RIP Google+


Way back on June 6th, 2012 in the third month of this here blog, I bemoaned the annexation of what was Google Places, by the then upstart social media initiative Google+. At the time I said:

“Google+ has enjoyed rapid growth with millions joining the platform, but there appears to be good data that although many join, the level or return visits and engagement is much lower than on Facebook. Given that, Google’s strategy of fusing Google+ with the very successful and highly used Google Places follows the Google Buzz strategy of essentially forcing a large number of users (in this case local businesses and reviewers) to join the game.”

And now, a few years later, even after they forced users and businesses to come, the party has never really improved. The good features they came up with — like the pics and video streams and hangouts — have been carved off the corpse, the leadership has gone, and we are left with a platform that doesn’t even require you to create a profile to use Google products like Gmail.

Although they haven’t officially stuck a fork in it yet, Google+ is clearly done. What contributed most to the demise? Well, it’s exactly as stated back in 2012: engagement (or lack thereof). I have a rule of thumb, the “ten minute test.” when I come across something interesting online, I dig in, then eventually end up deciding if I’m prepared to dedicate about ten minutes a day to it to stay engaged.

I’m probably a poor use case in the example of Google+, since I really don’t do much Facebook (let alone Google+). But what’s clear is that millions of users went there once, did what they had to do to keep using Google products, and never returned. The upside of that limited engagement was that it did get most of us permanently logged into Google in general, and it allowed Google to get a much stronger handle on local businesses in general. Beyond that, there simply wasn’t a compelling reason to keep folks going back. People are busy. Finding time to check Facebook, Twitter, and Instagram takes time. It feels like most of us held an informal meeting and agreed not to participate.

The other problem with Google+ (aside from its weirdly un-intuitive user interface) was how restricted it felt. Google keeps a famously vice-like grip on its products, and Google+ somehow felt like a social platform designed by the Post Office; not fun.

In any event, Google is clearly unwinding the remains. It’s not a huge problem for most people as they weren’t using it in the first place. Some marketeers were exploiting the heck out of it, but I imagine even they will get over it soon enough. It’s another example of Google being fabulously great at a handful of things (most importantly search and advertising) and stumbling over many others. They always want to play, but they typically show up late and demand by sheer willpower to be allowed into the game. They don’t always win.

When AI Meets PC


Google has spent many years and billions of dollars making its algorithms really smart. They detect behavior and the kind of content we look at, and then Google targets ads to users which its algorithms “think” fit them best. The recent calamity where it categorized African Americans as “Gorillas,” and more recently the one where it categorized a Chinese man as a “horse,” may be just the tip of a problem that most of us probably didn’t see coming.

If you make your algorithms super smart, the problem is they will behave as trained and work really hard to send the right ads to the best match, irrespective of the societal norms. In recent studies, researchers at Carnegie Mellon created fake male and female users and browsed sites which have a strong gender bias — think vs. It’s not an exact study, but it’s an interesting approach. They then set these fake identities to browse gender-neutral sites and tracked what ads Google presented to the various fake users.

The results were striking. Male users of news sites like CNN were more frequently shown job ads with higher salaries than females on the same sites. They also found that an image search for CEO only presented 11% female results. Frequent browsers about addiction were sent ads for rehab.

Some of these results can be explained empirically. The fact is there are many fewer female CEOs, so the number of images indexed is likely proportionately lower. Likewise, the addiction result. Google tracks your browsing history, so if you spend a lot of time looking at fly fishing sites, it’s likely that’s what you will get ads for on CNN.  The fact that it also works for addiction feels different to fly fishing, but it’s not really.

The jobs and gender question is much weirder. It suggests that Google is divining your gender from your behavior, then making a value judgment. If we can assume that job ads don’t/can’t in the vast majority of cases specify gender, then it’s odd and perhaps worrying that Google is doing the math for us. It seems unlikely that someone in Google sat down and came up with this as a neat strategy. But rather, it’s probably the Google ‘brain’ watching the kinds of jobs people of different gender apply for and preferring those kinds of jobs to that gender because they are more likely to earn that click. Google only gets paid when ads get clicked on, so it tries really hard to fit the best ad to what it knows about the user every time. That’s fine for fly fishing, but weird for jobs.

I suspect that this is simply the unintended consequence of the AI running the ad platform getting a little too good at its job. In society, we adopt behaviors and constructs to address what we perceive to be intrinsic societal problems. We make extra efforts to be inclusive; we go the extra mile. AIs don’t have the same PC conscience, unless it’s programmed into them. I imagine there will be a bunch of changes made in the near future to “level the playing field” on this topic. Ironically, it will probably reduce the overall effectiveness of Google’s ad product in some areas. But that might be a small price to pay to avoid the discrimination law suits.

