So it turns out the problems of AI are the problems of technology in general. Specification (making sure it does the job you want); robustness (the ability to cope with unexpected conditions); and the ability to monitor its performance and understand what it’s doing.
These are hard problems, even for a system designed from the ground up by humans, let alone one that has partly built itself.
Originally shared by Singularity Hub
DeepMind’s New Research Plan to Make Sure AI Is Safe
While a lot of doubt has been cast on the Sapir-Whorf Hypothesis (that we can only think about things we have words for) in detail, it does seem to hold to a certain extent. For example, people who don’t have words for numbers have only an approximate grasp on quantities, even small quantities.
Originally shared by Andy Brokaw
https://www.quickanddirtytips.com/education/grammar/languages-without-numbers
I have a story idea cooking about completely automated supply chains, where you can decide you want a whatever and say (in effect) “Accio whatever!” and it gets delivered to your hand in short order, without another human having touched it or even known about it.
Still looking for the storytelling angle, but I need to get on with it before real life overtakes me.
Originally shared by Singularity Hub
This Robotic Warehouse Fills Orders in Five Minutes, and Fits in City Centers
It’s a Peter Diamandis article, so it’s unflaggingly techno-optimist and a bit libertarian. But he does explore some interesting initiatives going on around the world (even if the Chinese ones are a lot more disturbing to me than they are to him).
I’ve used the idea of a central, trusted ledger that facilitates automatic taxes in my latest (not-yet-published) novel, Capital Crimes.
Originally shared by Singularity Hub
Could Tech Make Government As We Know It Irrelevant?
Not yet ready for prime time.
Originally shared by HACKADAY
A robot made my hamburger
The future is upon us and the robots will soon take over. Automated cars will put Uber drivers and cabbies alike out of work. Low-wage workers, like the people working behind the counter at McDonalds, will be replaced by burger-flipping robots. The entire…
http://hackaday.com/2018/10/11/i-ate-a-robot-hamburger-before-the-restaurant-went-out-of-business/
Originally shared by Singularity Hub
No Safety Driver Here—Volvo’s New Driverless Truck Cuts the Cab
When you’re solving a complex problem, try everything.
Originally shared by Singularity Hub
The 7 Wild Energy Technologies That Just Got a Billion-Dollar Boost
Originally shared by Kam-Yung Soh
Link includes a video showing the robots in action. “If fiberglass tubes suddenly started to sprout out of the ground, you might think you had stumbled on some alien invasion. But such tubes are a real thing, woven by newly developed autonomous robots to create large structures such as bridges and temporary shelters—with minimal human input.
Each “Fiberbot” has a winding arm that pulls fiber from a tank on the ground, mixes the materials in a nozzle, and winds the wetted fiber around itself like a silkworm cocooning. Next, the robot turns on an ultraviolet light to cook the fiber into a hard tube. Then, it deflates its body and uses a tiny motor and wheels to inch itself up on top of the hardened fiber, where the process begins again.
The robots can tilt and use different winding patterns to vary the thickness and the direction of the tubes. As they build, the Fiberbots communicate with each other through a computer network to avoid running into each other or other obstacles. Together, they can calculate the most efficient way to build a given structure.”
The project: hook sensors to a car to an AI to a receipt printer, and have it write a road novel as it travels. It’s tempting to dismiss it as a gimmick, and the output as nonsense. But it might be the beginning of something.
https://www.theatlantic.com/technology/archive/2018/10/automated-on-the-road/571345/
In concept, it’s kind of interactive Twitch.
Originally shared by Ninja On Rye
Two people send thoughts to a third person to play a game of tetris in a 3-person brainnet proof of concept.
It’s currently low-bandwidth, but as the description below really captures the wow aspect of this research:
Stocco and his colleagues have created a network that allows three individuals to send and receive information directly to their brains. They say the network is easily scalable and limited only by the availability of EEG and TMS devices.