Show me a Gary Marcus essay, I’ll show you a few new LLM “gotchas” that will be fixed by the next version. Season to taste with self-assured confidence that all these tech goobers really don’t understand how totally overrated AI progress is.
So it has been for 10+ years, so it will be at least 5 more.
It's hard to reconcile how 59% of devs in their survey are "confident" AI is improving their code quality, with prior empirical research that shows a surge in added & copy/pasted lines w/ a corresponding drop in moved (refactored) lines https://www.gitclear.com/ai_assistant_code_quality_2025_rese...
My experience (using a mix of Copilot & Cursor through every day) is that AI has become very capable of solving problems of low-to-intermediate complexity. But it requires extreme discipline to vet the code afterward for the FUD and unnecessary artifacts that sneak in alongside the "essential" code. These extra artifacts/FUD are to my mind the core of what will make AI-generated code more difficult to maintain than human-authored code in the long-term.
Wish more people would report their placebo experiments. I have periodically run them on myself, and am consistently surprised how I have been unable to differentiate substances I thought were helpful (adderall, kratom) end up indistinguishable from placebo over 20+ trials. I guess that my main takeaway was that it is hard to pinpoint when subtle drugs work. My second takeaway was that the data I generated would prob be useful to share, but with no examples anyone cared I opted against. This story inspires me to potentially revisit some of my past placebo tests and show my data.
So it has been for 10+ years, so it will be at least 5 more.