Breaking the “Capacity Limit” in Fiber Optics

With the amount of internet-connected devices increasing every day, the need for better, faster Internet is significant. University researchers have come up with a very clever solution to this problem.

Engineers at the University of California, San Diego have broken the “capacity limit” for fiber optic transmission, paving the way for faster, longer and potentially cheaper fiber networks. Currently, there is a limit on the intensity of light you can send through a fiber optic link, which arises from the fact that when you increase the intensity of light through a fiber cable, noise, distortion and signal dilution increases. This is called the optical Kerr effect, and it causes problems for fiber network designers.

Eventually the Kerr effect becomes so great that, at high intensity levels, distortion entirely prevents an outbound signal from being correctly interpreted by a receiver. The issue gets worse the longer a cable is, so to increase the transmission rate through a fiber optic link by increasing the intensity level while keeping noise at bay, you have to include signal repeaters along the way.

This new breakthrough helps increase the capacity limit of high-bandwidth fiber optic cables by conditioning signals with “frequency combs”, allowing the receiver to predict any noise that is introduced during the transmission. The ability to predict noise patterns means the receiver can reconstruct the intended data from the noisy signal. Thus, the amount of power that can be sent through a fiber optic cable can be increased significantly without worrying about the effects of noise. Researchers were able to increase the signal power in a fiber cable by 20 fold and still get data at the end.

The use of frequency combs also means data can be sent along longer cables without the need for repeaters, which has the potential to reduce the cost of fiber networks. For example, engineers were able to send data through over 7,000 miles of fiber cable with standard amplifiers and no repeaters. With this breakthrough, hopefully we can see the capacity of fiber cables increased beyond what is currently possible in real-world applications.

Does Google Have a Color Problem?


The story this week that has been setting social media on fire is the one about the algorithmic miss classification of some photos by Google+.  The problem is that it classified a picture of a nice African American couple as “Gorillas.” That’s horrible, and what’s worse is that the only way Google could find to fix the immediate problem was to remove the tag entirely. Obviously, it was a mistake and Google apologized profusely.

Image recognition is much tougher than you might think. I read recently that at any given moment, about 65% of our brain processing power is focused on visual processing. We are extremely good at it, and we evolved that skill over millions of years. We recognize things and people with incredibly few visual clues. Expecting software to be as good as we are is probably unrealistic.

Having said that, this oops moment does perhaps point to a larger problem. Silicon Valley is horribly white. Actually it’s white and Asian, by which I mean mostly Indian Asian. There is a sprinkling of Latinos and about 2% African American. In short, Silicon Valley is a privileged white, first-generation, super-educated immigrant mix. I don’t recall ever seeing a black face in a meeting. Companies like Google promote diversity, but any visitor to their cafeteria (a fantastic place BTW) will be struck by the mass of pencil-necked geeks and button-down collar-wearing senior nerds and the phalanx of Indians (the Indian food there is first class BTW).

If all the diversity you have is white and green card, it’s less likely that anyone is even thinking about checking things like algorithms for culturally disastrous mistakes like this one. Nobody looks like them, so nobody checked. I guarantee that there are special routines built in to check for things like Native American vs. Indian sub-continent. I wouldn’t be surprised if there is special code to try to differentiate various Asian groups, but clearly not black.

It’s a problem without an easy fix. Graduation rates, etc. are lower in some communities, but companies like Google could invest in urban tech programs. Maybe fund a few charter schools in cities that aren’t San Francisco. At the very least, they could check modules like the one in question for obviously offensive situations like this one. If you live in a bubble, this kind of thing is likely to keep happening.

Google Accused… Again



There are some arguments that just won’t go away. The “Google is skewing their results” argument has been kicking around for several years, and a new version has just popped up. I’ve commented multiple times on this deathly dull topic, but let’s review for old time’s sake.

Google skews its results. It’s a simple fact. They have done it for years. For example, if you use SafeSearch, it will skew away from adult content. When you make a local search, it skews results to fit your location. Nobody complains about that kind of skewing; it’s the type where they favor their own results or partners in reference to a competition that causes concern.

The FTC investigated this a year or more back in the US and found no fault. It’s unlikely to revisit this challenge, in part because it’s tricky to prove and in part because Google has a monster lobbying engine and the tech jobs issue (much like guns) is a third rail issue. The equivalent of the FTC in Europe is looking at taking on Google. Those EU guys hate Google with a passion and are always on the lookout for a stick with which to beat them.

