Showing posts with label artificial intelligence. Show all posts
Showing posts with label artificial intelligence. Show all posts

Thursday, June 23, 2016

Thursday: Rough Beast

Mere anarchy is loosed upon the world,
The blood-dimmed tide is loosed, and everywhere
The ceremony of innocence is drowned;
The best lack all conviction, while the worst
Are full of passionate intensity.

— excerpt, The Second Coming by William Butler Yeats, c. 1919

This lovely bit of atmospheric electronica by åpne sinn from the 2012 album entitled en seier is an odd fit for the anarchic theme. en seier is Norwegian for ‘one victory’, which is how åpne sinn‘s Steve Brand characterizes each day of life after surviving a heart attack, a personal apocalypse like Yeats’ rough beast of The Second Coming.

The last couple of weeks culminating in today’s Brexit referendum feel like the onset of a global heart attack. Hope we enjoy a victory after this strife, but it’s too soon for more than hope.

Still a little off, not up to any more reading today after staying up far too late watching the House Democrats’ sit-in protest last night. Here’s a few things worth looking at:

  • DIESELGATE: Fuzzy definition of ‘cold’ may have led to EU passenger diesel cars’ spewing more NOX than expected (Ars Technica) — This means ALL EU makers of passenger diesels, not just Volkswagen Group, are producing too much NOX at low temperatures within an unclear range of ‘cold’. We’ve already seen a lawsuit in the U.S. against Mercedes for this reason.
  • ARTIFICIAL INTELLIGENCE: ‘Concrete Risks in AI Safety’ (paper, PDF at arXiv.org) — Researchers from Google Brain, OpenAI, Stanford University and UCBerkeley looked at the potential risks of using AI. specifically

    …the problem of accidents in machine learning systems. We define accidents as unintended and harmful behavior that may emerge from machine learning systems when we specify the wrong objective function, are not careful about the learning process, or commit other machine learning-related implementation errors.

    Worthwhile read, but I see a problem already, though, a blindspot they come so close to resolving but fail to recognize. The research team is not particularly diverse, nor is the AI development community. They will program systems based on a world as they understand it, not as it is even though they believe they are programming for a rich and noisy environment. The results will be far worse than blue screens of death.

  • ELECTRIC VEHICLES: Siemens suggests trucks powered by catenary electric lines (QZ) — Interesting concept, using electricity a la late 1800s trolley cars versus expensive and weighty batteries. But the infrastructure required…I don’t know. But that’s a lot more potential profit for Siemens, eh?
  • ZIKA VIRUS: Lame Congress funding bill on Zika defense is lame (Jezebel) — Really can’t improve on Jezebel’s hed: ‘
    Amid Gun Control Protests, House Passes a Shitty, Ineffective Zika Virus Bill‘. Yeah, that, especially the part where religious fanatics put their personal faith ahead of suffering by ensuring no funds are used by Planned Parenthood for birth control to prevent conception by persons infected with Zika. Hey You Radical Fundamentalist Anti-Science Freaks In Congress: Get Ready To Pay For Many Microcephalic Babies’ Lifetimes And Guillain-Barré Syndrome Care. Pro-life my left arse cheek. More like pro-torture for women and babies.

That’s enough for today, I need to reserve my strength for the outcome of the Brexit referendum. Toodles.

Thursday, May 12, 2016

Thursday Not-Morning: Stupid

Jeepers. I need hip waders. There is just so damned much stupid over the last 24 hours. It’s a veritable flood.

The Future is here, and it’s stupid

  • Law firm “hires” first artificially intelligent lawyer (Futurism) — Oh how nice. Treat human misery like a fungible commodity by using IBM’s AI ‘lawyer’ Ross to process bankruptcies. Want to bet it’s cheaper to hire paralegals to do the work Ross does? Want to bet Baker & Hostetler’s Ross will be replaced by a competing internet-based firm processing bankruptcies even more inexpensively? Hey Congress: doesn’t it say something to you about the number and kind of bankruptcies when a ‘robot’ can process them?
  • Facial recognition expected to be $6 billion by 2020 (Curatti) — No invasion of privacy issues there, nor any security risks whatsoever. No chance at all two or more people have the same facial characteristics in terms of dimension.
  • Chinese tech company prepares for future where our consciousness lives forever in a computer (Bloomberg) — This is really creepy, and yet very much possible in the near-term future. If AI can nearly reproduce you from your social media, why can’t it replicate your consciousness?

