How Layer Breaks Down the Complexity of AI to Cut Through the Noise

AI is an increasingly crowded industry. Despite an economic slump in venture capital, funding for AI startups is soaring, and everyone wants a piece of the AI pie. 

One startup rising above the noise is Layer, a new Columbus-based startup helping companies build the next generation of AI copilots. 

Demand for copilots has increased exponentially in just the last few months, and perhaps you’ve used a primitive version of one before. If you’ve used an online customer service chat that was able to control your screen and take action for you, that’s the basic idea behind a copilot. In other terms, it’s an AI-powered personal assistant that can intelligently take action on behalf of its users and is operable using natural language. 

As technology shifts toward natural language interaction with computers (as opposed to dashboards), copilots like this are becoming more and more popular. Many large tech companies like Stripe and Shopify are pouring resources into building this kind of technology, which signals to other tech companies that natural-language software interaction is the direction in which to be going. 

When Jonah Katz and Andrew Hamilton, two best friends from Cleveland, were going through the Entrepreneurs Roundtable Accelerator in New York City, they met another founder trying to build a copilot at the time. This was when they realized there was no infrastructure in place to help him build what he wanted. Jonah and Andrew saw an opportunity: to make building this in-demand tool a whole lot easier for developers. 

“In the long run in AI, the best developer experience is what is going to win out,” says Jonah.

In 2023, Jonah and Andrew launched Layer and emerged out of stealth mode with a $3 million funding round led by Drive Capital — who also brought Jonah and Andrew back to their home state of Ohio. Now, they’re on a mission to build the most intuitive experience for developers looking to integrate copilots into their platforms. 

Breaking down the layers for simplicity

AI is a complicated industry. For one, there are many types of AI (machine learning, computer vision, generative AI, natural language processing, and more). For copilots, what becomes difficult is reliability. 

Large Language Models (LLMs) often “hallucinate” and produce incorrect information. Getting wrong information can happen with a simple Google search, but what’s more serious with copilots, is that the tool is trying to execute a task on your platform, often critical actions like updating a database or completing a financial transaction. 

“At Layer, we try to fuse the semantic understanding that an LLM has with this kind of rules-based approach, where we constrain what the LLM is actually able to do to give what we call a perceptibly reliable result to any user,” says co-founder/CTO Andrew Hamilton.  

For Jonah and Andrew, it’s all about building a reliable infrastructure that is simple for developers to use and produces high-quality results. 

“There’s a lot of impractical technology out there,” says Andrew. “We’re trying to make it practical. We also try to stay fairly interoperable. As new technology comes out all the time, we want to make sure that a developer who chooses to use our platform can actually use it.”

It’s exactly this kind of simplicity that drives the culture at Layer, too. With a flat hierarchy and a meritocratic mindset, everyone has an opportunity to make an impact. The development team does a lot of pair programming and works very hands-on with the technology. Andrew, as the CTO, is working every day writing code. 

“Here, we can distance ourselves from the noise and focus on building truly useful technology. ”

For new hires joining the team (there are a few who recently joined and Layer will be ramping up hiring again in 2024), it’s a huge opportunity to be on the ground floor of a company developing an innovative and advanced framework that enables companies to leverage the power of LLMs in their platforms. 

Layer is still in the pilot phase with a few select customers, and as they move forward in 2024, they’re looking to improve their self-service model, where any developer at any website can go to Layer and get their own copilot up and running in a few hours. In the meantime, the team is hyper focused on ensuring success for the select customers in the pilot phase. They work one-on-one with their development teams to make sure everything runs smoothly. 

As Jonah and Andrew look ahead, they’re both excited to be building the company out of Columbus. 

“There’s so much noise in the AI space, and it’s really easy to see that in New York. It’s hard to decipher which companies are actually creating value,” says Jonah. “Here, we can distance ourselves from the noise and focus on building truly useful technology. ”

For more information about Layer and to be the first to know about new job opportunities with them, follow Jonah Katz on LinkedIn to stay in the know, or join their talent network.