Welcome back! Thanks for reading. Get in touch with me by replying or emailing firstname.lastname@example.org. Erica, who writes Future on Wednesdays, is at email@example.com.
📺 Next on “Axios on HBO”: Iraqi President Barham Salih gives an exclusive interview, (sneak preview), IMF Managing Director Kristalina Georgieva talks socialism and taxes, and Sen. Kamala Harris gives her take on 2020.
Today's issue is 1,808 words, which shouldn't take more than 7 minutes to read.
Illustration: Eniola Odetunde/Axios
In the most exclusive AI conferences and journals, AI systems are judged largely on their accuracy: How well do they stack up against human-level translation or vision or speech?
Yes, but: In the messy real world, even the most accurate programs can stumble and break. Considerations that matter little in the lab, like reliability or computing and environmental costs, are huge hurdles for businesses.
Why it matters: Some stumbles that have tarred high-profile AI systems — like facial recognition that fails more often on darker faces, or medical AI that gives potentially harmful advice — have resulted from unanticipated real-world scenarios.
The big picture: Research often guns for incremental advances, juicing an extra percentage point of accuracy out of a previously proposed model.
That's one of several hidden considerations that can drive up the cost of creating an AI model that works in the real world. Among the other important factors accuracy doesn't capture:
"The machine learning model is just a tiny piece of the machine learning product," says Andrew Ng, founder of Landing.ai, a startup that helps companies set up AI processes.
One of the byproducts of the hidden costs associated with increasingly accurate AI systems is that they can make it hard for a new startup or cash-strapped university lab to compete.
What's next: Schwartz, Ng and others propose putting more effort toward solving problems more efficiently rather than making a slightly more accurate model at an enormous cost.
Go deeper: How not to replace humans
Illustration: Sarah Grillo/Axios
Machine vision is a crucial missing link holding back the robotization of industries like manufacturing and shipping. But even as that field advances rapidly, there's a larger hurdle that still blocks widespread automation — machine understanding.
Why it matters: Up against a shortage of workers, those sectors stand to benefit hugely from automation. But the people working in warehouses and factories could find their jobs changed or eliminated if vision technology sees new breakthroughs.
The big picture: Machine vision can help robots navigate spaces previously closed off to them, like a crowded warehouse floor or a cluttered front lawn. And it's critical for tasks that require dexterity, like packing a box with oddly shaped objects.
Driving the news: In a report first shared with Axios, LDV Capital, a venture firm that invests in visual technologies, predicts an upheaval in manufacturing and logistics, driven primarily by computer vision.
Yes, but: It'll take more than just high-fidelity cameras and fast AI perception to make an intelligent robot.
A broad understanding of the world helps us humans avoid confounding errors when we look around.
The big question: How much of the problem is solvable with incremental improvements in machine vision, before robots need better common sense?
The bottom line: "The Rubicon here, which we haven't crossed yet, is to not just be able to see objects," says Marcus. "It's interpreting scenes that will be the breakthrough."
Despite recession worries, demand for workers in the U.S. has climbed ever higher — and with it, the outlook for a bundle of jobs that could dominate future economies.
Why it matters: The fate of this seemingly future-proof work, much of it centered on interactions with increasingly intelligent machines, has looked rosy since 2016, but it's not clear how it would weather an economic downturn.
Driving the news: The latest numbers come from a quarterly report from IT consulting company Cognizant, provided first to Axios.
What's happening: Earlier this year, the "jobs of the future" index flattened along with larger hiring trends, suggesting that this work isn't entirely insulated from the rest of the economy.
Most future-oriented jobs are increasingly in demand, like health information managers and robotics engineers. But, surprisingly, tech consultants are among a few jobs that dove in Cognizant's most recent analysis.
Cognizant picked mostly white-collar jobs for the index, but several jobs outside that category are also good candidates for the list, too, either because they're byproducts of the tech industry or because they are difficult to automate with today's technology.
The big question: What happens when the economy eventually tanks?
Illustration: Eniola Odetunde/Axios
Detroit's gamble on the future (Joann Muller - Axios)
Google's contested claim of quantum supremacy (Elizabeth Gibney - Nature)
The AI face scans that make some hiring decisions (Drew Harwell - Washington Post)
Can you really be addicted to video games? (Ferris Jabr - NYT Magazine)
5G experiments on military bases (Roxana Tiron & Travis J. Tritten - BGov)
Photo: Tayfun Coskun/Anadolu/Getty
When the Golden State Warriors moved across the bay from Oakland to San Francisco this year, its players went apartment hunting.
Now, they have a familiar complaint, the San Francisco Chronicle's Connor Letorneau reports: The rent is too damn high.
Fun fact: Thirteen of the 14 Warriors have moved to San Francisco. The other one, Steph Curry, recently bought a $31 million property in Atherton, home to the country's most expensive zip code and an hour-plus trafficky commute to the new stadium.