migueluliMay 7
tante

"There are no more juniors. There was a funeral for their passing in 2024. Nobody came. The machine does what they do now, but cheaper. Of course, juniors weren't valuable for what they produced, they were valuable for who they would become: the senior engineer who knows where the bodies are buried. We optimized for output, and abolished apprenticeship. A few years from now, we'll wonder where all the seniors are. We shot them. Nobody will remember."

https://www.stvn.sh/writing/programming-still-sucks-fqffhyp

Programming Still Sucks. — Writing

Sorry Peter. — I'm at a birthday party, and while most people here also work in tech, there's always a Guy with a Real Job. You know, a physical job, building some or other thing people need. And this Guy always asks some variant of the same question: aren't you worried AI is taking your job? I glance around and see a few faces turning around toward us, rolling their eyes ever so slightly before returning to their previous conversation. Yes, this question again.

May 7 at 11:44amWeb
Show threadWhatisgoingonMay 7

@tante I never imagined that I would become one of the fabled mythical elder engineers just because there are no juniors following.

I will tell stories of hand written code around the campfire ...

Show threadPaul - Antifa. LGBTQ+ safe.May 7

@themipper Right there with you...

@tante

Show threadPoligofsky 🇨🇦 .. 🚫🤖May 7
@tante They have gone underground. They will return when needed.
Show threadMathias HasselmannMay 7
@tante Well written.
Show threadPoligofsky 🇨🇦 .. 🚫🤖May 7

@tante Why are managers—self-proclaimed “leaders”—the most vulnerable to group think? Or is it that group-thinkers are the most likely to go into management, with the wrong understanding, and the wrong incentives?

Meanwhile, the investors—assholes randomly selected from the same pool—are chasing unicorns, but don’t know what one looks like. So they throw mountains of cash at donkeys wearing dunce caps, until one of them shits out a golden bar of digital heroin.

Show threadRussellMay 7

@8r3n7 @tante

Incentives are aligned to the production of incentives. If it's easier to fake work than to actually work then folks are further incentivized to do less for more (or do more for more). It's also easier to ignore a problem than to fix it (in the short-term).

I think it's because humanity is already at or past the point of becoming post-scarcity. Part of that is society struggling to let go of the impulses and incentives born from scarcity.

Show threadMay Likes TorontoMay 7
@tante "The doll catches fire" is a perfect description of the state of product management over the past... 5 years. Maybe longer.
Show threadToni AittoniemiMay 7

@tante Yup. AI didn’t kill our jobs. Greed did. The same greed that outsourced everything to china and told us there’s be so much money ”trickling down” we wouldn’t need pensions..

Just a normal day in a neoliberal fiefdom called ”a democracy”.

Show threadDumbotron of LoxleyMay 7
@tante Captainpalooza 2025... lol, good story!
Show threadkramakerMay 7

@tante Probably the most literate summary of our times I've read. Nice work. Anyone that can wield natural language like this would have been a force with formal language(s) and context-free grammars.
You might enjoy: https://arxiv.org/abs/2601.05280v2

I certainly am.

On the Limits of Self-Improving in Large Language Models: The Singularity Is Not Near Without Symbolic Model Synthesis

We formalise recursive self-training in Large Language Models (LLMs) and Generative AI as a discrete-time dynamical system. We prove that if the proportion of exogenous, externally grounded signal $α_t$ vanishes asymptotically ($α_t \to 0$), the system undergoes degenerative dynamics. We derive two fundamental failure modes: (1) \textit{Entropy Decay}, where finite sampling effects induce monotonic loss of distributional diversity, and (2) \textit{Variance Amplification}, where the absence of persistent grounding causes distributional drift via a random-walk mechanism. These behaviours are architectural invariants of distributional learning on finite samples. We show that the collapse results apply specifically to closed-loop density matching without persistent external signal. Systems with non-vanishing exogenous grounding fall outside this regime. However, mainstream Singularity, AGI, and ASI narratives typically posit systems that become increasingly autonomous and require little to no human or external intervention for self-improvement. In that autonomy regime, the vanishing-signal condition is satisfied, and collapse follows under KL-based objectives. To overcome these limits, we propose neurosymbolic integration based on algorithmic probability and program synthesis. The Coding Theorem Method (CTM) enables identification of generative mechanisms rather than mere correlations, escaping the distribution-only constraints that bind standard statistical learning. We conclude that fully autonomous recursive density matching leads to degenerative fixed points, whereas externally anchored or mechanism-based approaches operate under fundamentally different asymptotic dynamics.

arXiv.org
Show threadThird spruce tree on the leftMay 7

@tante

This is fantastic and unfortunately   true.

