For the last few months, while networking or professionally socializing in any way, I keep finding myself having the same conversation. It goes like this:
- Even when this method is used appropriately, it’s not meaningfully valuable. I don’t understand why it’s a common practice. Ignore it. As an industry we should stop doing it.
- But we’ve found ways to make it faster and cheaper with “AI” tools and techniques.
- Even if it were immediate and free, I’d continue to ignore it. It’s no longer the right tool for the job.
- Why do you hate AI?
Oh, for fu…
(Deep sigh.)
This is where I change the subject or politely find a way to leave the conversation.
A few times, though, it’s been someone I know. Someone I’ve worked with in the past. Someone I would like to know more, even if we continue to disagree on the topic.
I continue exploring…
And they get dodgy.
(You see where this is going?)
Eventually I get frustrated and ask the blunt question.
(I never see where this is going.)
Do you believe this is a good methodology in this situation?
(As the words leave my mouth I slap my imaginary forehead.)
No.
(Why can’t I see this coming?)
Their answer is always no. They were hoping not to have to say it out loud. Now they’re irritated with me cornering them and making them say it.
It’s never about methodology. It’s never about choosing the right tool for the job. It’s never about the quality of the work.
It’s about demonstrating “AI” aptitude. It’s always about demonstrating “AI” aptitude.
Even when the methodology is inappropriate to the situation, if they are able to claim that a commonly used process is now more efficient with “AI” help, they win. That, more than anything, is what’s being rewarded in nearly every company right now.
I bet I would be rewarded for demonstrating more AI aptitude in my company even if it didn’t make anything better — heck, even if it made things worse! The problem is that I’m not smart enough to capitalize on this dynamic. I keep doing the dumb thing of trying to do a good job.
(Hang on. I know how that sounds.)
I’m not suggesting “AI” can’t help anyone do a good job. I’m saying a dumb process is a dumb process, even when it can be made more efficient with “AI” tools and techniques.
And we, as an industry, do a lot of dumb shit. Like, a looooooot.
(Which industry you may ask? At this point, does it matter? Does it? Does it?!?)
And, without fail, if they’re willing to be candid and continue, the “AI” improvements come at the expense of the quality of the larger body of work. Every single time.
(If you say I hate “AI”, I’m gonna lose it.)
That’s not always a bad thing!
Doing something with 75% of the quality at 25% of the cost might be the right choice. I wouldn’t call this a degradation. It’s a different tool with a different value and cost.
I won’t wager a guess at how much it happens, but many “AI” enablement projects could be 150% the cost and 50% of the value. Who knows? Rarely does anyone check. If we did, I imagine we’d be horrified.
What I worry about is that people don’t appear to realize this is what they’re doing. Or, even worse, they realize it but navigate conversations in a way that their boss won’t be able to spot the con. Make the local “AI” improvement numbers look good. Ignore the global consequences — nobody is going to verify. Then, reap the rewards.
Study systems in the laziest possible way and you will conclude that local improvements rarely aggregate into global improvements. That’s not how systems work. That’s how they fail.
But this is what we’re all doing. This is what we’re all doing all the time
This is not the way.
Even when we’re optimistic about “AI” tools and techniques, we all have to admit that we’re undermining ourselves by incentivizing them this recklessly. We are going to regret it. And I am beyond exhausted talking about it.
≈