5 Stories Our Readers Read Most
2022 was another extraordinary year. Within 12 months, we saw a new European war explode, crypto markets implode, AI become more chatty and real, data management concerns re-emerge with worries about cloud under-delivering.
These were some of the topics that we covered. But what were you, our readers, most interested in? For that, we went back to the future and uncovered the five stories you read the most. Below are these in no particular order.
AI hardware became a hot-button item
Paul Mah, our DSAITrends editor, looks at the modest rise of free and open source instruction set architecture.
The article examines the UC Berkeley-designed RISC-V and questions its suitability for powering AI models. Mah also studies the growing importance of GPUs, originally meant for graphics and gaming, as they become a mainstay of the modern AI infrastructure.
You can re-read the story here.
The tension behind TensorFlow wars
Mah’s earlier article on TensorFlow and its rival PyTorch examines why AI developers choose either framework to create their deep learning models.
The article dives deeper into the TensorFlow vs. PyTorch battle to uncover outdated and incomplete information. It also finds the struggle is really about industry users vs. researchers, with TensorFlow preferred by the former.
The article admits that there are no clear winners and that both frameworks are valuable. However, it also points out thatTensorFlow offers a more robust framework, while PyTorch is popular for creating state-of-the-art (SOTA) models. While the battle is still raging, Mah feels that TensorFlow offers AI newbies a chance at a career change, while PyTorch will continue to have solid traction in the research arena.
Follow the story to its conclusion here.
When humans become IoT devices
In this forward-looking article, Lachlan Colquhoun, our ANZ correspondent and NextGen Connectivity editor, feels we often overlook the human factor in IoT.
The article demystifies the notion that IoT devices are mindless sensors emitting data that IoT platforms capture and analyze. In fact, after COVID-19, location-based tracking is now making humans a “thing” in the IoT verbiage.
Some examples are indoor location marketing following improvements in BLE tags and smartphones, IoT beacons for workforce visibility, and contact tracing IoT trackers. Colquhoun ends with an optimistic role of IoT in fighting future pandemics and epidemics — natural or manmade.
Track the conclusions of the story here.
The wrong FinOps questions
This story that I penned came out of a recent interview on FinOps. It ponders whether we are framing our FinOps questions incorrectly.
The article argues that FinOps should be a strategic tool to improve cloud resilience and planning. Instead, we use it as a tactical tool in reaction to the sticker shock faced by rising cloud costs. More importantly, it can be an excellent tool to show how much it will cost you to pivot or cloud-ify your infrastructure, which is valuable as we head toward lean times.
We also look at the role of Technology Business Management (TBM) frameworks and why venture capitalists and investors now demand FinOps be embedded in all startup operations.
Optimize your FinOps thinking by reading the article here.
FOMO is not what you think it is
No, we are not talking about fear of missing out here. Instead, Mah points out it can also mean Faster Objects, More Objects.
FOMO, in its alter-meaning, is a new machine learning technique for real-time object detection, tracking and counting on low-powered microcontrollers. It is 30 times faster than conventional techniques — or so its founding creators claim.
Mah looks at how TinyML underpins the FOMO design, allowing machine learning models to run on power-constrained devices. The article also considers it an ideal middle ground for computer vision but highlights its various limitations.
Find out more about FOMO limitations here.
2023 will be reading different
The stories are only a glimpse of what our readers are reading and their main concerns in 2022. And as we enter 2023 and face new challenges and opportunities, the topics of interest are shifting.
More on these next year!
Winston Thomas is the editor-in-chief of CDOTrends and DigitalWorkforceTrends. He’s a singularity believer, a blockchain enthusiast, and believes we already live in a metaverse. You can reach him at [email protected].
Image credit: iStockphoto/Prostock-Studio
