Archives

Google

Big Tech drastically reverses course on coronavirus ad policies

Big Tech drastically reverses course on coronavirus ad policies

When tech companies roll out new policies, it’s usually after long periods of meticulous internal discussion and review. Big reversals aren’t the norm. 

Then Covid-19 hit.

As the looming threat of the coronavirus began hitting the U.S. in February, many tech companies started to roll out new ad policies.  

Google banned most nongovernmental Covid-19 advertising in early February. YouTube followed suit shortly after, demonetizing videos about the coronavirus so its advertisers wouldn’t see their ads placed on unpleasant coronavirus content. 

Both companies, owned by Alphabet, classified the coronavirus as a “sensitive topic,” which is typically reserved for things like school shootings or natural disasters — short-term events that bad actors could capitalize on to earn money from Google’s advertising partners. Read more…

More about Twitter, Google, Youtube, Amazon, and Coronavirus

Google research makes for an effortless robotic dog trot

As capable as robots are, the original animals after which they tend to be designed are always much, much better. That’s partly because it’s difficult to learn how to walk like a dog directly from a dog — but this research from Google’s AI labs make it considerably easier.

The goal of this research, a collaboration with UC Berkeley, was to find a way to efficiently and automatically transfer “agile behaviors” like a light-footed trot or spin from their source (a good dog) to a quadrupedal robot. This sort of thing has been done before, but as the researchers’ blog post points out, the established training process can often “require a great deal of expert insight, and often involves a lengthy reward tuning process for each desired skill.”

That doesn’t scale well, naturally, but that manual tuning is necessary to make sure the animal’s movements are approximated well by the robot. Even a very doglike robot isn’t actually a dog, and the way a dog moves may not be exactly the way the robot should, leading the latter to fall down, lock up, or otherwise fail.

The Google AI project addresses this by adding a bit of controlled chaos to the normal order of things. Ordinarily, the dog’s motions would be captured and key points like feet and joints would be carefully tracked. These points would be approximated to the robot’s in a digital simulation where a virtual version of the robot attempts to imitate the motions of the dog with its own, learning as it goes.

So far, so good, but the real problem comes when you try to use the results of that simulation to control an actual robot. The real world isn’t a 2D plane with idealized friction rules and all that. Unfortunately, that means that uncorrected simulation-based gaits tend to walk a robot right into the ground.

To prevent this, the researchers introduced an element of randomness to the physical parameters used in the simulation, making the virtual robot weigh more, or have weaker motors, or experience greater friction with the ground. This made the machine learning model describing how to walk have to account for all kinds of small variances and the complications they create down the line — and how to counteract them.

Learning to accommodate for that randomness made the learned walking method far more robust in the real world, leading to a passable imitation of the target dog walk, and even more complicated moves like turns and spins, without any manual intervention and only little extra virtual training.

Naturally manual tweaking could still be added to the mix if desired, but as it stands this is a large improvement over what could previously be done totally automatically.

In another research project described in the same post, another set of researchers describe a robot teaching itself to walk on its own, but imbued with the intelligence to avoid walking outside its designated area and to pick itself up when it falls. With those basic skills baked in, the robot was able to amble around its training area continuously with no human intervention, learning quite respectable locomotion skills.

The paper on learning agile behaviors from animals can be read here, while the one on robots learning to walk on their own (a collaboration with Berkeley and the Georgia Institute of Technology) is here.

Want to survive the downturn? Better build a platform

When you look at the most successful companies in the world, they are almost never just one simple service. Instead, they offer a platform with a range of services and an ability to connect to it to allow external partners and developers to extend the base functionality that the company provides.

Aspiring to be a platform and actually succeeding at building one are not the same. While every startup probably sees themselves as becoming a platform play eventually, the fact is it’s hard to build one. But if you can succeed and your set of services become an integral part of a given business workflow, your company could become bigger and more successful than even the most optimistic founder ever imagined.

Look at the biggest tech companies in the world from Microsoft to Oracle to Facebook to Google and Amazon. All of them offer a rich complex platform of services. All of them provide a way for third parties to plug in and take advantage of them in some way, even if it’s by using the company’s sheer popularity to advertise.

Michael A. Cusumano, David B. Yoffie, and Annabelle Gawer, who wrote the book The Business of Platforms, wrote an article recently in MIT Sloan Review on The Future of Platforms, saying that simply becoming a platform doesn’t guarantee success for a startup.

“Because, like all companies, platforms must ultimately perform better than their competitors. In addition, to survive long-term, platforms must also be politically and socially viable, or they risk being crushed by government regulation or social opposition, as well as potentially massive debt obligations,” they wrote.

In other words, it’s not cheap or easy to build a successful platform, but the rewards are vast. As Cusmano, Yoffe and Gawer point out their studies have found, “…Platform companies achieved their sales with half the number of employees [of successful non-platform companies]. Moreover, platform companies were twice as profitable, were growing twice as fast, and were more than twice as valuable as their conventional counterparts.”

From an enterprise perspective, look at a company like Salesforce . The company learned long ago that it couldn’t possibly build every permutation of customer requirements with a relatively small team of engineers (especially early on), so it started to build hooks into the platform it had built to allow customers and consultants to customize it to meet the needs of individual organizations.

Eventually Salesforce built APIs, then it built a whole set of development tools, and built a marketplace to share these add-ons. Some startups like FinancialForce, Vlocity and Veeva have built whole companies on top of Salesforce.

Rory O’Driscoll, a partner at Scale Venture Partners, speaking at a venture capitalist panel at BoxWorks in 2014 said that many startups aspire to be platforms, but it’s harder than it looks. “You don’t make a platform. Third party developers only engage when you achieve a critical mass of users. You have to do something else and then become a platform. You don’t come fully formed as a platform,” he said at the time.

If you’re thinking, how you could possibly start a company like that in the middle of a massive economic crisis, consider that Microsoft launched in 1975 in the middle of recession. Google and Salesforce both launched in the late 1990s, just ahead of the dot-com crash and Facebook launched in 2004, four years before the massive downturn in 2008. All went on to become tremendously successful companies

That success often requires massive spending and sales and marketing burn, but when it works the rewards are enormous. Just don’t expect that it’s an easy path to success.