“I’d start the retraining now, but not in a panic. I’ve been automating my own work on purpose, and the more I do, the clearer it gets what’s actually mine: taste, judgment, connecting dots. That’s the part to protect.”

“I would retrain now, but not because the demo is real. Most of what I have seen demos do in minutes still needs a human in the loop for taste, judgment, and brand safety when it ships.”

“I would retrain now, but not in a generic learn AI way. The budget slides mean the organization is already deciding what roles look like next, and waiting means letting that decision happen without you.”

“I automated half my own design process once and it backfired: I lost the boredom that actually produced the good ideas. Start retraining now, but retrain toward the part AI can't fake, which is taste and judgment, not more tool speed.”

“I've watched AI demos look like replacements and then fall apart on the real work, but I've also bet my company on the idea that taste and storytelling are what survive when everyone can make things. Start moving your team up the ladder now: from making assets to directing the system, owning the brand story, and deciding what good looks like.”

“I wouldn't wait for the budget to become a conversation. At Anthropic I watched the design process shrink and split: half the work is now jamming with engineers on working code, the other half is pointing the whole team at a three-month vision instead of a two-year deck. The people I want around me are the ones already building with the tools, not the ones waiting to be told the threat is real.”

“I would retrain now, but not because the demo is guaranteed to work. At Noesis we run every service person with their own agent, and the real leverage comes from the operator, not the tool. The people who learn to steer, vet, and redirect these systems become the ones the team cannot replace.”

“I got laid off in 2020 and spent months waiting for the floor to stop shifting under me; the people who moved fastest were the ones already building the next thing before the decision was made for them. Retrain now, but do it in public: learn the agent stack, run a real internal pilot, and make yourself the person who can judge whether the demo is theater or not.”

“I built a Claude-powered PM agent at PlusPlus that replaced a six-figure hire, so I know the budget slide is real. But the half it demos is the half that looks impressive: the agent can't yet judge what should be built, only describe what was asked for. I'd retrain now, not because layoffs are certain, but because the designers who survive are the ones who learn to direct the agent instead of competing with it.”