“links” Category

Andrew Ng’s Deep learning Quest, Google and Apple

Andrew Ng is helping lead a group at Google dedicated to making giant advances with neural networks:

It was a shift that would change much more than Ng’s career. Ng now leads a new field of computer science research known as Deep Learning, which seeks to build machines that can process data in much the same way the brain does, and this movement has extended well beyond academia, into big-name corporations like Google and Apple. In tandem with other researchers at Google, Ng is building one of the most ambitious artificial-intelligence systems to date, the so-called Google Brain.

Pretty good piece about the increasing overlap of neuroscience and neural network research for technological purposes, but what I want to emphasize is how much Google has invested in neural networks (or “artificial intelligence” generally, if you’d rather, but that term is pretty misleading). Both Apple’s and Google’s futures depend heavily on using user data and other data sources to provide value for users, and Google has a huge advantage here because they’ve been investing heavily in it for a very long time. It’s just as important to Apple, but Apple had to acquire the Siri team to gain the capability. That’s a huge disadvantage.

This isn’t just about speeding up voice recognition or making it more accurate, although that is an advantage Google Now has over Siri—using voice recognition in Google’s iOS search app feels much faster than Siri because it shows you what it thinks you’re saying as you say it. It’s much more than that; since this has been something very important to Google for a long time, and something of an intrinsic organizational competence, Google can move much quicker to develop the capability in Google Now than Apple can. Apple must move even quicker to make it a skill for Apple, too, and to take advantage of their own unique resources that Google doesn’t have.

May 9th, 2013

Paper for iPad’s Pinch-to-Zoom

Paper for iPad now has pinch-to-zoom support.

As usual with Paper, it’s a bit different than normal pinch-to-zoom. Rather than zoom the entire canvas, pinching places a loupe-like circle around the area you’re zooming on so you can make your changes while still retaining the entire canvas’s context. Smart.

One slightly related note, though: I still can’t get Paper’s two-finger undo gesture to work reliably for me. Gestures can make applications much better, but they can also make it maddening.

May 9th, 2013

iWork ’09

Peter Cohen:

On any given day, a quick check of the top-selling paid apps list in the Mac App Store will reveal Apple’s Keynote, Pages and Numbers in the top ten. It’s surprising, given that each of those apps was originally bundled as Apple’s iWork ’09 productivity suite, released in, you guessed it, January, 2009. It makes me wonder when or if we’ll ever see an update to them.

May 9th, 2013

Riker Sits Down

Something I sure didn’t notice while watching the Next Generation: Riker sits down like a crazy person.

May 9th, 2013

‘Before Midnight’ Interview

Ethan Hawke, Julie Delpy and Richard Linklater discuss “Before Midnight”:

Hawke: The first two films are so much about romantic projection. The third had to be the opposite of that. We couldn’t play that trick again.

Delpy: But it couldn’t be totally taken away from that romantic idea — otherwise it’s depressing.

One of the movies I’m most excited for this year.

May 8th, 2013


Thinglist is an iPhone app for remembering things you want to try. Movies, books, restaurants, music, places. Great idea, and the app looks beautiful.

May 8th, 2013

Data Analyzing Your Way to a Box Office Hit

For just a small fee, you too can have your screenplay analyzed for maximum box office effectiveness:

A chain-smoking former statistics professor named Vinny Bruzzese — “the reigning mad scientist of Hollywood,” in the words of one studio customer — has started to aggressively pitch a service he calls script evaluation. For as much as $20,000 per script, Mr. Bruzzese and a team of analysts compare the story structure and genre of a draft script with those of released movies, looking for clues to box-office success. His company, Worldwide Motion Picture Group, also digs into an extensive database of focus group results for similar films and surveys 1,500 potential moviegoers. What do you like? What should be changed?

“Demons in horror movies can target people or be summoned,” Mr. Bruzzese said in a gravelly voice, by way of example. “If it’s a targeting demon, you are likely to have much higher opening-weekend sales than if it’s summoned. So get rid of that Ouija Board scene.”