This latest stick comes from a study commissioned and paid for by Yelp. That in of itself probably makes it worth less than the paper it’s printed on, but it’s an interesting study nonetheless. Essentially, they showed Google search presentations with and without the Google “Focus on the User Listing” OneBox container. In their testing, they found that users clicked on the plain results rather than the optimized version 45% more frequently. Their conclusion was that Google is deliberately sabotaging the end user by making the results less easy to use, forcing users to search again or…what, give up in tears? For a complaint of this kind to prosper, you have to show harm, and their version of harm here is a lower click through rate.

They aren’t claiming that a Google OneBox business being preferred is actually benefitting Google financially. Rather, by featuring OneBox registered businesses, Google is harming the end user. The overall claim is at best paper thin.

There are several quite good reasons why Google often prefers OneBox results. To start with, OneBox businesses have taken the time to validate their business, confirm the info is correct and, in many cases, create a nice business profile page that the searcher may well click through to. Either way, what does seem to be clear is if this is annoying users, it’s doing so to the extent where they are changing their search engine. Most people have bigger fish to fry.

Search and how that search is presented (much like the weather) is a big, complicated topic. If you stare hard enough, you are bound to find things that might look a bit off. Keep staring and it will change. Complaining about it to an FTC which has been bought and sold by Google won’t get you very far. I imagine these researchers will be mailing their results to the EU commissioners right away.

Scan Your World, Then Live in It

Untitled-1As Virtual Reality inches close to an everyday reality, we need clever people coming up with the components to make it real and useful. One of those groups has just received a large chunk of venture funding, and I’m delighted.

These guys have come up with a clunky-looking camera, which can scan the interior of your house (or pretty much anything) quickly and cheaply. It’s fairly expensive to buy; at $4,500 per it’s not exactly discretionary income level. But it’s fast, and the subscription rates and monthly fees are really reasonable. They are predictably targeting the Realtor market, thinking that having a fully realized high definition virtual interior will make selling upscale property that much easier. Panoramic tours of property are not new, but fully navigable VR 3D at this price point and availability probably is.

unnamed (1)

It’s also a fascinating opportunity for museums and galleries that can offer an easy-to-navigate, high-quality virtual tour on any browser. The results are very easy to navigate and understand; if you can use Google Maps Street View, you can use this. It’s supposed to be compatible with VR headsets. I haven’t have the chance to experience that yet. I can imagine game developers must be salivating at the prospect of being able to digitize in 3D and physical location. If they can get access to that location one time, they can build an immersive experience around it and populate it with virtual people.

What’s exciting is that these guys, Oculus Rift, Google, GoPro and many more are coming out with VR implementations that make sense and are reasonably affordable. Non-individual deployment is the full solution, but each step takes us closer.

Wearable Meets Medical


We are all familiar with wearable tech like Fitbit that screams, “Look at me, I’m a jogger! Be impressed by my fitness!” But in reality, they aren’t much use as medical devices. Medical practice has used lab monitoring for many years, but the moment the patient leaves, so does most of the opportunity to monitor the treatment in a meaningful way. That might be changing. The Google X research division — home to Glass and WiFi by balloon — has announced plans for a wearable device designed to monitor patients with the kind of accuracy and reliability medicine would find useful.

It’s a brilliant idea that, in many ways, makes more sense than most of the consumer wearable applications littering the market right now. Initially, they are targeting clinical trials, but I could see this going much wider. Imagine a not-too-distant future where your doctor dispenses prescriptions along with the watch needed to track the efficacy of the treatment in real time. I have no idea which parameters could be tracked without compromising the skin, but we already track things like blood oxygen level and alcohol that way, so I imagine the opportunity is out there.

At first glance, this looks like something of a niche play, but if you factor in the enormous growth in conditions like diabetes and childhood asthma (both of which benefit from constant measurement of critical levels), I wouldn’t be surprised to see the early versions of this kind of tech seeing very rapid adoption. It’s good to see yet another technology imagined in Star Trek becoming reality. Now if only we could figure out matter transportation too.

The Subscription Game


I was very impressed to read today that Dollar Shave Club just raised an additional $75 million to support their growth. I love these guys with a passion. I have to buy razors and I only like the good ones, but paying a fortune for the Gillette product has gotten up my nose for as long as I can remember. There are two marketing triumphs of the past 20 years: Starbucks, who got us used to paying $5 for a cup of coffee and Gillette, who has convinced the American male to spend $7 a blade. Both brilliant moves.

The Dollar Shave guys (and their imitators, who now include Gillette) simply mail four really great blades to your house each month for just $9. It’s simple, it’s brilliant and I can now forget about buying blades. They are one of the latest successes on the subscription circus, and they won’t be the last.

