[guywithahat]

> CEOs keep taking (mostly misguided) about how GenAI will replace their people. The thing they miss, and this highlights, is that customers will also expect to pay far less for GenAI produced workloads, which likely more than eliminates any cost savings

Is this bad? This seems good. The end goal being once it's cheaper, you can do more revenue. Also while I obviously don't stand by this Deloitte scandal when done right GenAI work is better

[AlotOfReading]

"GenAI done right" comes across like the famously elusive "agile done properly" and "safe C". If everyone is holding it wrong, are they?

[guywithahat]

Most implementations of GenAI are done correctly from what I’ve seen. A huge amount of production code is written by AI, and I use it every day because it’s fast and significantly improves my productivity.

[AlotOfReading]

Fast and productive is very different than good. I'd probably agree with a description of AI code as the fast food of code, but I certainly wouldn't call Taco Bell "better" if the wedding caterer showed up with it.

[guywithahat]

I think you're looking for lexical gaps where there are none. If you're not using LLM's while coding, you're behind the industry. While GenAI may be overhyped, it's not a fad, and if you don't incorporate it into your (what I assume is a) software based job you're going to fall behind your coworkers. I can't list every positive adjective here but it's an incredibly useful tool which everyone should be using.

[AlotOfReading]

I'm using them, but frankly only for things where my quality bar is lower. I've seen little evidence that anyone is routinely getting work that's comparable to competent humans, excepting incredibly recent work like Joshua Rodgers' [0]. And I'll note that even Joshua's work was mainly successful in identifying issues, but actually fixing them involved much more human intervention.

[0] https://joshua.hu/llm-engineer-review-sast-security-ai-tools...