nikhil.io

Assorted Deep Thoughts™ on LLMs

To use them effectively as a programmer,

  1. You still need to know what to tell it1.
    Be explicit. Especially with success/failure conditions. Especially when laying out a brand new project or module.

  2. You need to check all the work.
    “Trust But Verify.” Did you solve the right problem? Are your tests testing the right thing? Are the docstrings and comments actually helpful (and correct)? Is the code generated maintainable apropos feature integration and bugfixes?

  3. Vibe Coding is bullshit for anything other than proofs of concept.
    If your app goes big, thank you for ensuring others’ job security.

With all three, Wisdom and Experience will still matter. Both generalists and specialists will still matter. Replacement fears are mostly unfounded and driven by greedy and/or uninformed people. Don’t listen to them. Hone your craft daily, be as good as possible at it, keep your chin up, and you will be fine.

I am very sorry if you are an entry-level programmer. I still think you will be okay.

The Future

The free-use, ad-free gravy train will end soon, especially with OpenAI and Anthropic. You will pay by the token or at least be inundated by frustrating and privacy-looting advertizements. You will understand the true cost of inference and learn the depressing reason behind how non-Google companies continued to attract investment (Radix Malorum est Cupiditas.)

The hot gas that the stock market currently sits on will be released soon enough and it won’t be pretty for all except the wealthy. All will finally see how little we’ve grown as an economy across all sectors other than this one.

Any AI that is both unobtrusive and effective 99% of the time (as endorsed by the user) is what will endure. Think of OCR when you hit spacebar on a Mac, but at a very advanced level. Design is hard and people don’t like it when you shove your shiny new toy into their faces (especially with the splendid job you’ve done with the PR around it.)

On-device LLMs that are ‘good enough’ and take you 80% of the way there are the future.

  1. Man asks old mechanic to fix his car’s engine that’s making awful rattling noises. Mechanic pokes around and, after 10 minutes, hits the engine with a giant hammer. Problem solved. Mechanic asks for $1,000. Aghast, man asks for itemized receipt. Sure, says mechanic. “Hitting engine: $10. Knowing where to hit engine: $990.”

    Know where to hit the engine.↩︎