Indian service companies can train some of their intake with COBOL, some obscure printer programming language, Clojure etc and give them anxiety about getting into a career dead end.
15 years back I got an affiliate marketing ebook, and it had lessons on getting hundreds of blogspot subdomains and using a desktop tool to publish the same crap about, say, tropical fish to all of them. There was/is also a captcha solving service which got a nice landing page about employing people.
Now that code is cheap, I ensured my side project has unit/integration tests (will enforce 100% coverage), Playwright tests, static typing (its in Python), scripts for all tasks. Will learn mutation testing too (yes, its overkill). Now my agent works upto 1 hour in loops and emits concise code I dont have to edit much.
Totally get it, and I think we’re describing the same control loop from different angles.
Where I differ slightly is: “100% coverage” can turn into productivity theatre. It’s a metric that’s easy to optimize while missing the thing you actually care about: do we have machine-checkable invariants at the points where drift is expensive?
The harness that’s paid off for me (on a live payments system) is:
- thin vertical slice first (end-to-end runnable, even if ugly)
- tests at the seams (payments, emails, ticket verification / idempotency)
- state-machine semantics where concurrency/ordering matters
- unit tests as supporting beams, not wallpaper
Then refactors become routine, because the tests will make breakage explicit.
So yes: “code is cheap” -> increase verification. Just careful not to replace engineering judgement with an easily gamed proxy.
I'm writing my side project as if I can afford only a 70B model in 2028, even if I have VC-subsidised unlimited GLM-5 now. I'm trimming away most of the generated code and generating more tests.
Would prefer if 2028 models are concise and generates perfect refactors.
I have semi unlimited budget forAI tools. And doing it purely with Claude code or similar has been in effective.
The architecture is too simple or unmaintainable rubbish.
One can say that I don't need to care about that, but after few weeks of work, every small chance needs to read everything which quickly gets expensive to do anything.
Using human code and doing similar tasks is incredibly cheaper and more effective.
I run few simultaneous green field projects and I can evaluate basically any tool I want. And I will say that quality wise, all ones I tried are disappointing.
Solar panels are tougher than cutlery: "Most monocrystalline and polycrystalline solar panels are rated to withstand 25 millimeter (0.98 inches) diameter hail falling at 50 miles per hour"
reply