Month: May 2016

An illustration of what some of the new exoplanets announced today may look like, along with their host stars. (credit: NASA)

It is one thing to observe the periodic dimming of a star’s light, as NASA’s Kepler Space Telescope has done for thousands of planet “candidates” since its launch in 2009. However, to confirm that such dimmings are in fact due to a planet passing in front of a star, as opposed to any number of false positives such as a binary star companion, requires intensive follow-up work with ground-based instruments, most often a measurement of radial velocity to determine the object’s mass.

To ease this workload, planetary astronomers have devised a few different statistical techniques, but none have been fully automated until now. Princeton University researcher Timothy Morton has developed software that can, within a few minutes, asses the orbital period and other data gathered by Kepler to assign a statistical probability that planet “candidates” are, or are not, planets. When tested on previously confirmed exoplanets and false positives, the new technique worked almost flawlessly.

Described in the Astrophysical Journal, this new method allowed scientists to rapidly assess Kepler’s planet candidate catalog from July, 2015, which identified 4,302 potential planets. Of these, the new technique found that 1,284 were planets at a confidence level of 99 percent or greater. An additional 1,327 are likely planets, but did not reach the 99 percent confidence threshold.

(credit: Ann Oro)

Windows users woke up to something that doesn't happen every day: the disclosure of two zero-day vulnerabilities, one in the Microsoft operating system and the other in Adobe's Flash Player.

The Windows bug is being actively exploited in the wild, making it imperative that users install fixes that Microsoft released today as part of its May Patch Tuesday. Cataloged as CVE-2016-0189, the security flaw allows attackers to surreptitiously execute malicious code when vulnerable computers visit booby-trapped websites. In the days or weeks leading up to Tuesday, it has been exploited in targeted attacks on South Korean websites, according to a blog post published by security firm Symantec. Technically, the vulnerability resides in the JScript and VBScript engines, but IE is the vehicle used to exploit it.

Separately, Adobe officials warned that a newly discovered Flash vulnerability also gives attackers the ability to remotely hijack machines. It was first reported by researchers from security firm FireEye, and exploits exist in the wild. Adobe said it planned to release an update as soon as Thursday.

(credit: Amazon)

Amazon is launching a new video service called "Amazon Video Direct." The new service entices professional video creators to upload their videos to Amazon, where they will be displayed on the Amazon Video site alongside studio-created TV shows and movies. The videos will be viewable by "all Amazon customers" via an ad-supported model, shown to Amazon Prime Video subscribers (presumably without ads) or available as a one-time rental or purchase. The service is launching in the United States, United Kingdom, Germany, Austria, and Japan.

It's easy to see "Ad-supported video" and label Amazon Video Direct as "a YouTube competitor," but Amazon is clearly only aiming for the "professional" end of the YouTube spectrum. Uploading a video requires that users first create an account (a regular Amazon account won't work) with a "company" name. It's also mandatory to connect a bank account and submit tax information so Amazon can distribute all the money you'll be making. The paperwork required just to upload a video takes it out of the running for the viral cat videos that pop up on YouTube—this service would be more for the Machinimas or Finebros of the world.

The launch partners give a good idea of what the market Amazon is aiming for. The press release states "AVD launch partners include: Conde Nast Entertainment, HowStuffWorks, Samuel Goldwyn Films, The Guardian, Mashable, Mattel, StyleHaul, Kin Community, Jash, Business Insider, Machinima, TYT Network, Baby Einstein, CJ Entertainment America, Xive TV, Synergetic Distribution, Kino Nation, Journeyman Pictures, and Pro Guitar Lessons." (Disclosure: Conde Nast owns Ars Technica, so maybe our videos will show up there someday.)

(credit: Flickr/Jeffrey Smith)

Measuring ‘drunk’ is pretty easy; the more alcohol someone drinks, the more alcohol shows up in that person’s blood and the more impaired that person becomes, falling somewhere on a scale of tipsy to wasted. Measuring ‘high,’ on the other hand, is far hazier—much to the dismay of some states' law enforcement.

Blood tests that try to quantify marijuana use are in fact useless at assessing how impaired a driver is, according to a study by the AAA Foundation for Traffic Safety. In other words, the study found that people with low blood amounts of THC—or delta-9-tetrahydrocannabinol, the main psychoactive component of pot—may still act as if they’re really stoned. On the other hand, some people may have THC measurements off the charts yet still act normally.

The finding is critical because several states have already set legal limits for the amount of THC a person can have in their blood while driving. AAA concluded that such limits are “arbitrary and unsupported by science, which could result in unsafe motorists going free and others being wrongfully convicted for impaired driving.”