Why Labor as a Service is as Cloudy as it Gets


When thinking about cloud computing, it’s easy to get caught up in the technological innovations. Often overlooked, however, is the tool that made all this high technology possible: the human brain. What if we could access that resource via the cloud, too? As I detail in my weekly GigaOM Pro column, we can. Some call it “labor as a service” (I like LaaS), others call it “labor-on-demand,” but everyone should call it cloud computing.

Perhaps invented by Amazon Web Services (s AMZN?) in the form of its Mechanical Turk offering, the LaaS market is now taking off. Startup CloudCrowd this week announced $5.1 million in Series B funding for its service, while fellow startup CrowdFlower was the subject of an in-depth interview with O’Reilly Media. The spectrum of use cases for solutions is broad, ranging from book reviews to editing to software testing, and beyond. If something requires human judgment that can’t be replicated in an algorithm, it’s a prime candidate for LaaS.

Cloud computing is more about a set of capabilities than it is about any specific technology set, so who’s to say silicon has to do the actual computing? Developers turn to traditional infrastructure as a service because they can get CPU resources when they need them, for as long as they need them, and without having to go through the expense and bureaucracy of purchasing, installing and managing physical resources.

LaaS does the exact same thing with employees (only application development is far easier: “Do this.”). Via GUI, submission form and/or API, customers designate how many workers they need, for what task and for how long, and pay accordingly. The LaaS provider handles everything else: finding, training, evaluating and paying the workers, then delivering the results to the customer.

Ultimately, I think models like LaaS will force us to look at cloud computing far beyond the current scope of IaaS, PaaS and SaaS. We’re seeing services take hold now that don’t fit nicely into the definitions we’ve ascribed to these terms, but certainly fall under the cloud computing umbrella. I’d love to hear what you think about this topic: What other types of emerging services warrant their own “aaS” acronym, and how far can we expand the definition of cloud computing before it breaks?

Read the full post here.

Image Source: Flickr user James Cridland



Pingsta has been doing this for the network engineering since 2008 under the tagline of expertise-in-the-cloud and intelligence-as-a-service. Pingsta’s work anywhere, anytime through any media vision is the future of work.

a Pingsta ICE member

David Cowan

I liked the idea so much I invested in CrowdFlower. Founder Lukas Biewald blew me away in our very first meeting, during which he ran a sentiment analysis for me on the last 500 tweets about BillShrink, completing it with 96% confidence per item before the pitch was done. (It cost me $13.)


It was an interesting read ..

When you look at the Managed service business ( recruiters r business development managers ) currently providing service on demand. It lends itself to LAAS…..


Imran Anwar

I am sorry, but I don’t understand how you can call that cloud computing. The idea of labor on demand is compelling, but I do not think one can quite compare that as a functional model to cloud computing/IT. In the IT world we can slice jobs, or processes, or datasets, based on additional CPU/Memory/storage requirements and distribute workloads or hire more fairly standardized units of capability/capacity.

How does that work in anything remotely creative or judgement driven? One can’t expect a great novel to be produced by hiring 100 authors online. I wouldn’t even expect basic functionality apps to get created that way. Perhaps you have better examples that can help shed light on how “LaaS” or “LoD” is “Cloud Computing”. Thanks.

Derrick Harris

I think David’s example below provides an exemplary example of the cloud connection. Just like a company might launch a temporary web site on EC2 or run a quick batch analysis on some data, someone might use LaaS to carry out an ad hoc task that requires human judgment. The real similarity is in the model — it’s fast, easy, cheap and doesn’t require going to HR (or, in the case of servers, IT). Essentially, the only difference is brain cycles vs. CPU cycles.

You wouldn’t use it to create anything, but rather to analyze in ways that the vast, vast majority of algorithms cannot. Or to edit, debug, cleanse data, etc.

Comments are closed.