AWS re:Invent 2020 — What got my attention the most?

Lior Alperovitch
2 min readDec 16, 2020
A screenshot from the announcement of Monitron during AWS machine learning keynote

Last week during their annual event, AWS introduced two new AI devices that can help streamline the work required to apply machine learning (ML) with IoT devices:

AWS Panorama — To process images and videos from existing IP cameras.

Amazon Monitron — To measure vibration and temperature.

AWS takes a step further with its IoT offerings and provides both hardware and the relevant infrastructure combined!

This isn’t a typical move for pure cloud companies such as AWS, which usually focus on remote services rather than on-premises hardware devices.

It looks to me like AWS understands that it’s difficult for startups and more prominent companies to build not only the ML models, but also appropriate software and hardware that will run on the edge.

Will the new offerings be a breakthrough, similar to the Amazon Echo for consumers? It seems that AWS would like to have a foot in the door and shift industrial enterprises to start using its public cloud offering by providing end to end solutions regardless of their existing legacy systems.

A screenshot from the announcement of AWS Panorama during AWS machine learning keynote

However, processing ML algorithms on the edge, as provided in AWS Panorama, is beneficial because of several constraints:

Real-time responses — Sometimes, we can’t afford the roundtrip to the cloud.

Protecting sensitive information — Data from highly regulated devices, such as medical devices deployed at hospitals or controllers for autonomous vehicles. They both capture and process highly personal, life-critical information.

Work without an internet connection — We can’t afford to stop working when the internet is down.

Those constraints are applicable, especially for domains such as safety and security, logistics, healthcare, agtech, automotive, etc.

It would be interesting to see if companies will use these services to build unique and innovative applications based on camera and sensor data. If so, will we start to see an ecosystem of edge appliances powered by other cloud companies such as Google and Microsoft?

If you want to watch the full AWS re:Invent2020 ML keynote the link is here: https://virtual.awsevents.com/media/t/1_07cg4srl/186983933 (registration is required) the video will start after the sponsor's advertisement.

--

--

Lior Alperovitch

product person | tech enthusiast | passionate about creating products that users love