How long before a similar process just writes screenplays? And, really, what would be lost?

May 8th, 2013

Racism At the White House Correspondents’ Dinner

Seema Jilani had this experience while attending the White House correspondents’ dinner:

As I left the hotel and my husband went to the ballroom for the dinner, I realized he still had my keys. I approached the escalators that led down to the ballroom and asked the externally contracted security representatives if I could go down. They abruptly responded, “You can’t go down without a ticket.” I explained my situation and that I just wanted my keys from my husband in the foyer and that I wouldn’t need to enter in the ballroom. They refused to let me through. For the next half hour, they watched as I frantically called my husband but was unable to reach him.

Then something remarkable happened. I watched as they let countless other women through — all Caucasian — without even asking to see their tickets. I asked why they were allowing them to go freely when they had just told me that I needed a ticket. Their response? “Well, now we are checking tickets.” He rolled his eyes and let another woman through, this time actually checking her ticket. His smug tone, enveloped in condescension, taunted, “See? That’s what a ticket looks like.”

When I asked “Why did you lie to me, sir?” they threatened to have the Secret Service throw me out of the building — me, a 4’11″ young woman who weighs 100 pounds soaking wet, who was all prettied up in elegant formal dress, who was simply trying to reach her husband. The only thing on me that could possibly inflict harm were my dainty silver stilettos, and they were too busy inflicting pain on my feet at the moment. My suspicion was confirmed when I saw the men ask a blonde woman for her ticket and she replied, “I lost it.” The snickering tough-guy responded, “I’d be happy to personally escort you down the escalators ma’am.”

May 7th, 2013

Amazing Fact! Science Proves Health Insurance Doesn’t Improve Health (Maybe)

One interesting side-story to the Oregon Medicaid study story is that the researchers initially published positive results for Medicaid. As a result, looking at how pundits responded to those initial findings and today’s final findings is… interesting. Peter Suderman’s done a good job of it:

Aaron Carroll, an influential health policy blogger at The Incidental Economist, emphasized the rigor of the study. “I’d like to reiterate that this was a randomized controlled trial,” he wrote. “An RCT is pretty much the best way to prove causality, especially if it’s well done.” And because it’s an RCT, he concluded, “we can even start talking causality.” Ezra Klein published a column touting the study with the headline, “Amazing Fact! Science Proves Health Insurance Works.” He explained why the randomized study was so valuable: “The gold standard in research is a study that randomly chooses who gets a new treatment and who doesn’t. That way, you know your results are unaffected by differences in the two populations you are studying.”

Now, well, it’s all a little less clear. “The problem with the Oregon study,” Klein wrote this morning,” …is we don’t really know what we’re learning.” Carroll, who was ready to start talking causality when the first study was published, is now counseling caution. “So chill, people. This is another piece of evidence. It shows that some things improved for people who got Medicaid. For others, changes weren’t statistically significant, which isn’t the same thing as certainty of no effect. For still others, the jury is still out.”

I like Klein and Carroll’s caution today. I just wish they had it when the results were, um, a bit more friendly to their prior beliefs, too.

May 6th, 2013

Barro on the Oregon Medicaid Study

Josh Barro’s take on the Oregon Medicaid study is good:

Yesterday, we got results from the two-year Oregon Health Study, which randomly assigned some low-income people to receive Medicaid coverage while others did not. The study found that Medicaid led people to consume more health care and was effective in reducing both financial strains due to medical costs and depression. But it did not find significant effects on the physical health measures that were tracked.

Despite efforts to spin it to the contrary, this is bad news for advocates of the Medicaid expansion. While Medicaid is clearly good for some things, it was supposed to be good for all of the measures tracked.

May 6th, 2013

Adobe’s Project Mighty smart stylus and Napoleon ruler

Adobe has been working on a pen and ruler for tablets, and the combination looks awesome.

