Uber announced that it was expanding its autonomous taxi service to San Francisco—but it’s already getting into trouble

unnamed

Uber announced that it was expanding its autonomous taxi service to San Francisco—but it’s already getting into trouble.

Within hours of its cars hitting the roads, there were at least two reports of the company’s self-driving vehicles running red lights. The company is investigating the incidents, calling safety a “top priority,” but hasn’t confirmed or denied whether the cars were driving autonomously at the time.

Meanwhile, the company also appears to be running regulatory red lights, too. California officials warned the ride-hailing company to halt its autonomous operations in the city because it doesn’t have a permit to test the vehicles.

Uber was aware that this issue might surface, but insisted that it had “looked at this issue carefully” and decided that, on balance, it didn’t require a permit to start testing.

City officials have threatened the company with legal action it it doesn’t stop tests.

Posted in: Infographic of the day | Leave a Comment

This is the tiny super-computer companies are desperate to get their hands on

15openai_lead_image_web
Companies grappling with complex data increasingly want to use machine learning, so it’s perhaps not surprising that they’re scrambling to get hold of Nvidia’s new “AI supercomputer in a box.”

Its anonymous grey case is stuffed with eight of the chip maker’s highest-end graphics processing units, promising data scientists the ability to run deep-learning models without much of the hassle usually associated with building a heavy-duty AI rig.

The price tag: a cool $129,000.

Early adopters of the machine seem to think it’s a price worth paying.

Companies that have bought one say that it lets them train their analytical models faster, enables greater experimentation, and could facilitate breakthroughs in science, health care, and financial services.

Not bad for a dull-looking box. Time to save up, or add it to the Christmas list.

Posted in: Infographic of the day | Leave a Comment

Facebook unveils new tools to crack down on fake news

160721113604-zuckerberg-aquila-smile-780x439

Facebook has unveiled new tools to crack down on fake news. Journalists from various organizations can now tag stories as “disputed by 3rd party fact-checkers,” and users can flag content as a “fake news story.” The company’s algorithm has also been tweaked, and stories deemed fake can’t appear as paid ads.

In other news, Mark Zuckerberg did not attend the summit meeting between Silicon Valley leaders and Donald Trump. The only other tech leader not to appear was Twitter’s Jack Dorsey. A Trump spokesman said the site was not “big enough” to justify an invite. Unnamed sources at the company say it was because CEO Jack Dorsey refused to make a #CrookedHillary emoji.

Trump of course still uses Twitter regularly.

Posted in: Infographic of the day | Leave a Comment

Investors question Facebook board behavior

160721113604-zuckerberg-aquila-smile-780x439
Facebook investors are accusing Marc Andreessen, one of the company’s board members, of colluding with Mark Zuckerberg rather than representing their best interests—bad news for a company already facing intensifying public scrutiny.

Recently unsealed court filings reveal that Andreessen gave Zuckerberg regular updates of board discussions about proposals to allow the CEO to sell his Facebook shares while maintaining majority voting status, as well as enabling him to retain control of the social network if he pursues government service.

Investors argue that, by passing information via text messages to Zuckerberg and helping him negotiate with the board, Andreessen betrayed their trust.

The news is poorly timed for Facebook, as the company comes under questioning over issues of fake news and its role as a media organization.

Many commentators are calling for people to stop taking Facebook’s word at face value.

As the Guardian suggests, it may soon find its power being challenged—potentially from numerous angles.

Posted in: Infographic of the day | Leave a Comment

Facial recognition science has grown up

A software tool called Face2Gene can identify a distinctive constellation of facial features to diagnose genetic conditions.

Meanwhile, a more controversial machine learning approach from  Shanghai Jiao Tong University in China has been used to identify criminals from their faces, with a claimed accuracy of 89.5 percent.

And the app Blippar is using facial recognition to drive augmented reality software that allows users to, quite literally, put a name (and rather more besides) to a face.

Our features can clearly reveal plenty about us, and algorithms will only get better at reading them: a recent MIT study, for instance, will help computers identify faces from unusual angles, much like humans can.

Posted in: Infographic of the day | Leave a Comment