Early-stage founders share the ups and downs of building a company
At New York Tech Week, small founders discussed lessons learned about tech stacks, co-founders, and AI.
• 4 min read
New York’s Tech Week is full of events where execs boast about dizzying growth and AI revolutions.
At a convening of early-stage founders near the top of One World Trade last week, the stated goal was to air a more honest accounting of the trials and troubles of getting a new tech company off the ground.
It was for founders who’ve maybe reached “the point in your founder’s journey where you’re rolled up like a child on the floor, hoping that your company doesn’t succeed,” as Lori Mazor, founder and CEO of AI education company Synthetivity, put it in introducing the session.
True to that aim, the panel began with a question about each founder’s biggest mistakes and lessons learned. Elizabeth McCalley, founder and CEO of a marketing platform for early-stage startups called StartStak, said she wished she’d performed more due diligence of people and tech platforms when starting out.
“We never saw an AI framework we didn’t love,” McCalley said. “We fell in love with all of them—big mistake. Some are not ready, some are still not ready. Define frameworks that work for you.”
These days, StartStak has built its tech stack around the chance that the fast-changing AI landscape can shift at any moment.
“We’re able to switch in and out frameworks in days—not that we want to do that,” McCalley said. “Pretty much the entire market is trying to figure out what the market is, what’s important, what’s staying, and that’s how we designed our architecture, around that first principle.”
Eugina Jordan, CEO and founder of integration platform YOUnifiedAI, said her biggest mistake was picking a co-founder who didn’t work out—“Just like with any marriage, those issues will cost you some paperwork.” And Dan Coates, CEO and co-founder of youth market research firm Ypulse, said his was an acquisition that had some issues under the hood.
“It turned out to be nothing like what we expected,” Coates said. “What we managed to get out of it was a lot of net operating losses, which we’ll use to defer against our taxes in the future, but that’s all we got.”
Every company is built on hard choices.
Founder Brew is our twice-weekly newsletter covering how great ideas and entrepreneurial spirit grow into real businesses. We examine what it takes to build, the tradeoffs founders face, and what keeps them going.
By subscribing, you accept our Terms & Privacy Policy.
Coates said after that acquisition failed to deliver the tech capabilities the company needed, his company now only works with tech providers with a proven track record in the exact capability needed.
“We were inundated by emails and calls from companies that were saying they could help us realize our AI dreams, and we had the same answer every time, which is, ‘Show us something that you’ve done that’s in production that’s like what we’re trying to do,’ and that really winnowed down the field,” Coates said.
Jordan said her startup taps mostly Google products and Gemini AI in addition to ChatGPT. The company originally applied to Microsoft’s startup program and “got some small amount, and then they killed the whole program, and that was not scalable for us.” They then negotiated with Google and eventually became part of Google Cloud’s startup program.
“There’s a tier that Google doesn’t advertise, it’s called Ecosystem Tier, and you can get up to $25,000 if you have a bootstrap startup with that tier in Google Credits,” Jordan said.
All three founders said the way they operate has been reshaped by AI, particularly in the last several months. Ypulse has been using its stockpile of youth-focused market data to build AI integrations and skills for customers, Coates said. McCalley said her company builds videos with AI avatars. Jordan’s team has built different skills—Google Gems, in this case—to manage marketing content for different social channels.
They also expect their business models to transform in the next five years. In Ypulse’s case, for instance, Coates said it could be fully integrated into data marketplaces within the big AI platforms.
“We may not even have a go-to-market function,” Coates said. “You’ll find our data source within a marketplace of data sources within Gemini or Claude or within ChatGPT, and you’ll basically say that you’re making decisions in certain spaces, and then you’ll have the opportunity to plug into our repository and benefit from our insights.”
Every company is built on hard choices.
Founder Brew is our twice-weekly newsletter covering how great ideas and entrepreneurial spirit grow into real businesses. We examine what it takes to build, the tradeoffs founders face, and what keeps them going.
By subscribing, you accept our Terms & Privacy Policy.