As a keen podcast listener, I’m constantly assFullSizeRender (1)ailed by subscription offers. I’ve been a member of Audible for close to a decade, and I have about 1,000 audio books in my library. I joined Blue Apron a few months ago, and ever since then, I’ve been cooking pretty much Cordon Bleu meals for two to three nights a week. It includes every single ingredient and the recipe to make it happen for $10 per person, per meal. I have no idea how they do it for the price, but it’s amazing. It’s also much cheaper than shopping for two and watching produce I don’t need go off. Here’s a pic of the very tasty salmon cake burgers I made last night:

Some of these services save money, all of them save time to some extent or another, and some are just wildly convenient. It feels like this is a trend that is building steam. Online shopping, especially Amazon Prime (a subscription service) has seriously hurt shopping malls. They are closing by the thousands all over the nation. Online music wiped out all physical copies of music as we knew it, and live music became the dominant way artists make money. Even if all that these services do is free up time for us to binge watch Orange is the New Black on our Netflix subscription, that’s a good thing in my mind. I’m doing less of what I hate and more of what I like. I’m in. Next up: Trunk Club for buying clothes.

VR Fans: It’s Essay Time


It’s about now that I wish I’d spent time developing a strong YouTube presence rather than developing a company or having a life. Google is looking for YouTube stars to test out their new Jump rig, which is essentially 16 GoPro cameras bolted onto a lazy Susan with software to stitch the video together into a 360 experience. It’s not quite true VR, but it’s an interesting, relatively low budget way of moving the concept forward.

What Google’s doing, in true Google form, is asking end users like you or me to fill out an application form to be one of those early creators. There’s a bio section, a short questionnaire and the all-important essay section, where no doubt bearded hipsters and their female equivalents will try to pitch the perfect combination of approachable elitism and cuteness we have come to expect from things like Google commercials. There will be hundreds wanting to take the rig up in a plane and jump out; there will probably be virtual safaris and virtual tea ceremonies and kittens.

Since I don’t have nearly enough social juice to qualify, I’ll save my breath to cool my porridge (as my dear mother would say). But here’s what I’d apply for:

I’d put the rig on a shoulder-mounted harness and go visit places like Rio at peak Mardi Gras, or the Red Light district in Amsterdam or Bangkok. I’d strap a rig to a Seal team member doing ISIS interdiction. I’d cut a deal with a leading NASCAR or F1 team and follow an entire race season. Let’s put a rig on the pitcher’s mound at Yankee stadium. Sure, I’d base jump and water ski with it, but I’d also put it on the Dakar Rally and tour the Vatican and Machu Picchu. Let’s get a rig on the set of Game of Thrones and let’s see how that gets put together.

Then I’d make every adventure available online as they happen. Anything that gets VR moving into our mass consciousness has got to be a good thing. Let’s not waste cycles being cute and elitist; let’s get the people what they are actually interested in ASAP. A bread and circuses approach, which plugs right into the things which we already care about enough to buy season tickets and airfares to do will move the dial much quicker. Google, please let’s not be Silicon Hipsters about this. Let’s do something that people actually care about.

Using Your Tongue to “See”


A new device uses your tongue to help the blind become aware of their surroundings, and has recently been approved by the FDA.

The BrainPort V100 has three parts: a camera that has been attached to a pair of dark sunglasses that takes photos of what is in front of the person, a hand controller that receives the image data from the camera and then sends them on the stimulation device in the form of low intensity electrical stimuli.

The stimulation device is a square lollipop-like device that can produce up to 400 electrical stimulation points, aimed at the person’s tongue. On a normally developed person, this would feel like tactile stimuli on the tongue, but for those that are blind, it seems that the visual cortex can actually be stimulated.

The device does not make a blind person see actual visual images, but rather these stimulations on the tongue help the blind interpret what object is in front of them. This can offer the blind some independence, because they can learn how certain object feel on their tongue and become aware of what is present in front of them.

The device can help blind people tell the difference between certain objects that they have in front of them, such as a banana and a ball. Furthermore, it could even enable them to read some small words, such as EXIT, that bares great importance in their independence.

The visually impaired can practice with the BrainPort device in specialized areas, so that they can learn how it works and how to use it. One type of practice rink is an oval track which they can learn to walk in, without assistance. By practicing, the blind can identify certain patterns in their usual routes much better than using a walking stick.

The initial price of the BrainPort V100 is steep, and will be $10,000. Hopefully, as the technology is improved and its use becomes more widespread, the price will become more affordable.