Radar trends to watch: April 2020
Coronavirus, real time transcription, quantum computing, and regulating cryptography.
Since early in March, technology news has been all Coronavirus, all the time. That’s a trend we expect to continue through April and probably beyond. So let’s start with Coronavirus news, and hope that we have something different for next month.
- The Coronavirus pandemic is forcing reconsideration of how private data is used. Maciej Ceglowski’s post “We need A Massive Surveillance Program” is important, particularly since Maciej has a long history as a privacy advocate. At the same time, many other privacy advocates are saying, “Be careful what you give up, because you won’t get it back,” including Edward Snowden.
- A number of organizations are using blockchains as a way of sharing coronavirus data. I don’t think this will be the blockchain killer app (it’s too specialized), but it might be the killer demo.
- While the maker movement of a decade ago has died back, it’s worth noting that the coronavirus has spawned a lot of maker projects—from facemasks to ventilators, and many things in between.
- 24,000 Coronavirus research papers in one archive: Now the question is how researchers will use this archive effectively. There’s really only one answer: automatic summarization and intelligent search using AI.
- Apple has made biometrics on watches an essential feature. Other companies with smart watch products will be forced to follow–especially since doctors are now replacing in-office visits with telemedicine. A lot of cultural change is needed before doctors will accept ambient data detection, but Coronavirus may force that change to happen.
Operations and Infrastructure
- Rolling updates for Kubernetes with Kublr: Rolling updates are an essential feature for groups that are practicing continuous deployment. There have been some hackish workarounds, but Kublr attempts to provide a real solution.
- AWS has a Linux-based operating system for containers called Bottlerocket. Bottlerocket’s most important feature is that it streamlines the update process, making updating possible for container orchestrators.
- Monitoring production systems is an essential practice. m3 is an open source monitoring tool from Uber that is effective at huge scale. It is being commercialized by Chronosphere.
- Microsoft buys npm: This certainly isn’t Steve Ballmer’s Microsoft. And, along with Microsoft’s acquisition of GitHub, it makes Microsoft a dominant player in much of the open source movement.
- Chrome has new tools to help develop for the visually impaired; they simulate what the page would look like with different vision problems. This is an important step forward for developers working on accessibility. Mozilla also has accessibility checking.
Artificial Intelligence and Machine Learning
- Realtime transcription and translation with Google Translate: This feature is rolling out to Android now, and will be delivered to iOS later. There are lots of issues that they will have to think about—for example, there are significant variations in Spanish from country to country—but it’s an impressive accomplishment for natural language technology.
- We’ve known for some time that AI-based image classification can be tricked. Researchers have shown that it is also possible to spoof LIDAR, which could have a big effect on the development of autonomous vehicles.
- A data set is a world view. This isn’t a new idea, but it’s important. A must-read article.
- Facebook has developed a new system called Deep Entity Classification for detecting fake accounts. It’s based on connection patterns between users and also seems to take advantage of machine-generated labeling.
- TensorFlow Quantum integrates quantum computing into TensorFlow to jump-start research into machine learning on quantum computers. While TensorFlow does not directly support quantum computing, this makes it a tool for simulations and prepares the way for supporting real quantum computers.
- Honeywell hasn’t been part of the quantum computing picture so far, but at the beginning of March, it suddenly announced that it had built a quantum computer. They’re claiming it will be twice as powerful as IBM’s machine.
- There is legislation in the US Senate that would have the effect of restricting encryption. While this is framed as a bill to combat child sexual abuse, it would have drastic effects on computer security of all kinds.