The Past remains, and it’s stupid, too

  • Staffing company Portico sent home a receptionist for not wearing high heels (BBC) — A petition emerged in response, asking Parliament to outlaw such policies; 100,000 signatures mustered overnight. They’ve reversed their position today after a furor arose about their policy requiring women to wear 2-4 inch high heels on the job at a PriceWaterhouse Cooper facility. PwC says it’s not their policy. Come on now — it’s 2016, not 1956. It’s just plain stupid to ask workers of a specific gender to wear attire for looks — attire which causes discomfort and is not recommended by doctors.
  • Belgian beer company changes iconic American brand name to pander to voters (AdAge) — Take one of the oldest and most recognized U.S. brands on which hundreds of millions of dollars have been spent to entrench an immigrant’s name into the American psyche. Then remove it and replace it with the country’s name for six months. My gods, the stupid on this one. Fortunately a West Michigan brewer is taking advantage of this opportunity with ‘Murica! I could use one right about now.
  • Some SAP accounting software users attacked because they screwed up in 2010 (The Register) — Talk about time travel. I’m sure there’s some folks who’d like to go back to 2010 and execute that security patch correctly this time before hackers smite their business to smithereens.

The Present’s no gift

  • Don’t feed the sea turtles (Scientific American) — Surprise! When tourists feed junk food to sea turtles, the turtles’ health mirrors that of humans fed the same crap.
  • Study: Ransomware cybercriminals provide better, faster service than internet service providers (Nature) — Not even a rational comparison next to Comcast. Seems like there’s a market opportunity here; if crooks held a machine hostage AND offered a PC tune-up, would PC owners happily fork over cash? Hmm.
  • Marijuana use during pregnancy increases risk for pre-term birth (ScienceDaily) — What a surprise that a psychoactive drug combined with toxic by-products from smoking a plant product might have negative effects on pregnancy.

Ugh. Hope tomorrow is kinder to us. See you in the morning!

Monday, March 21, 2016

Monday Morning: Synthesized Brain

When you need a break this hectic Monday morning, take five minutes and watch ANA from Factory Fifteen. I'm intrigued by the props and set -- how much is CGI, and how much is actual production line? What company allowed this production company access to their equipment? Though snappy and visually engaging, the story's not realistic -- yet. But much of the equipment on the production line is very close to that used in manufacturing today. And just as depicted in this short film, the weakest link is the human. Worth keeping in mind this week as we plow deeper into the conflict at the intersection of humans and devices. Speaking of which... Apple-heavy week ahead
  • Hearing in California tomorrow in front of Judge Sheri Pym over the San Bernardino's shooter's iPhone. Be sure to read Marcy's take on the hearing and witnesses.
  • WLTX of Columbia SC posted a timeline of #AppleVsFBI events -- unfortunately, it starts on February 16 with Judge Pym's order to Apple.
  • NYT reported last week that Apple employees may quit if Apple is ordered to cooperate and write security-undermining code. But is this a deliverable in itself? The article offered an incredible amount of detail about Apple's operations; if employees quit, any entities observing the technology company will know even more. Has this shakedown been designed to yield information about Apple's operations, while risking corporate and personal security?
  • Apple will release information about new products today at a media event. The buzz may be less about the new products than the hearing tomorrow.
  • An iPhone 6 bursting into flames during a flight to Hawaii didn't help Apple. One might wonder why this particular phone flamed out so spectacularly as it's a relatively new device.
HEADS UP TECH USERS
  • Kindle users: Amazon is forcing a mandatory update across all its older Kindle reader devices. Deadline: TOMORROW MARCH 22 -- after that date, users will have to manually update devices and download books via PC and not over the internet.
  • Tweetdeck users: Owner Twitter will kill the Windows app on April 15th. After that time, Windows-based users will need to use a browser. Can't blame Twitter--it's ridiculously expensive to write and service so many apps when the same devices usually have a browser.
  • Android users: 1) Protect your privacy and security by checking these settings; 2) Check this setting, stat, to prevent unauthorized access.
  • Nexus users: Make sure you have the latest patch issued last week. All other Android users should nag their equipment makers for their version of the same patch.
Before the machines complete their occupation of our world...
  • Nice read on law emerging with the rise of robots. Too bad none of them really incorporate Asimov's Three Laws of Robotics. (The Atlantic)
  • Want to bet the overlords will argue workers should be paid less because they don't have to work as hard wearing an exoskeleton -- like these at Panasonic? (By the way, DARPA, that's yet another commercially-developed exoskeleton near release; where's yours/ours?) (Mashable)
  • Artificial intelligence already pitted against humans by those bloody banksters. Watch this video and ask yourself if this guy from Global Capital Acquisitions realizes there are humans at the nodes of the investment network whose lives are affected by his blah-blah-blah-babbling about artificial intelligence. STG he could be a machine himself. (Bloomberg)
  • Myths about AI busted - another solid read. Combined with the preceding Bloomberg bankster video it reinforces AI threat awareness. (Gizmodo)
After watching that video at Bloomberg, I think we're a lot closer to ANA than we realized. Watch your backs -- Monday is certainly gaining on you, if robots aren't.