I am Sara, tunnelling under Mordor with a USB stick. I have attempted to document the cron job and institutionalize the periodic nudge it needs to run payroll... but then I get yelled at for not videcoding enough new featureslop. There are no juniors for me to explain the cron job too.

Perhaps the AI may one dayabsorb the wiki page about the cron job. Hopefully someone else thinks to ask the AI about why payroll didn't run.

Show threadDarwin WoodkaMay 7

@tezoatlipoca @tante

There are no more spoons, we sold them all to pay for the AI

Show threadStephen 🌈 (he/him)May 7
@tante So right…
Show threadSteven 🏳️‍🌈May 7
@tante oh hey I wrote this! Thanks for sharing!
Show threadArmin HanischMay 7
@stevendotjs @tante Thanks for writing this and congratulations on your craftsmanship with words! Hits the perfect tone and while English is not my first language, reading this was pure bliss.
Made it out of programming (it’s just a hobby now) and glad about that. 😎
Show threadSteven 🏳️‍🌈May 7
@Linkshaender @tante ahw thank you! Happy for you you made it out! If you wanna tell me about it, I'm all ears.
Show threadArmin HanischMay 7
@stevendotjs @tante I‘ll do, just need a little time to explain why I actually still like to program as a hobby and why you’re right nonetheless.
Show threadArmin HanischMay 10

@stevendotjs @tante Wrote a blog post about the reason programming is still fun, but just a hobby for me now. Let's call it a shift in seeing purpose. It's all about people.
As English ist not my native language I tried my best and then used Deepl and Kagi Translate to make it into something readable (both of them had astonishingly few corrections 🤗). Let me hear what you think of this, please.

https://www.arminhanisch.de/2026/05/programming-is-now-only-a-hobby/

Why Programming Is Still Fun, but It’s Just a Hobby Now

Why Programming Is Still Fun, But It’s Just a Hobby Now I used to be a software developer. I even trained other software developers and gave talks at conferences. For a few years now, I’ve only been programming as a hobby. This wasn’t an active “escape” from software development, but rather a realization and a desire to learn something new again. Keep reading if you want to know what I actually find interesting about programming and why coding ("vibe" or not) is only a small part of it.

Show threadSteven 🏳️‍🌈May 11
@Linkshaender @tante ok I made a start and it's excellent so far! I'll read more tonight when my little one is in bed.
Show threadArmin HanischMay 11
@stevendotjs @tante Take your time, our „little one“ turns 21 this summer and if I think back it can only be weeks ago she learned to walk. 😆😂
Show threadtanteMay 7
@stevendotjs it's fantastic! Thanks for writing it
Show threadJoy_intlMay 7

@stevendotjs i applaud you, kind sir
- I hilariously joined a web dev program in 2022, and when I completed it in 2024, I could see it was pointless trying to get a job as a Jr..

Your story resonates (also bc I'm a career changer so old enough to have other experiences that are similar) - it's both hilarious and stark.

Well done 👏

@tante

Show threadSteven 🏳️‍🌈May 7
@Joy_intl @tante I... I'm so sorry. I'm glad you realized what was happening, cut your losses and did the right thing, but damn that's harsh.
Show threadJoy_intlMay 7

@stevendotjs i feel for my cohort who tried to move forward by going for their masters.. I hope it works for them. I hope there are enough companies who see the mess AI creates cannot be trusted.. but humans tend to take the path of least resistance..

In truth, I was never a coder, so let's hope I can find other ways to feed myself and my family.. it's rough out here for everyone.

@tante

Show threadSteven 🏳️‍🌈May 7

@Joy_intl @tante it really is, I have a fiance and newborn to provide for as well, and it's making me anxious.

Take care ❤️ reach out for help if it becomes too much please 🙏

Show threadJoy_intlMay 7

@stevendotjs That is very kind. I return your kindness and encouragement. We humans do better when we act in solidarity.

✊️

@tante

Show threadD2May 7

@stevendotjs kudos. so very on point.

Having ridden out the implosion of the dot-com crash, I’m staring at all this buildout and wondering what this time smells like. Capital spend crushed a few big names. Example: Lucent went to a nickel on the dollar.

Show threadSteven 🏳️‍🌈May 7

@InkomTech oh yeah, pulling up a chair to roast marshmallows over the bonfire.

And thanks ❤️🙏

Show threadD2May 7
@stevendotjs I don’t fear dioxin nor microplastics, but I ain’t eating anything cooked over that blaze.
Show threadDale HagglundMay 7

@stevendotjs @tante Great article... I checked out out the link and the pckt.blog page mentioned there but I only saw this article: I'd love to read more of your writing, either past or future.

