The “hard truth” sounds like “AI isn’t replacing programmers anytime soon.”
A nice summary provided by the authors:
How developers are actually using AI. Very different usages for “bootstrappers” versus “iterators.” Perhaps a reason why one tool is unlikely to work equally well for both groups?
The 70% problem: AI’s learning curve paradox. Lesser-talked-about challenges with AI: the “two steps back paradox,” the hidden cost of “AI speed,” and the “knowledge paradox.”
What actually works: practical patterns. AI-first draft, constant conversation, and “trust but verify” patterns.
What does this mean for developers? Start small, stay modular, and trust your experience.
The rise of agentic software engineering. A shift to collaborating with AI, multi-modal capabilities, autonomous but guided approaches, and an “English-first” development environment.
The return of software as a craft? The lost art of polish to return, and the renaissance of personal software.
Additional thoughts. A good time to refresh what software engineering really is and how it has been the dream of needing no developers since the 1960s. And still, demand for experienced engineers could well increase in the future, rather than decrease.
So when AI replaces junior devs as a partner for the boring stuff for seniors, it will be interesting to see how juniors can even get to a senior level. I guess that whole reviewing, understanding and refactoring AI generated code needs to be a guided learning by a senior.
The idea that junior devs will go away is a lie theyll feed us till the end. My job is to dictate how these tools should be used for hundreds of employees and there’s a huge clash in expectations vs reality.
The AI slop should help the juniors, make them more useful “out of the box” by essentially rewording and customizing a google search answer, and save time for seniors explaining. Even the best models are lying and cannot efficiently have all this context, not any time soon. It currently takes enormous amounts of power to do what a brain can do with the energy equivalent of a peanut. That’s just not cost effective.
The only reason they are able to do it is because they are dumping stolen profits disguised as investments and “a bad economy/inflation”. Theyre essentially doing this off the back of tax payer. Once electricity starts costing what it should cost right now without the massive investments, it would never work. The current elite will take us for a ride until the next schtick though, and we’re all worst off for it. Senior engineers are just doing 1.25x the work and C suite is lying about AI.
Let’s hope it does a metaverse / web 3
I’d say a lot of this isn’t really sustainable. Also you need to strike some balance. The senior devs also can’t spend substancially more time correcting the AI’s mistakes. I mean AI time is practically for free. But if that weighs down on the senior devs, they’re the ones with a decent salary. It has to make them more efficient, or in the end it’ll be more expensive.
Also same applies to the rest of the world. Sure we’d need to change how junior/senior works. But I’d argue we’d also need to change the whole society if entry level jobs and low income jobs get replaced by machines. I mean these people need food on their tables, too. So we’d really need to change workplace hierarchy, education, plus the entirety of how salaries and income works.
What actually works: practical patterns. AI-first draft, constant conversation, and “trust but verify” patterns.
So… being an expert who communicates, being an expert in copying templates, and being an expert who knows what they’re doing?
Yes, it turns out that “Effective Leadership,” contrary to what people selling you books about AI and business are saying, remains the same as it’s been for thousands of years:
- Listen to the people who do the work
- Support them in doing their work
- Get yourself and other obstacles out of the way of people doing the work
Ah but that won’t sell books or convince the Board that you and you alone have the Sauce which will counteract all collective human knowledge and make them infinite money.
Salesbros and SEO and the takeover of tech by business have utterly destroyed what it means to be an engineer, an expert, or even just a person.
D
