Addressing the question in the title, "What happens if an entire class of workers loses faith in their careers", look at what happened to printers. That was a good skilled trade for centuries. Then it just went away. Phototypesetting took the high volume end. Desktop publishing took the low end. Computer-controlled presses took press operators. And then the newspaper business died.
I've heard homeless former printers talking about the good days at the Burger King near the cable car turntable at Powell and Market. They thought they had careers.
Now they're struggling to survive.
Knitting is not the answer. This is a survival issue. Somehow I suspect the author inherited money.
The developer's astrology app containing tarot reading features was rejected from Apple's App Store that prohibits astrology apps.
He then replaced the app's content with a clone of, if not the identical open source astronomy app "Dark Hours", copying even the name. The original is here: https://darkhours.app
The developer was then in contact with John Gruber, who wrote an article claiming Apple's reviewers confuse a nice, scientific astronomy app with astrology. The developer reviewed the article before publication.
It then turned out that the developer forgot to mention that originally he did upload an astrology app for review. Gruber felt duped, and noted that it's unclear whether an updated build containing the new astronomy app was ever submitted to Apple for review: https://daringfireball.net/2026/08/retraction_app_store_reje...
Now the developer seems to claim that he merely accidentally cloned that open source app when working with Claude.
I believe there are some programming jobs in which the code is absolutely the easier part. Not all of us work in signal processing, integrated systems or have to push upstream to Linux kernel because the company we work for really needs a memory allocation optimization for its data centers.
Navigating customer requirements and building something that satisfies both market's needs and company strategy can be an incredibly difficult and frustrating problem to solve. Especially if you need to also oversee the execution of the strategy. So not only you have to predict what they want or know the domain deeply enough to understand what they say they want is not what they really want, you also have to come up with a plan for executing your solution in a corporate environment.
There is a reason that books like "the staff engineer's path" cover topics such as local maximums, communication, establishing support for executing a plan or creating alignment on big efforts. In large corporate environments with multiple international customers, code is most of the time not the hardest problem.
This is already the case for Master's degrees and above in Denmark. I've been the opponent examining students. We ask Master's students to give US instruction on a topic (drawn randomly on a card in front of them, from a list they know a few days in advance). They do a chalk and talk on a topic for about fifteen minutes with three to five professors being "dumb students". It works fantastically and at the end it's very apparent to both student and teachers how well a topic is understood. Marks are rarely contested and in my experience most students think the process is fair.
Of course, woe betide those with social phobias....
Just in the last few weeks, D'addario, one of largest guitar string manufacturers, posted on Instagram what is clearly an AI-song to advertise its new line of guitar strings.
After being called out on it being AI, they posted a mobile-recording of the song recreated in a DAW, but it didn't even match the timing or the notes played.
After being called on it AGAIN, they blamed it on LANDR and Apple Mastering "changing the notes", but those only do "AI knob tweaking", they're not generative-ai audio synthesis. LANDR itself replied.
After being called on it AGAIN, they finally caved and said one of their workers used Suno without their knowledge.
After someone contacted the worker, he said he was pretty clear in his communication with D'addario that he had used Suno.
EDIT: The worker claimed it was "his first time using AI" and "he didn't know", but after someone did some digging in his Instagram profile, there's a few Suno songs and a lot of AI posters and avatars.
This sort of ass-covering is getting crazy with AI...
> If coding is easy, how come programmers were in high demand, and have demanded large salaries for years (even before ZIRP)?
Because programmers have generally been forced to wear additional, invisible hats that are essential to making the code happen in the first place.
Writing code is not hard. Writing correct code is. Knowing what is correct in a setting with paying customers generally involves interacting with those customers. Either directly or worse. The gigantic salaries paid to the most prolific employees is not due to their ability to write code. It is due to their ability to interrogate the shit out of the customer until they finally reveal the true requirements.
Side comment about the difference between “My server is a phone now” and “My phone is a server now.”
In linguistics, the terms “theme” and “rheme” refer to the status of information in a sentence [1]. The theme of a sentence is something that is known or can be assumed from the context (that is, old information), while the rheme is information that is newly introduced in the sentence, often as a comment on the theme. The tendency in many (most? all?) languages is for the theme to precede the rheme in connected discourse.
Without context, a reader is therefore likely to assume from “my X is a Y now” that X is something the writer had previously and that Y is something new. Because most people have a phone but few have their own servers, I was initially confused by the title “My server is a phone now.” I had to read the article to realize that the word order is in fact appropriate: for the writer, the server is old information (something he has had for a while) while the phone (as a server, at least) is new.
In English, theme and rheme might be inferred not only from the position in a sentence but also from how a noun is marked. Possessives like “my” are generally used before nouns referring to things that are previously known or assumed; the definite article “the” is similar. In contrast, the indefinite article “a” (or “an”) usually marks information being newly introduced, so it is more common with rhemes.
I worked as a Japanese-to-English translator for many years. Because the grammatical structure of Japanese sentences is very different from that of English, sentence-by-sentence translation often ends up with paragraphs where the information flow is awkward and unnatural. I gradually learned that I could produce smoother translations by reorganizing my English sentences into an old-to-new information sequence.
