Tune out the AI noise and get stuff done

Whenever a new model or tool comes out, everyone jumps on it. Within a day your feed is full of people trying it, ranking it and arguing about it. Then come the scary versions. It broke out of its sandbox. It could hack you, or hack the world. AI is going to kill us all.

The hype side is just as loud. You can run a fleet of agents. You can have an agent running an agent that's running twenty more. Most of that doesn't do what it's supposed to do, and I'd say 90% of people don't need it anyway. We run Galang AI as an AI native company, and most of us use one or two agents each.

Even people who live deep in this stuff will tell you the same thing. Matt Shumer, who says he runs hundreds of agents a day, put it this way in September: "these insane setups are fun and all, but they don't actually make you more productive. You just end up spending more time building the system than actually doing real work."^[1]

We have a customer who's very AI forward. He keeps up with AI news and he's on X, which is where I think most AI news breaks first. When we visited his office, he was completely set up. He had his own Mac mini system on its own Tailscale network, running pretty much every agent out there: OpenClaw, Hermes, Grok Bot, Meta's Muse.

When we did our audit, we asked him what problem all of it was solving for his business.

He said it wasn't. He could do more things himself, but he couldn't get his employees to adopt the new systems, or even want to use them. So the time he saved by using agents went right back into managing his new AI stack.

I don't count that as a failure. He got the ball rolling, and he knows more about what these tools can do than most people. But you have to find a balance. One person getting faster doesn't move a business very far when the rest of the team is still working the old way and that one person is spending his extra time keeping the setup running.

There's another trap, and this one shows up when AI does work. People get so amazed by what they're able to do with it that they get carried away with everything it can do. You don't need to do every single thing AI can do. You just need to do enough that it moves your business forward.

A BCG team writing in Harvard Business Review in March surveyed 1,488 full time workers in the US. People who went from using one AI tool at a time to two said they got a lot more productive. A third tool still helped, but less. "After three tools, though, productivity scores dipped."^[2]

When something works, let it run for a while before you go looking for the next thing to add.

I'm not saying don't try new things when they come out. Try them. Be open-minded about it.

Learn your stuff, too. Spend real time with the tools so you know what they're good at and where they fall down. You can't judge whether something will help your business if you've only read about it.

Then focus on the work. Before you add anything to how your business runs, ask yourself a few questions:

1. What is this actually going to accomplish?
2. What gap does it fill in the processes and procedures we already have?
3. Is it going to save us time, and where will that time go?
4. Will the rest of the team actually use it?

If you can't answer those, it can wait. RAND found the same thing when it interviewed data scientists and engineers about why AI projects fail. The most common root cause they named was people misunderstanding or miscommunicating "what problem needs to be solved using AI." They also called chasing "the latest and greatest advances in AI for their own sake" one of the most frequent paths to failure.^[3]

On that last question, pick things that fit how your team already works, and use them yourself where your team can see it. That's how you avoid ending up like our customer, with a great setup that only one person uses.

This is how we try to run Galang AI, too. We're doers. We work fast, and we focus on AI that actually gets the ball moving for a business. That doesn't mean we cut corners. We don't give up quality or security to move faster, and someone on our team checks the work before it goes out.

The people we work with have noticed. Dr. Mark Klein, co-founder and president of ColoWatch, said the speed at which we finished two very different kinds of projects "bordered on amazing, especially since the quality of the work matched that speed." Christopher Gapasin, founder of Mighty Handy and Crafty, said we "cut through the AI hype and showed me what actually works."

When Dr. Sandeep Pawar joined our board of advisors, he sent us a note that sums it up. He said people think the speed in this new age of AI comes from the AI itself, but it's also "speed with which people now need to execute on all things related to their business." Then he added: "I love that folks at Galang are so efficient (but not reckless) in all your dealings."

AI is a really powerful tool. Don't let anyone tell you it's anything more than that, or anything less. Learn it, and then use it to get your work done.

If you want help figuring out where AI fits in your business, book a free 15-minute intro with us.