AI Advice Must Come From the Trenches, Not the Sidelines

There's a common saying that I dislike: "Those who can't do, teach." If teaching isn't informed by practice, it tends to drift into the theoretical. Especially in fast-moving fields such as AI, with numerous pitfalls and nuances, only those who also practice are qualified to teach and advise.

How else can you help a company pick the correct path to pursue with their AI initiative if you haven't walked that path before? In slower-moving fields, you can rely on a wealth of accumulated knowledge, slowly absorb it, and then disseminate it: You don't have to lay brick or cut wood to be a successful architect; you can learn the relevant material properties from a textbook. After all, wood, stone, and concrete, among others, don't undergo dramatic progress every few months.

But with your AI projects and initiatives, you'll want to get advice directly from those who are building and doing, not just theorizing. This, of course, goes against the common industrial-scale consulting model: A hyper-effective sales team creates FOMO (fear of missing out) and/or promises untold riches if only you buy their services, then hands you off to their engineers who're now tasked with doing the impossible (or delivering something over-built). Or, a seasoned strategy consultant draws up all sorts of ares where your company would be guaranteed to benefit from AI, but their suggestions are completely unburdened by actual feasibility because that person has never gotten their hands dirty.

So when you're shopping around for advice on AI, seek out the nerds and hackers and doers. You might not get as polished a PowerPoint deck, but you might just save yourself a lot of headache and wasted effort.