AI Weekly: The 3-pronged AI startup test
摘要：Hey, everybody!When I talk with people about being an AI reporter, one of the topics that tends to come up is the issue of evaluating companies that claim to be working with AI in some capacity. It can be hard to evaluate how different companies compare, especially with new startups cropping up on a weekly, if not daily, basis.
When I talk with people about being an AI reporter, one of the topics that tends to come up is the issue of evaluating companies that claim to be working with AI in some capacity. It can be hard to evaluate how different companies compare, especially with new startups cropping up on a weekly, if not daily, basis.
To help with my own workload, I’ve developed a three-point test, which I apply to every company that lands in my inbox. To pass, companies need to have these three elements: AI experts on staff; experts in the field that they’re focusing on; and a unique dataset. An early-stage company that is weak in one of those areas may still turn into a remarkable business, but I’d want to see some level of self-awareness about either bringing in the needed expertise or data.
Here’s the reasoning behind my criteria. Experienced staff is necessary because building an AI system is hard and requires specialized skills, which calls for people who have past work in the field. Depending on how sophisticated and novel a company’s application of AI is, they’ll need varying degrees of expertise in that field. If you’re trying to build a machine learning system using techniques nobody else has, the people building that system better be at the top of their game.
Subject matter experts, meanwhile, are needed to make sure that the AI application being built is solving the right problems. People don’t go out and buy two quarts of AI. They buy a product to solve a problem, so the application of artificial intelligence should create results that outperform predecessors and fit into customers’ existing workflows. People who understand the market that a product will enter are critical to getting that piece right.
The creation or collection of a unique dataset is all about ensuring that a product is defensible. If it’s possible for a team at Google, Salesforce, Microsoft, or Amazon to come along and replicate or exceed a company’s results without much effort, that’s bad news. Businesses should be working to acquire and protect unique information that can then be used to build better AI systems.
If a company only meets two of these three criteria in the long run, they’re open to attack by a better prepared competitor. Businesses without sufficient AI expertise will have a hard time developing the intelligent applications they need. Those without relevant subject matter experts will have to learn what their customers need in order to apply AI to it. And without the right data, the systems won’t provide the best insights that they could.
Thanks for reading,
Blair Hanley Frank
AI Staff Writer
P.S. Please enjoy this video:
FROM THE AI CHANNEL
As a longtime Destiny player, I’ve occasionally found myself wishing for the companionship of one of the game’s Ghosts — floating robots that offer commentary, guidance, and the occasional wisecrack, along with persistent resurrection. Bungie, the online shooter’s studio, set out to offer players something like this (minus the functional immortality) with a limited edition […]
The software development kit (SDK) for Google Assistant is expanding to additional countries today, including Australia, Canada, Germany, Japan, and the United Kingdom. As part of the expansion, Google Assistant SDK is now available in English, French, German, and Japanese, according to a Google blog post. The SDK makes it possible for hardware makers to […]
Alongside the growth in adoption of conversational AI like Amazon’s Alexa and Google Assistant, social robots are beginning to enter the home to help people get things done and play games together. Unlike other electronics most often treated as inanimate objects, with social robots there’s more to consider, such as their personality or how they […]
The concept of AI has been a subject of fascination for almost as long as computers have existed. However, we’ve only recently begun to see what a future with AI might actually look like and — despite the grim picture that is often painted in sci-fi stories — the outlook is pretty exciting. Luckily for […]
Amazon’s Alexa has long had games inspired by Batman, SpongeBob SquarePants, and Jeopardy, but today, as Echo Buttons go on sale for the first time, Amazon is introducing its first accessory for multiplayer gameplay. At launch, Echo Buttons work with voice app games like Fourth Down Football Trivia with Philip Rivers and Sounds Fun with Mike Epps. Developers […]
Without a doubt, smart speakers like the Amazon Echo or Google Home will be some of the most popular gifts this holiday season. But unlike last year, when there were only a handful of choices, you now have dozens of ways to speak with AI assistants like Alexa, Google Assistant, and Cortana. Some speakers offer […]
Silicon Valley Is Turning Into Its Own Worst Fear
We asked a group of writers to consider the forces that have shaped our lives in 2017. Here, science fiction writer Ted Chiang looks at capitalism, Silicon Valley, and its fear of superintelligent AI. (via BuzzFeed)
DeepMind claimed this month its latest AI system – AlphaZero – mastered chess and Shogi as well as Go to “superhuman levels” within a handful of hours. (via The Register)
Over the past year, we’ve been comparing Google Maps and Apple Maps in New York, San Francisco, and London—but some of the biggest differences are outside of large cities. (via Justin Obeirne)
Last week, Netflix decided to have some holiday fun courtesy of its user data. So the streaming service took to Twitter to pose the question, “To the 53 people who’ve watched A Christmas Prince every day for the past 18 days: Who hurt you?” (via The New York Times)