Sectors
Capabilities
Specialisms
Yesterday I went to Nokia Bell Labs, where the transistor was originally invented in 1947, for the OverdriveAI Summit. Fitting, because what that building represents is the same thing AI needs most right now: sharp environments
The plaque at the entrance hits it best. It quotes Harold D. Arnold, Bell Labs’ first director of research:
“Invention is not to be scheduled or coerced… it follows research through the operation of genius and the best that any department can do to promote it is provide a suitable environment.”
That stuck with me throughout the summit. We’ve got the tools. The infrastructure is being built at a pace we’ve never seen before. A community of early adopters eagerly shares best practices and thoughtful experiments all day, every day on LinkedIn. The real leverage now lies in the spaces we build around teams, customers, and ecosystems to get the most value out of the technological leap forward that comes with AI.
Here are three themes that stood out.
Mike Wilner from OpenAI led one of the most grounded sessions of the day, focused on how startups can differentiate in an era when building is easy and copying is even easier.
“It’s never been easier to build. Now the most important question is, what do we build and for whom?”
Whether you’re a startup founder or a marketer at a scaled company, the pressure is the same: you’re not going to outrun the speed of generative tools. Even pretty good content is a commodity. But you can outfox your competitors in insight, by getting closer to your customers than they are.
Wilner also made a strong case that no one should be in a holding pattern waiting for the next breakthrough in their model of choice:
“If research on new frontier model capabilities stopped today, there would still be ten more years of growth” from startups leveraging the tech to make products that solve their customers’ problems.
The models are ready. The opportunity gap isn’t technical. It’s about focus, specificity, and who knows their audience well enough to build something that matters.
One of the most compelling enterprise use cases came from Mano Mannoochah, Chief Data, Analytics and AI Officer at Verizon. Their team built an internal AI-powered “podcast” for execs—an audio summary of millions of customer service interactions, built to surface issues, patterns, and sentiment in a way humans can absorb.
It’s not flashy, slickly produced content. But it doesn’t need to be! They’re using AI internally to get signal from noise.
For marketers, it’s a good reminder that not every AI play needs to be content generation. Some of the most strategic applications are upstream: spotting real-time friction points, segment shifts, and unmet needs, before they show up in lagging indicators like sales.
A session titled “Why New Jersey just might win the AI race” took a refreshingly regional view of the tech landscape. There’s no end to breathless, “which country will win the AI race?” coverage, but within the U.S., states are now quietly competing too.
Talent, energy costs, data center density, regulatory posture are all variables that will shape where future Fortune 500 companies do business.
Corey Sanders from CoreWeave put it well: when Google acquired DoubleClick in New York, it didn’t just buy an ad tech company. It created a center of gravity. The people who built that success stayed, invested and became the seedbed for the whole Silicon Alley startup ecosystem.
We’re still early enough in the AI timeline for those same gravity wells to form throughout the country.
The transistor didn’t change the world because it was clever. It changed the world because Bell Labs built the kind of place where clever people could focus, experiment, and stick around long enough to make something useful.
That hasn’t changed.
If AI is going to live up to the hype, it won’t be because one more model drops. It’ll be because enough teams – and enough regions – created real gravitational pull that keeps talent close, problems specific, and invention inevitable.
See all