"As is now generally admitted, over a limited range of operation, machines act far more rapidly than human beings and are far more precise in performing the details of their operations. This being the case, even when machines do not in any way transcend man's intelligence, they very well may, and often do, transcend man in the performance of tasks. An intelligent understanding of their mode of performance may be delayed until long after the task which they have been set has been completed. This means that though machines are theoretically subject to human criticism, such criticism may be ineffective until long after it is relevant. To be effective in warding off disastrous consequences, our understanding of our man-made machines should in general develop _pari passu_ with the performance of the machine. By the very slowness of our human actions, our effective control of our machines may be nullified. By the time we are able to react to information conveyed by our senses and stop the car we are driving, it may already have run head on into a wall."
"In neurophysiological language, ataxia can be quite as much of a deprivation as paralysis. A patient with locomotor ataxia may not suffer from any defect of his muscles or motor nerves, but if his muscles and tendons and organs do not tell him exactly what position he is in, and whether the tensions to which his organs are subjected will or will not lead to his falling, he will be unable to stand up. Similarly, when a machine constructed by us is capable of operating on its incoming data at a pace which we cannot keep, we may not know, until too late, when to turn it off."
EU data regions are a reflexive action by companies that try to hold on to their EU customers (and more and more are leaving, surprisingly the larger ones seem to be leading here). Realize that as long as you are still hosted on US owned infrastructure or that if there are US (or: five-eyes) owned companies anywhere in the stack your data can still be forcibly pulled and often without you being aware that this happened. There are only very few such stacks that are 100% owned by EU entities.
Something I've wondered... if you publicly say that a domain is for sale and someone has a trademark for it would you automatically lose in arbitration?
Around 1998 I registered a domain. Sony registered a trademark with the same name a few years after that. Someone on a Gmail account asked if he could buy it - I later found out he worked for Sony. I told him no - it's for a game I've been working on. I went on a vacation for a month and when I got back there was a fedex package filled with documents from Sony saying I'm violating their trademark and they'll take the domain.
I got a lawyer and he told me that I shouldn't offer it for sale as that would show the arbitration board that I don't need the domain. But he also told me that in order to fight the trademark it would probably cost a couple of hundred thousand dollars. So I could keep the domain but not use it for commerce...
In the end I ended up selling it to Sony but through my lawyer - I never stated it was for sale. But this was early on for domains and I wonder what the process is like now.
I don’t even think most of it is deeply philosophical.
They just got uberrich and now worry more about the concerns of the uberrich than the concerns they had when they were normies.
Rule of law, functioning government providing services, functioning public commons, free press, freedom of speech for all, any social safety net, etc don’t matter when they are so far removed from the public they can just go buy yourself an island/citizenship/president.
Suddenly small differences in marginal tax brackets, antitrust enforcement & limits on the ability to simply buy the government they want are the biggest encroachments on freedom they feel.
> Writing code is not hard. Writing correct code is. Knowing what is correct in a setting with paying customers generally involves interacting with those customers.
That's like saying "building a car is not hard, building a real car that you can use and that passes regulation is".
IOW, writing code is hard in every reasonable context.
> But that’s not what happened at all. Godier proceeded through a series of escalations up to the App Review Board and the Review Board responded that they determined the original rejection was valid because, I shit you not, “We understand that the app includes a live tarot reading feature.”
> It has no tarot function, no horoscopes, and nothing that I, or anyone else I’ve asked, would associate with astrology.
I have dealt with the Byzantine App Store bureaucracy before. What you do here, is you submit a new version, pretty much identical to the previous one, with the comment “tarot function removed”.
Everything in AI seems to be focused on LLMs lately. But in my opinion, powerful problem-specific models like this are even more interesting. The SOTA AI models used in weather forecasting are already outperforming the classic NWP models while being orders of magnitude more efficient (inference). Most are based on multi scale (hierarchical) Graph Neural Networks, an architecture which is not often talked about. The original Graphcast paper is worth a read if you think this is interesting: https://arxiv.org/abs/2212.12794
For anyone reading this and jumping to a broader conclusion: the Nixpkgs core team disbanding does not mean Nixpkgs or Nix is dying.
It does mean that this particular structure was not sustainable, very important contributors burnt out and we need to do better, faster. We need to continue learning from this and continue building a stronger ecosystem that prioritizes the contributors who are the only reason any of this is possible.
Personally, I'm sorry and grateful. Sorry that it ended in core folks being burnt out. Grateful since they did some of the most amazing work, more than anyone can imagine two people doing.
This was a big thing for drone people based out of Eppley AFB in Omaha. Wake up, drive to work, attack enemies, drive home, sit there eating dinner with your family trying to pretend you didn’t kill someone and can’t talk about it. Combat troops get to lean on each other after a firefight: “dude, good shot. That guy was about to get us!” You get validation that you did something to safe you and your friends, and it’s still a heavy burden, but you have people who were there telling you that you’re a good person and did the right thing.
The Air Force did something shockingly reasonable (according to the newspaper coverage): they require all pilots to attend mandatory counseling sessions. Then there was no sign of weakness or any of that toxic nonsense. “Ugh, guess I have to go talk to the doc today, and I don’t want to, but guess orders are orders.”