I started work as a full-time software dev in 1988 (or a few years earlier if you count summer jobs), and I was lucky enough to enjoy it pretty much through to retirement a few years ago, right around the first chatgpt release. One of the parts I found most personally fulfilling was mentoring younger devs (which definitely included me learning from them and them challenging my old-guy reflexes) and to see what's happened in the very few years since is profoundly discouraging...

And thanks for @tante for forwarding this...

Show threadSteven 🏳️‍🌈May 7

@DaleHagglund @tante oh man yeah congrats on getting out in time!

My site is honestly barely half finished, I wasn't really expecting this to go anywhere, and this was the first thing I published in years. I'll be sure to keep writing ❤️🙏

Show threadDale HagglundMay 7

@stevendotjs @tante Thanks, and I'll look forward to more articles. As far as retiring, I can't claim any any special foresight; it was just the right time to leave for various reasons. Without at all realizing it at the time, I just happened to be in the industry at a time when software devs had substantial power versus management, if we chose to use it. It was much closer to retirement, and more-so in the years since, that I started to realize that was historically unusual, and changing rapidly for the worse in software.

Every now and then I read articles by a couple software dev bloggers I trust who are, I think, looking carefully at *how* to use LLMs in a way that might (just maybe) set devs up to be "centaurs", ie the drivers of technology, versus Doctorow's "reverse-centaurs", ie, meat-puppets directed by the system a la Amazon drivers. I don't have any particular faith that those can be achieved in any broad sense in the current context of the industry, though...

Show threadFelix EckhardtMay 7
@tante we might not need any human coder, when no senior is available anymore. It was similar with other professions.
Show threadNfoonfMay 7
@felix_eckhardt @tante this is the same like saying there is no difference between a taylormade suit and 2$ underpants made by slave labor in Bangladesh.
Show threadFelix EckhardtMay 7
@Nfoonf @tante yeah, that's a fair comparison. Doesn't mean it isn't going to happen.
Show threadTor Iver WilhelmsenMay 7
@felix_eckhardt @tante Yup, no one becomes cooks any more after we got microwave meals in the freezer section. Why bother?
Show threadFelix EckhardtMay 7
@toriver @tante I would compare this to the job of a weaver. The waevers became machine assistants and at some point the job just disappeared. Not saying this is a good thing or makes the world a better place. But i see the possibility that this will happen.
Show threadJevon Wright 💙💛May 7
@toriver @felix_eckhardt @tante except we still need cooks to come up with recipes for those microwave meals...
Show threadDMTomMay 7
@tante heh - my wife, who (encouraged by our friend who did the same switch ~5/6 years ago) decided at the end of 2021 to quit her supply chain management career and become a developer (so we both could have remote jobs ant travel) - ignored the funeral and tried, and tired and tried till 2025, when she said fuckit and became qa engineer … just in time to train the machines 🤦‍♂️
Show threadBill, organizer of stuffMay 7
@tante Gawd this is good.
Show threadD2May 7
@tante #noLiesDetected #notHyperbole #whyAmILaughing
Show threadquixoteMay 7
@tante And if anyone does remember they will have eyes rolled at them for being olds....
Show threadFlipper 🐬🏳️‍🌈May 7

@tante i was discussing the same thing this morning. AI coding is an extractive technology; you do not learn by using Claude code.

If you're going all in on AI coding tools, then you need senior engineers who have learned everything, so you can mine that experience. Juniors become LOC generators and are never going to become senior engineers.

For the first time since I started this career, nearly 30 years ago, I can't wait to be out of it.

Show threadRémi LetotMay 8
@tante
I don't understand how this is not more obvious to more people... no matter the field, if you have no juniors now, you won't have seniors later...
Show thread3Jane Tessier AshpoolMay 10

@tante well that was f’ing bleak. And they’re trying to do it to my non-technical but nevertheless SME team of CAPA specialists. Let them use AI to write Five Whys. Let them us AI to judge if the record is compliant. Let them write their self evils and goals using AI agents. What could go wrong? Except that doesn’t teach you what good looks like or how to write a rationale. AI can’t write a rationale; it can write things that look like one, that’s what it’s good at. But rationales rely on good thinking.

I’m just hoping we outlive this idiotic AI bubble but I’m afraid it may be here longer than that.

I’m afraid of interviewing for a new job, should a good opportunity come up, because I know they’re gonna ask how I feel about AI or how I’ve used it to solve problems in the past. And all I’ve got is, there are good applications of AI but I feel we need to look for the problems and then consider AI to solve them, rather than try to apply AI as a solution to every problem. And if they’re bought in to the lie, IDK how that will go.

Show threadDuncan BluesMay 27
@tante I feel that story so very much. 😩