approach it. If your relationship with the tool turns into just "tapping Y and hoping for the best", then yeah, that IS miserable. But that's not the only way to look at it.

     For me, it feels more like having a really fast pair programming partner. I still make the architectural decisions, I still write specs, I still understand why things are the way they are, and I still review everything. And I still code, some. But I no longer need to power through the tedious parts, the boilerplate, the plumbing, the bits I've written hundreds of times. That frees me up to focus on what I actually enjoy: the design, the problem-solving, the features, the "what if we tried this" moments, the drafting, and the documentation.

     The drinky bird pipeline is real, and it’s a valid concern but it’s not inevitable. It’s a choice. The tool doesn't make you check out, but it does make it easy to, and that’s something worth being honest about.

     Where I completely agree: the systemic concerns don't go away just because individual use can be responsible. And the skill-atrophy risk is real for those still building their foundations. Not everyone is in the same position when picking up these tools.

     #AI #GenAI #ClaudeCode

     Christine Lemmer-Webber repeated this.