I see much more histeria and false but extremely high hopes, than real deal from where I sit (deceloper at faanglike high tec company).
Looks like the higher the management, the farther away from real engineering work — the more excitement there is and the less common sense and real understanding of how developers and llms work.
> Are you 10x more efficient?
90% of my time is spent thinking and talking about the problem and solutions. 10% is spent coding (sometimes 1% with 9% integrating this into existing infrastructure and processes). Even with ideal AGI coding agent id be only 10% more efficient.
Imagine a very bright junior developer. You still are heavily time taxed mentoring him and communicating.
Not many non technical people (to my surprise) get it.
Based on posts and comments here there are plenty “technical enough” people who don’t understand the essence of engineering work (software engineering in particular).
Spitting out barely (yet) working throwaway grade code is an impressive accomplishment for TikTok, but it has very little to do with complex business critical software most real engineers deal with everyday
On the contrary, I would class myself a mid to high skill dev, I have a CS degree and about 10 years of Java/C++/Rust/Python under my belt (focused on Financial Market applications).
I would consider myself today 2-3x more effective than where I was 12 months ago.
I can grok on a new code bases much faster by having an AI explain things to me only a grey beard could previously, I can ask Gemini 2.5 (1M context length) crazy things like “please create a sprint program for new feature xyz” and get really good high quality answers. Even crazier I can feed those sprints to Claude Code (CI/CD tests all running) and it will do a very good job of implementing. My other option is I can farm those sprints out to human dev resources I have at hand and then spend 90% of my time “thinking, hand holding and talking about code and solutions” and working with other devs to get code in prod.
Imo this is a false victory, emphasis should be on shipping. Although each domain / pipeline / field needs and prioritises different things and rightfully so. AI lets me ship so much faster and for me that means $$$.
I think I am a realist and your last point about “engineering” - is a contradiction. Maybe try better tools? Lastly:
“While the problem of ai can be viewed as, “Which of all the things humans do can machines also do?,” I would prefer to ask the question in another form: “Of all of life’s burdens, which are those machines can relieve, or significantly ease, for us?”
Richard Hamming, pg.43 The Art of Doing Science and Engineering: Learning to Learn
How often you grok a new code base per year? If that's the core of your work, then yes - you benefit from ai much more than some other engineers.
Every situation is unique for sure.
> I would class myself a mid to high skill dev
It's not about your skill level, rather about the nature of your job (working on a single product, outsourcing company with time framed projects, r&d etc.)
Are you able to qualify a “2-3x” improvement? That’s a honest question. The anecdotes out there are wildly all over the place, and don’t match up with my own experience or that of my peers. I’ve only seen a marginal uplift, which includes productivity offsets caused by mistakes and hallucinations, not only for my own work, but from LLM assisted output from coworkers.
It's like saying "robots are replacing civil engineers". Asphalt laying is about 10%? of the work required in commissioning a road. The deciding whether to build a road at all, the costs, where to build it, the math all need to be done by a civil engineer.
The bulk of Software Engineering is feasibility study, requirements gathering, detailed design (architecture) then finally the implementation phase where AI comes in.
Those stages are in order of importance. Getting it wrong in especially the first two results in a high quality shiny white elephant at best.
The implementation phase is at most 20% but on average 10% of the work required to commission reliable maintainable software.
It makes sense. The business stakeholders always want to get things done ASAP, and they don't really care about how it is done. This is especially true if the stakeholders want to do many one time trials.
I think those stakeholders are the true engine of promoting AI.
okay, but code still has to be written. you can be a master architect and if the codebase requires X lines they have to come from somewhere. i'm just having a hard time grokking how you you can spend 1-10% of your time coding and actually ship anything at speed. esp if you imply you're not far away from the real engineering work.
or maybe at these companies the product is pretty stable or you're in an area where it's more optimizations vs. feature building?
Looks like the higher the management, the farther away from real engineering work — the more excitement there is and the less common sense and real understanding of how developers and llms work.
> Are you 10x more efficient?
90% of my time is spent thinking and talking about the problem and solutions. 10% is spent coding (sometimes 1% with 9% integrating this into existing infrastructure and processes). Even with ideal AGI coding agent id be only 10% more efficient.
Imagine a very bright junior developer. You still are heavily time taxed mentoring him and communicating.
Not many non technical people (to my surprise) get it.
Based on posts and comments here there are plenty “technical enough” people who don’t understand the essence of engineering work (software engineering in particular).
Spitting out barely (yet) working throwaway grade code is an impressive accomplishment for TikTok, but it has very little to do with complex business critical software most real engineers deal with everyday