A little while ago, I saw someone post a Tweet that said something to the effect of, "In our haste to adopt AI, we created the urgency without the infrastructure." Naturally, the rest of the post was a tech bro showcase of how they "solved it" with standardized agents or whatever else, but it was the idea of urgency without infrastructure that I found to be a crucial observation.
This is especially pertinent given that I feel this viscerally now at work. The pressure is mounting, the deadlines are shorter, and the expectations are higher.
But is anything actually improving?
Speed is every business's favorite vector to try to increase. Gotta get more numbers earlier. Gotta push twice the code in a day. Gotta produce more images than will ever be used in a few hours. Speed is also the most dangerous—and despite the old entrepreneur's mantra of "greater risk, greater reward," I don't think that tracks onto speed as often as we're told to think. Speed should be applied at appropriate times, and not often (in my opinion) so as to reduce the risk of crashing in all possible ways.
Urgency puts pressure on people to apply more speed, and that always entails cutting corners—whether it's not fully finishing pieces or not being as careful as usual. Since AI is our massive multiplier in all systems right now—as well as the pressure to adopt, implement, and use AI—many of us are faced with this conundrum of urgency without infrastructure, so we lean on AI to fill in the gaps. AI increases expectations, AI is used in an attempt to fulfill them.
What then, is infrastructure? Guardrails, red tape, processes, strategy, methodology, design. It's not just setting up AI-agent skills and prompt libraries. It's doubling down on the emphasis of quality. It's insisting on breaks, on regulations, on reviews, on thoughtful action.
"Haste makes waste," my wife always says when she notices that she's rushing. This is an adage she learned while working at McDonald's, the business that is the classic case study in process and strategy that enables a huge volume of people to be served in very short periods of time—consistently. I think this applies to even more situations now, not just employee productivity or efficient processes. The public calls it "AI slop" when we see pointless, incorrect, deceptive, or ridiculous content produced by AI—waste indeed—created by the haste to pump out more content for social media algorithms (AI content for AI judgment and distribution). It is only thoughtful, intentional intervention that can slow things down.
I'm exhausted this week since there was no week-end for me at work, so it's a very short article today. The only thing keeping me going is the excitement of the new album from Evanescence in the next few days ;)
But I really think this is critical to think about. If you're able to advocate for more strategy do it. If you're able to draw boundaries about how much work you can ethically do even with AI, you should do it now—with relentless advocacy for yourself and your teammates—because slowing down is not what society is choosing for us right now, and that's a dangerous place to be.