The Resistance to Using Any AI...

When folks say "I can't find a use for AI", I think far too many of them are overthinking the use-cases, or expecting a much more grand difference in their lives.

Without actually relying on LLMs to do the thinking for me, I can say without a doubt that AI has fundamentally changed the way that I work over the past couple of years. No, it hasn't made me 10x more productive or anything like that, but what it has done is GREATLY reduced the number of mistakes that slip through, and added an unbelievable amount of quality of life tools to my belt. The mistake reduction alone is probably the biggest benefit, and why I just can't see myself really going back.

To give a few examples:

• Almost every major AI player has some form of "Deep Research"; the short version is that it goes Google deep-diving for you to pull a source-cited response to your query. If you've ever been stuck on something, or needed confirmation on a line of thinking? This is huge. Rather than spending 30+ minutes Googling, I can now find the source I need within 5 minutes or less. It has rarely failed me at this.
• When you're working alone on something, mistakes are far more prone to happen; this is especially true for developers. When you're on a bigger team, you have some ability to offset this by getting someone else to put a second set of eyes on it. But the reality is that isn't always possible- sometimes the team is busy, or you are tackling an issue alone, or they just don't have the domain knowledge to assist. Is AI perfect/does it catch everything? Absolutely not. People don't either, though. At least it's more than JUST you.
• It's a veritable toolbelt of quality of life features. It used to be that if I hit a post online that was in a language I didn't know/understand, and if the browser didn't offer a translate feature for that page, I'd have to Google Translate the text manually. Now I can just quickly screenshot the page, drop it into a chatbot, and have it tell me. Is it perfect? Of course not. But I don't need perfect if I'm just trying to understand what it's saying. Similarly for sanity checking a large corpus of text for a fact- "here's what the conclusion I'm trying to verify, here's the text, help me find the information relevant to it within this text." kind of thing.

There are other benefits I get as a developer beyond these. But I think that part of my hang-up when I see folks completely write off AI is that just the minor benefits of having it on-hand pay off in spades for me.

Is it easy to quantify that value for a corporate cost benefit analysis? No. But I think that's a problem corporate leaders need to learn to deal with in their own way. You know that it has value, and if you aren't able to estimate exactly what that value is up front, then you really need to decide if you're willing to let that roadblock you from such value, or if you're going to go out on a limb, let folks try it, and calculate the return after the fact.

Though, on the last note- if you do that, please train your folks on it. There is a particular way to handle AI that makes it more valuable, and if you aren't using it that way then it could become more of a hindrance than a help.