Commentary: While you may not be able to run like Google, there is one important way others can emulate its engineering success.
The problem with enterprise IT consistency (i.e., with implementing a single software stack across the organization that will bring order to chaos) is that enterprise IT isn’t static. Because technology isn’t static. As Google’s Kelsey Hightower put it in an interview with Comcast’s chief software architect Jon Moore, “Most organizations learn over time. So whatever you build today as the [default stack] is going to branch out based on new learnings.”
So how should an enterprise standardize, thereby reaping cost savings and productivity gains? You don’t, said Hightower. At least, not once and for all: “Standardize where you can, but allow things to grow apart and then re-standardize those things as you go and just follow the path of evolution.”
But how to standardize?
If your company has gotten behind in its big data management, there are ways to get current using automation.
Not all companies have the resources to develop their own big data management systems, hire the staff to manage them, and glean all the information they can from it. With large amounts of data coming in, this can lead to massive data management challenges.
SEE: Report: SMB’s unprepared to tackle data privacy (TechRepublic Premium)
“There are companies like Netflix and Twitter that immediately understood the value of big data and that had the resources needed to develop big data staffs and applications at the onset of the big data movement,” said Monte Zweben, CEO of Splice Machine, a company that provides a SQL-database platform that’s specifically tailored for accelerated big data modeling and deployment. “Unfortunately, most of the Fortune 2000 companies couldn’t compete with these efforts, so
Machine learning typically requires tons of examples. To get an AI model to recognize a horse, you need to show it thousands of images of horses. This is what makes the technology computationally expensive—and very different from human learning. A child often needs to see just a few examples of an object, or even only one, before being able to recognize it for life.
In fact, children sometimes don’t need any examples to identify something. Shown photos of a horse and a rhino, and told a unicorn is something in between, they can recognize the mythical creature in a picture book the first time they see it.
Now a new paper from the University of Waterloo in Ontario suggests that AI models should also be able to do this—a process the researchers call “less than one”-shot, or LO-shot, learning. In other words, an