Golioth is (Not) Pivoting to AI

If you regularly read our blog, or follow any of the Golioth social media accounts, you likely saw our “Golioth for AI” announcement post and video this week. We have been building towards this launch for quite some time, but astute readers may have noticed that there were actually very few new features announced that were exclusively dedicated to artificial intelligence or machine learning. This was by design.

Depending on where you sit, the recent explosion in the use of AI across industries has been either the largest technological shift since the internet, or the biggest hype cycle since the dot com boom. We prefer to think of it as something in the middle: a new capability already demonstrating usefulness for a subset of tasks, while also exhibiting severe short-comings in others. One thing we can all agree on is that the pace of innovation is rapid, with new models being developed, new hardware being built, and new companies being started. Choosing where to invest is complicated in the face of an uncertain future.

That’s why over the last year as we have rethought the architecture of the Golioth platform, we have been thoughtful in our approach to building modular, composable product features. This has been most apparent in our launch of Golioth Pipelines, which has allowed us to introduce new functionality every week, while enabling our customers to incrementally integrate those features into their own architecture. We believe that flexibility is your most valuable asset in the face of uncertainty. This is the same ethos we are bringing to enabling our customers to leverage AI.

Rather than wrap the API of one of the many AI platforms, as so many others are doing today, we have chosen to expose those platforms to you directly, ensuring that building on Golioth does not restrict your ability to access the latest and greatest the market has to offer. Rather than building a proprietary model training service, we have invested in open source solutions and partnered with leaders like Edge Impulse.

We’re not pivoting to be an AI company because providing access to the best tool for the job has been our mission all along. Investing in leveraging AI one way today should not preclude you from investing in another way in the future. With Golioth, you can be sure it won’t.

Dan Mangum
Dan Mangum
Dan is an experienced engineering leader, having built products and teams at both large companies and small startups. He has a history of leadership in open source communities, and has worked across many layers of the technical stack, giving him unique insight into the constraints faced by Golioth’s customers and the requirements of a platform that enables their success.

Post Comments

No comments yet! Start the discussion at forum.golioth.io

More from this author

Related posts

spot_img

Latest posts

OTA: Working with Multiple Images

OTA usually makes people think of updating firmware. But OTA also covers many of other applications, like distributing new machine learning models, UI image assets, or firmware updates for other "downstream" processors. Golioth has the flexibility to support all these use cases. Here's how to use the power of the Golioth OTA service when you need to deploy more than just one binary.

Using the NXP FRDM-RW612 with Golioth

The new NXP FRDM-RW612 board comes with built-in Ethernet, and a tri-band radio. We were able to get it up and running all of the Golioth Firmware SDK sample applications in just a few hours. Here's what goes into the process.

A Sneak Peek at the Bluetooth-to-Cloud Early Access

Golioth's Bluetooth-to-Cloud is in private access currently, but this post lets you peer behind the curtain to see how it works on some development boards.

Want to stay up to date with the latest news?

Subscribe to our newsletter and get updates every 2 weeks. Follow the latest blogs and industry trends.