Lifeboat Foundation: Safeguarding Humanity
Toggle light / dark theme

Blog

Category: computing

https://youtube.com/watch?v=a_mYvSVG3QY

For those unfamiliar, SXSW is a week-long, trendy, if not seriously geeky festival of film and culture, panels and discussions. This year, one of the strangest – and either most disturbing or most compelling, depending on where you stand – talks was delivered by Hiroshi Ishiguro, a Japanese inventor and roboticist. The Osaka University professor was speaking about human-like androids and what roles they might fill within society in the near future. Ishiguro discussed his greatest and most marvellous creation to date: a “Geminoid” (robot in his own likeness) whose human appearance has been deftly created through with a plastic skull, a metal skeleton and silicon skin – and is controlled by an external computer. It would be hard, at a glance, to tell the two apart. In fact, the Geminoid held an autonomous conversation in Japanese, on stage, in front of an audience of hundreds.

Geminoid is not Ishiguro’s first uncannily human robot. In 2005, he developed a female android named Repliee Q1Expo, telling the BBC, “I have developed many robots before, but I soon realised the importance of its appearance. A human-like appearance gives a robot a strong feeling of presence. Repliee Q1Expo can interact with people. It can respond to people touching it. It’s very satisfying.”

At SXSW on Sunday, Ishiguro discussed how he imagined these human-looking robots might become a part of the everyday sooner than we think; as receptionists, language tutors and museum-guides. In fact, he discussed how he and his team have tried and tested the robots in everyday situations. “Japanese males hate to talk to the shopkeeper because it signals they want to buy something,” he explained. “But they don’t hesitate to talk to the android.” He then jokingly added that it helps that “[a] robot never tells a lie, and that is why the android can sell lots of clothes.” Which begs a couple of questions including why do Japanese males have problems interacting with shopkeepers, and what happens to the shopkeeper in this scenario?

Read more

Have you ever wanted to mess with a video by making its cast say things they never would on camera? You might get that chance. Researchers have built a face detection system that lets you impose your facial expressions on people in videos. The software uses an off-the-shelf webcam to create a 3D model of your face in real time, and distorts it to fit the facial details in the target footage. The result, as you’ll see below, is eerily authentic-looking: you can have a dead-serious Vladimir Putin make funny faces, or Donald Trump blab when he’d otherwise stay silent.

This isn’t about to reach software you can buy, but the implications for video creation are big if it becomes more than a university project. You could use the tool to mess with your friends by having celebrities say audacious things, or have famous figures recite dialogue in movies without needing to painstakingly animate faces frame by frame. In other words: get ready for an era when even the most plausible videos aren’t safe from a little computer trickery.

Read more

And Steve Jobs was not yet back at Apple when he gave a remarkably prescient interview to Wired’s website the same year. Although the iMac, iPod, and iPhone were still years away, and Jobs was working at NeXT, he clearly saw where the computing industry was headed.

And although his later work at Apple clearly influenced the way things turned out, he still offered a slew of predictions that are shockingly accurate today.

Here’s what Jobs got right:

Read more

The EU-funded COLUMNARCODECRACKING project has successfully used ultra-high fMRI scanners to map cortical columns, a process that opens the door to exciting new applications, such as brain-computer interfaces.

Cortical columnar-level fMRI has already contributed and will further contribute to a deeper understanding of how the brain and mind work by zooming into the fine-grained functional organization within specialized brain areas.

By focussing on this, the project has stimulated a new research line of ‘mesoscopic’ brain imaging that is gaining increasing momentum in the field of human cognitive and computational neuroscience. This new field complements conventional macroscopic brain imaging that measures activity in brain areas and large-scale networks.

Read more

I did an interview on AI and politics for CBC, which also went out on NPR yesterday.

This week, Google’s artificially intelligent computer, AlphaGo won a tournament in the complex board game called Go. American presidential candidate Zoltan Istvan says it’s that in a matter of 10 to 15 years A.I. will be advanced enough to be president of the United States of America.

Read more

Researchers in Finland have figured out a way to reliably make quantum computers — technology that’s tipped to revolutionise computing in the coming years — even more powerful. And all they had to do was throw common sense out the window.

You’re almost certainly reading this article on a classical computer — which includes all phones, laptops, and tablets — meaning that your computer can only ever do one thing at a time. It reads one bit, then the next bit, then the next bit, and so on. The reading is lightning fast and combines millions or billions or trillions of bits to give you what you want, but the bits are always read and used in order.

So if your computer searches for the solution to a problem, it tries one answer (a particular batch of ones and zeros), checks how far the result is from the goal, tries another answer (a different batch), and repeats. For complicated problems, that process can take an incredibly long time. Sometimes, that’s good. Very clever multiplication secures your bank account, and faster or more efficient equation-solvers put that in jeopardy.

Read more

Now, we’re hitting Terminator mode with this.

If you’re worried that artificial intelligence will take over the world now that computers are powerful enough to outsmart humans at incredibly complex games, then you’re not going to like the idea that someday computers will be able to simply build their own chips without any help from humans. That’s not the case just yet, but researchers did come up with a way to grow metal wires at a molecular level.

At the same time, this is a remarkable innovation that paves the way for a future where computers are able to create high-end chip solutions just as a plant would grow leaves, rather than having humans develop computer chips using complicated nanoengineering techniques.

DON’T MISS: iPhone 7: Everything we know so far

Researchers from IBM’s T.J. Watson Researcher Center are working to create wires that would simply assemble themselves in chips. The scientists use a flat substrate loaded with particles that encourage growth, and then add the materials they wish to grow the wire from.

Read more