Take Care of What You’ve Got

A sixth-generation Oregonian’s case for using AI wisely, and why good stewardship and good business are the same move.

My family has been in the Northwest for six generations. We came out on the Oregon Trail and homesteaded in Douglas County, in the southern part of the state. One of my relatives served as mayor of a small Oregon town, back when the hardest fights in local government were about water rights, who got how much of a river that did not care about anyone’s plans for it.

You learn something growing up in a family like that. Not a political something. A practical one. You take care of what you have got. The soil, the water, the equipment, the people next to you. Not because the world is ending, but because these things are finite and it is foolish to waste them. You leave the place a little better than you found it, because someday it belongs to your kids, and their kids after that.

I have spent my career a long way from that farm, running large technical systems and now building an AI company. But that ethic never left me, and lately it keeps showing up in the one place I did not expect it: our power bill.

The intelligence we are racing to deploy runs on my river

Here is something most people do not think about when they use AI. It runs on real, physical things. Electricity. Cooling. Water. And a lot of it runs right here, in my backyard. The data centers along the Columbia River are powered in no small part by the same hydro and Gorge wind my family has lived alongside for a hundred and fifty years. The intelligence the whole world is racing to deploy runs, in part, on my river.

I am not here to tell you the sky is falling. I will be honest, I am skeptical of a lot of the doom framing around this. But I do not need a catastrophe to know a simple thing: when a resource is finite, wasting it is not clever. Using it wisely is.

And here is the part that surprised me. In AI, using the resource wisely is not a sacrifice. It is the smart business move too.

Most companies are burning premium compute on ordinary work

The dirty secret of enterprise AI right now is how much of it is waste. Teams point their most powerful, most expensive models at every task, from drafting a summary to reformatting a list, the same way you would fire up a diesel generator to charge a phone. It works. It also costs a fortune and runs slower than it needs to, and most of that horsepower is spent on work a smaller, right-sized model would have handled just as well.

That is not being cutting-edge. That is leaving the truck idling in the driveway.

The wiser approach is the one any good steward would recognize. Match the tool to the job. Use the powerful model when the work actually calls for it, and a lean, efficient one when it does not. You get the same result, for a fraction of the cost, and usually faster.

This is why we built Swa the way we did

Swa gives your team every major AI model in one place, and it is built to make the efficient choice the easy one. Efficient models run as the sensible default. Your admins can set the guardrails. And you can see exactly where the spend is going, instead of finding out at the end of the month.

We do not just preach this. We run on it. Over one ninety-day stretch, our own team of twenty-six people ran more than a hundred million tokens of real work through Swa. About seventy percent of that work ran on efficient, right-sized models. The result was roughly half the compute cost of an all-premium approach, for the same output. Same work, done well, at half the bill.

Good for your bottom line, good for the world we share

This is the part I love, because it does not require anyone to choose. Lower AI costs go straight to your bottom line. They free up budget to put AI where it actually makes you more competitive, instead of burning it on busywork. Efficiency does not mean using less AI. It means being able to afford more of it, in the places that matter.

And that same wise use is easier on the finite things underneath it all. The power, the water, the backyard we are handing to the next generation. You do not have to believe the world is ending to see that this is just common sense. It is good for your business, good for your people, and good for the world we all share. Those are not in tension. They are the same move.

Take care of what you have got. It turns out to be good advice for a farm, for a river, and for how you run AI.

Ben Pickett, Co-Founder and COO of Swa