A startup that learns how to code has been hailed as a success by the company it was born out of.
But how is a startup that doesn’t actually code going to succeed?
Forsyth, which uses artificial intelligence to teach software to recognize objects, was founded in 2013 by former Google engineer Chris Gaffney.
Now it has raised more than $3 million in venture capital.
“It’s not just a business,” Forsythes co-founder and chief executive, Alex Mott, told Business Insider.
“We’re trying to create a business that is more than just a startup.”
To do that, Forsys founders said, they have to understand how to build a platform that can deliver a product and help other people do the same.
“You can’t just make a business out of it, and you can’t simply build a product on top of it,” Mott said.
“The first thing that people should look at is what are the problems that you’re solving.”
A startup needs to understand where the problems are, Mott told Business News Daily.
And then it needs to figure out how to solve those problems.
Forsyts founders built a prototype to test its technology in a lab.
“In a lab, it was just a bunch of cardboard boxes,” Mett said.
The boxes could hold either a computer with software that the startup uses to train its neural networks, or a device that connects to the cloud.
In the first week of testing, the startup saw a significant reduction in errors, which they attributed to the device.
“There was a drop in error rate by half,” Matt said.
They then turned their attention to building an app for a mobile device, and quickly found out that, while Forsyths platform can recognize objects on a map, it also has to work with the smartphone app.
“One of the things we’ve learned over the past couple of years is that when we’re building apps, it’s all about the user experience,” Motta said.
“[But] it’s not necessarily about the app.
We’re building a platform, we’re not building an interface.
We don’t want the app to do the work, we want the platform to do it.”
To solve this, the Forsysts team has been experimenting with different ways of teaching the system to recognize a variety of objects.
“I think it’s a very interesting approach because it takes a different way of thinking,” Motti said.
Motta says that the team is trying to make the platform even better than the prototype they had when they first built it.
“What we want is a machine that’s capable of learning from data that’s already been built by human beings,” he said.
Forsys uses artificial neural networks to train neural networks that recognize objects.
Image: Forsytheys website Forsytemthes team is now working on a new platform for mobile devices, and Mott is hoping that it will be ready by the end of 2018.
The company hopes to have it ready to take advantage of the next wave of machine learning and artificial intelligence advances.
Mott says that his company wants to build forsyths product into a cloud-based service that can be used by hundreds of millions of people around the world.
Mett says that their main goal is to “do the right thing” and not only forsytemto solve a problem but also to build products that people will actually use.
“Ultimately, we think the future of machine intelligence and artificial learning is going to be in the cloud, in the Internet of Things,” Mitts said.