The pen is pressure-sensitive, but the coolest part is the “ruler,” which allows you to draw straight lines or precise arcs. Really, really cool.

The demonstration, though, shows a very large lag between drawing and the line actually appearing on screen. It’s strange because it’s much, much worse than I’ve experienced with any drawing or sketching app on iOS.

In any case, it’s heartening to see Adobe working on meaningful projects like this.

May 6th, 2013

Enstitute, apprenticeships for where college fails

Enstitute is a group providing apprenticeships for college-age people:

How did she catapult from dropping out of college to landing a plum job? She became an apprentice to Hilary Mason, chief data scientist at Bitly, through a new two-year program called Enstitute. It teaches skills in fields like information technology, computer programming and app building via on-the-job experience. Enstitute seeks to challenge the conventional wisdom that top professional jobs always require a bachelor’s degree — at least for a small group of the young, digital elite.

“Our long-term vision is that this becomes an acceptable alternative to college,” says Kane Sarhan, one of Enstitute’s founders. “Our big recruitment effort is at high schools and universities. We are targeting people who are not interested in going to school, school is not the right fit for them, or they can’t afford school.”

Colleges are incredibly expensive, the cost continues to rise, and yet they are increasingly less effective at preparing people to be successful. There’s absolutely value in a liberal arts education (in fact, I think there’s even more value now), but many schools don’t even do a good job of exposing students to a variety of disciplines to make them more well-rounded. Universities crank students through, make them take class after class with lecture-midterm-lecture-midterm-lecture-final, put them tens of thousands of dollars into debt, and leave many of them not much better off than they were before entering.

So new education organizations like this should be welcomed. Perhaps they’re not exactly what we need to replace universities, but we don’t need to replace universities—we need different options, different paths, different ideas that allow people to take a route that fits them better, and places pressure on our bloated, staid education system to change.

May 6th, 2013

David’s Brisket House

David’s Brisket House:

The new David—a nickname he inherited along with the deli— turned it into his family’s business, and they’ve built that business into the kind of thing you’d only see in New York: a Jewish deli, run by Muslims with Brooklyn accents, for the benefit of Bed-Stuy.

May 6th, 2013

PDFpenPro 6 from Smile (Sponsor)

Big thanks to the people at Smile Software for sponsoring this week’s RSS feed. If you work with PDFs, you owe it to yourself to check out PDFpen.

PDFpenPro is the advanced version of PDFpen. PDFpenPro does everything that PDFpen does, such as add signatures, edit text and images, perform OCR on scanned documents and export Microsoft Word documents. It also has the ability to create a PDF form, build a table of contents, and convert HTML files to PDF.

The new PDFpenPro 6 adds document permission settings. When you share a PDF, you can restrict printing, copying, and editing of your PDFs. You can also use the new automatic form field creation tool to convert a non-interactive form into an interactive PDF form with text fields and checkboxes automatically added.

PDFpenPro 6 is available on the Smile Store and the Mac App Store for $100. A free demo can be downloaded on the Smile site. Find out why Macworld calls PDFpenPro “the crème de la crème of PDF editing and annotating applications.”

Sponsorship by The Syndicate.

May 6th, 2013

Bill Ayers defends Weather Underground bombings

Bill Ayers defended the Weather Underground bombings over the weekend:

“To conflate a group of fundamentalist people [in Boston] who are nihilistic in some way with a group of people who spent their lives trying to oppose the murder of 6,000 people a week … and still the killing went on. And still the killing went on. What would you have done?” Ayers said. “There’s no equivalence [with Boston]. Property damage. That’s what we did.”

In his talk to the crowd, Ayers mentioned that in 1970, he lost three friends in the Weather Underground, including his lover, Diana Oughton. He did not explain in his talk how they died – they were killed when nail bombs they were making in a Greenwich Village townhouse blew up.

Nail bombs are a hallmark of bombings intended to damage property, and not to harm people, I’m told.

May 6th, 2013