Part of my career encapsulates a period where I was a PLC programmer, installer, commissioner, and troubleshooter for massive build outs (sky scrapers, data centers, laboratories, factories). Interestingly, this was after years of teaching myself software engineering, eventually participating in large open source projects. The clash of entering the PLC world was _extremely_ harsh.
Let me give an example: I once worked with an integrator who was working on an AHU feeding an extremely critical portion of a datacenter (I was a lead by this point and mostly played babysitter). During certain points of the day you couldn't open the door to this room due to negative pressure because the logic was over-ramping the exhaust fans. As I watched this contractor work, I saw him open his laptop, with Windows on it (because Microsoft has had a death grip on this industry for decades now), and proceed to backup the PLC program into a massive folder with God knows how many other "customer projects" he was carrying around in this thing. He then proceeded to go do some physical checks in the field, came back, and prepared to upload the fixed program. As I watched, I noticed he _grabbed a backup from ANOTHER customer_ and I immediately had to intervene. Who knows what untold damage I saved from that single move.
I tell this story to demonstrate just how far into the dark ages this industry is. I vividly recall coming into the data center for a fortune 50 company, one everyone here would know, and being astonished that they had never heard of Network Attached Storage or RAID and why they might want to consider a disaster recovery plan for their multi-million dollar mechanical plant.
This industry is in _desperate_ need of strong technical help, but unfortunately the "higher ups" tend to be the same people who are "comfortable" with the way thing are and refuse to move. I literally tried for a decade before giving up and moving into software engineering proper.
So, anyways, just imagine the most archaic and barbaric set of IT software, controls, and procedures, dumb that down even further, and you've landed on the infrastructure/teams that operate probably half of critical infrastructure.
My own PC died and now I’m locked out of my entire Steam library for the foreseeable future. A $2000 PC is literally a downgrade from what I bought 10 freaking years ago.
Surprised no one has made the extremely obvious observation that it took literal centuries for printing technology to eliminate the trade of printers, which could not have possibly have employed more than 0.01% of people at its peak.
Compared to the potential obsolescence of 40% to 60% of the US workforce over the course of 10 - 20 years?
You can argue about whether the obsolescence is real or not, but if it is, the plight of the printers is totally incomparable.
Question: What happens when you drive a car at 90 mph into a concrete wall?
Answer: Well, actually just this morning pulling into my parking spot, I bumped into the barrier beneath my tires. It was fine!
I would be really curious to hear from devs at Databricks what the experience of development is like internally. I work at a small startup with essentially unlimited AI spend budget - the entire point is that I should be turning to it at every opportunity since our human labor is so expensive relative to tokens. So generally it's like:
- Spend most time prioritizing/discussing what to do.
- Once that's agreed, use Fable 5 High + 5.6 Sol XHigh come up with a design + plan. Agree on the high level plan. (Usually this just comes down to choosing where the change belongs on the spectrum between minimal patch <-> full redesign)
- Use Opus 5 or Sol Med to execute
- Auto-fix bugs and CI until green + thermonuclear review skill x3.
- Manual interrogation of change/nits
- Come up with QA plan and have Codex Computer Use execute on it
- Manually spot check the final result (usually a sizable diff, thousands of lines, complete feature E2E, etc)
I probably spend like $80 a day at least but I produce the output of 3 or 4 2022 engineers and probably at better quality. So it's easily worth it. Would I save money by switching to GLM 5.2 and such...perhaps? IDK. At our scale it's not worth the time spent building the eval harness to actually understand the performance tradeoff.
My grandfather was a printer not to far from that Burger King, in a building now owned by the academy of art. In 1963 he had 5 children, a house in Daly City and 3 apartments across the street from Dolores park as well as a brand new home in Terra Linda. He died early, but extended family stayed in the business. It was union work, he as the youngest operator of the 4 color press. His wife my grandmother didn’t work. The printing business was cooked by the early 90s. I was laid off from Broadcom a year and a half ago- I think about the printing industry to Ai analogue all the time. I got lucky and got a job a startup making less than half of what I made before. I know much more capable and smart people laid off in the same round- that are still wondering what the hell to do.
Farewell, Etaoin Shrdlu documents the last day of hot type at The New York Times before the digital changeover. It's one of my favourite documentaries, everything feels so vivid and alive:
The Nixpkgs Core team was established in Sept 2025 (https://discourse.nixos.org/t/establishing-the-nixpkgs-core-...), so it is a fairly new concept and idea. They've accomplished some good things as highlighted in the post, but are now stepping down. Yes, we'd prefer for the team to exist, but we've also functioned without one for ~20 years. It may take multiple iterations to bootstrap the concept and to figure out the right structure - or we may find it isn't needed. So no, this is not a critical emergency. It's a return to the status quo of late last year.
That being said, I think it is still a concept worth pursuing.
I've heard homeless former printers talking about the good days at the Burger King near the cable car turntable at Powell and Market. They thought they had careers. Now they're struggling to survive.
Knitting is not the answer. This is a survival issue. Somehow I suspect the author inherited money.