Studies don't mean anything when newer models that change the game come out every two months.
I'm pretty in the middle on this, I agree on the expansion, I disagree on the skills that will be needed.
I think that AI has changed the nature of programming so much that it will not be the same going forward, and the "expansion" will be a titanic shift of ALL WALKS OF LIFE towards more technical skills to 'develop', 'configure', 'manage' these AI tools, where the engineering work is for the building building the underlying frameworks.
I don't think this is a bad thing for learning to program, programming skills (or at least the thinking skills behind the code) might become table stakes for a lot of jobs that currently aren't considering 'engineering' roles.
Engineering might eat the world, not software.
I can give me personal opinion on this comment only that is unrelated and not an statement regarding Codesmith and or its leadership.
I hate others say "this time is different" but I feel like that's the case...
If an engineer produces 10X the raw output (and potentially more) for the same salary, then even in a world where there is demand for the skill of programming, we might just not need as many.
Like if the productivity output is so much higher, I think it can be enough to both see an expansion of the need for code-writers but not a need for many more engineers.
Maybe 50% of the engineers we have now can write all the code humanity needs, and the others need to do a new job.
I can't give my opinions on him but I can share opinions in summarizing the podcast that the overall discussion doesn't propose any concrete solutions or have any concrete examples. It's mostly him talking about himself and the host talking about himself and extrapolating to high level opinions about the industry.
One of the points the founder makes against this is that new people to the industry were effectively being paid to learn and develop tacit knowledge and eventually they do and are a valuable engineer. And he is concerned that with AI there will be no reason for a company to hire someone in those early phases because the math doesn't add up anymore (AI replaces all of their value for $100 a month instead of $10K a month). I believe he mentions he thinks the government needs to step.
Codesmith Founder on becoming a SWE in the world of AI: "I don't know what the route is to that level [of tacit knowledge] for people not already in the system".
Source Podcast (March 12th, 2026): [https://www.youtube.com/watch?v=eggWeDjCFdA](https://www.youtube.com/watch?v=eggWeDjCFdA)
I’m currently researching a 2024 lawsuit involving Codesmith and its investors, so I’m not sharing opinions on Codesmith right now.
Other direct quotes from the discussion:
“domain knowledge is built by experience”
“if more of the programmatic building is done by AI, how do you build the tacit knowledge?what's the route in for people”
Anthropic's research indicates AI is indeed replacing human programmers already and likely will replace far more soon: [https://www.anthropic.com/research/labor-market-impacts](https://www.anthropic.com/research/labor-market-impacts)
Codesmith's blog and Bloomtech's blog are marketing blogs and not a source of fact. Anthropic's research is academic research.