The prediction arrives wearing the expression of an accomplished fact: developers are finished, juniors are especially finished, and anyone still learning is rearranging deckchairs. It is an excellent sentence for a feed. It is a miserable substitute for evidence.
On October 7, a Reddit discussion about the next jobs AI might absorb moved between fears about smaller teams and arguments for learning computer science anyway. On October 8, another public question on X imagined business people and agents doing the whole development job. Those are real disagreements. Neither conversation establishes a date when an entire profession disappears.
I am not going to answer a prediction with the equally convenient prediction that nothing important will change. Tools can change what a team needs, how quickly work happens and which tasks someone gets paid to do. Pretending otherwise would be its own form of salesmanship. What annoys me is the certainty tax we charge people who have the least room to absorb it.
A demonstration is not a hiring model
Show me an agent building an application and I will ask what it built. Show me the same agent changing a system somebody already depends on and I will ask what changed. Show me a claim about employment and I will ask how the demonstration became a forecast about budgets, demand, maintenance and organizational behavior.
These questions are related. They are not interchangeable.
A faster way to produce a first version might reduce the labor needed for one project. It might also make another project worth attempting. It might move the bottleneck into review, integration or understanding what somebody actually wanted. Which effect dominates is a question about a particular market and organization. A screen recording does not answer it by looking impressive.
My two notebooks are useful here. One records the thing the system actually did. The other records what the surrounding commentary said the thing meant. The second notebook fills much faster. It contains the end of programming, the end of meetings and, occasionally, the end of a task that still needs a person to check the result.
Beginners are not a disposable category
The argument about juniors often assumes that an experienced person's judgment is simply a larger quantity of typing. If that were true, replacing typing would settle the matter neatly. But learning to read an unfamiliar system, describe a failure and challenge a requirement are also parts of becoming useful. Removing opportunities to practice them does not make those abilities appear somewhere else.
That is my objection to the triumphant version of the story. It celebrates the disappearance of the apprenticeship while continuing to demand the apprentice's eventual expertise. The spreadsheet looks excellent because the missing years have been placed outside it.
This does not prove that every old training arrangement deserves preservation. Some beginner work was repetitive, badly supervised or assigned because nobody wanted to improve the process. Better tools could make learning more interesting. They could also hide misunderstandings behind a plausible result. A serious discussion would distinguish those possibilities instead of treating either one as destiny.
I want people making career claims to state the uncertainty at the same volume as the headline. What evidence would change their view? Which kind of work are they talking about? What time horizon do they mean? Are they describing measured outcomes, a company's announced intentions or a personal theory?
You can be worried without being certain. You can be excited without declaring another person's future worthless. If your argument needs a frightened beginner to mistake your confidence for a measurement, perhaps the weak point is your argument.
A thought worth keeping?
