It’s a skill just like any other. Without practice people will absolutely waste time. But knowing which problems are appropriate for an LLM can save days of labor.
As an engineer in the field for 30 years, this tech beats every man page, sdk, google search, stack overflow post or head banging situation by light years.
The rebuke of it is surprisingly antithetical to the progress of the industry. It reminds me of the decade+ of those who refused to use an IDE and favored notepad[++] or vanilla vim instead. I suspect in 5 more years those similarly-minded folks will finally adapt in the same way we’ve all grown accustomed to IDEs today. Or those who adapt will just outpace those who don’t.
The rebuking is largely focused on using AI-written code (and other content) in ignorant ways, not on using AI to learn or to get example code upon which to base your own, or for simple objective code you can confidently audit, which are in fact more reliably good and less depressing/degrading/market-damaging use cases.
As an engineer in the field for 30 years, this tech beats every man page, sdk, google search, stack overflow post or head banging situation by light years.
The rebuke of it is surprisingly antithetical to the progress of the industry. It reminds me of the decade+ of those who refused to use an IDE and favored notepad[++] or vanilla vim instead. I suspect in 5 more years those similarly-minded folks will finally adapt in the same way we’ve all grown accustomed to IDEs today. Or those who adapt will just outpace those who don’t.