Startcrowd: a Blockchain network for Artificial Intelligence projects
Startcrowd is a social network facilitating AI projects. On Startcrowd, requesters can post their projects, and other participants can complete them. In a future iteration of Startcrowd, contributors will be able to get paid with a blockchain token. The Startcrowd product is already launched (without the blockchain feature).
Current projects cover various applications of artificial intelligence, like smart agriculture, green energy, or drug discovery. There are also projects in fundamental research, combining Boltzmann machines and ergodic theory. More examples can be found on the Startcrowd website.
The AI for drug discovery project is the most active, with one research paper. Drug discovery is a popular theme on crowdsourcing platforms (Innocentive, Open-Source Drug Discovery, Dream Challenges…), and it’s also an important topic on Startcrowd.
Types of projects: open-source or proprietary
Open-source is the default posting mode. Keeping a project open-source allows to maximize visibility, and minimize the barrier of entry for potential contributors.
Proprietary projects are also possible on Startcrowd. This option can be chosen by entrepreneurs willing to build Intellectual Property. In this case, requesters can select contributors for their private project, and they can make them sign Non-Disclosure Agreements.
It might seem hazardous to crowdsource proprietary innovation to strangers around the world. However, with the help of a well-executed rating system, this practice will certainly become widespread, like sleeping at an AirBnB home, or jumping into an Uber car.
It might be good to recall that an open-source project is not necessarily unpaid: sometimes, it can be sponsored. Likewise, a proprietary project is not always paid: labor can often be performed in exchange of the mere hope for a future job or equity. That’s how unpaid internships are functioning.
How participants will earn Startcrowd tokens
Tokens can bring incentives to diverse tasks, including:
Building data science models, and the software around them (front-end…)
Selling data for training models
Powering computations of data scientists
Staking to secure the proof-of-stake Startcrowd blockchain
Providing consulting services. Experts can help define the problem, collect the data, price the project, or settle Intellectual Property disputes. A new class of consulting services will emerge for Startcrowd, akin to Search Engine Optimization for Google, or Social Media Marketing for Facebook.
Why blockchain?
Using a blockchain allows the full decentralization of the network. Traditional crowdsourcing platforms often suffer from friction between a decentralized crowd and a centralized platform.
A token also facilitates shared ownership. This should reduce the risk of building an exploitative platform, which can always happen for uberization projects.
Moreover, blockchain is a better fit for access to investment capital. A decentralized project is more comfortable with decentralized investors. People building Startcrowd already span on 3 continents (Africa, America and Europe), and we are welcoming contributors wherever they want to be. Traditional Venture Capitalists do not appreciate this decentralized organization. They prefer to keep founders at their hand, and sometimes do not even look beyond 20 miles from the Stanford campus. For example, it is notorious that Y Combinator expects founders to move to San Francisco Bay Area.
On the other hand, blockchain is empowering a new generation of crypto-investors, unencumbered by any specific geography or regulation. For non-accredited investors, or in countries with depressed local economies like Nigeria or Venezuela, blockchain is the only access to lucrative tech investments.
The specific form of the token is not decided yet. It might be built on top of ethereum, on top of an Ethereum competitor (NEO, EOS, Cardano…), or be a completely independent blockchain. This topic remains to be discussed by potential users and crypto-investors.
Why now?
There is a growing demand for data science services. It is predicted that demand for data scientists will soar 28% from 2017 to 2020.
On the supply side, online courses are training a new generation of data scientists.
However, this new supply and demand don’t always match. It’s hard to find a job after having completed online courses. Students have to overcome a lot of barriers, for which they were not prepared. They have to apply for a job, relocate where the job is, and then commute to the office every morning.
Startcrowd vision is that jobs must come to talent, instead of the other way round. With a platform like Startcrowd, online learners can find a familiar setting, in which they already excelled. They can make the most of their online education, during which they developed their initiative, self-learning, and self-organization capabilities.
Moreover, online courses remain pretty basic, they rarely go beyond the undergraduate level. After completing these introductory courses, students often ask 'what next’, and they remain clueless. On Startcrowd, they can find research-level projects, and receive an online graduate education.
To get a feel of this job market inefficiency, I launched groups on Facebook and LinkedIn for remote data scientists jobs, and candidates outnumber job postings by several orders of magnitude. So talent shortage in data science may be just a myth.
Startcrowd vs. competitors
A way to better understand the Startcrowd product is to make comparisons with some competitors.
Startcrowd is organized differently from platforms like Freelancer and Upwork. On freelancing platforms, freelancers are not allowed to communicate with each other. They can only communicate with their potential customers. That’s why freelancing often remains confined to small tasks, which can be completed by a single individual. On the other hand, Startcrowd is a full-blown social network, and startcrowders are encouraged to communicate with each other. Open-source projects invite them to do so. They can build teams, and tackle more ambitious projects.
Moreover, user experience is often poor on those freelancing platforms. Good projects and good freelancers run away from them, resulting in a market for lemons. Startcrowd has the potential to better execute, by being laser-focused on the AI and data science niche.
The first difference is the user interface: Github is made mostly for coders, with a user interface less visual than Startcrowd. A suitable interface is important for different crowds to gather and exchange different skills and tokens.
The second difference is that Github does not seem interested in building a marketplace around repositories. Projects are either free and public, or they live on private repositories invisible to the outside world (Github Enterprise). In all cases, there are no transactions occuring on the Github platform. There are people making money by contributing to Github, but they are doing so out of Github. There’s currently no platform successfully combining Freelancer and Github, and Startcrowd is trying to fill this gap.
Kaggle is mostly known for competitions, which attract lots of participants. On Kaggle, only the staff can organize challenges, and there are relatively few of them, compared with the number users. Startcrowd is different, because challenge organization is also crowdsourced. Everyone can post a challenge. This should reduce the number of participants per challenge, and so reduce talent waste. On Kaggle, only a small fraction of submissions contribute to the final solution, the rest is useless.
Moreover, Kaggle competitions are exclusively focused on model building. Startcrowd is involved in the whole product pipeline around data, in order to deliver a sellable product. As a result, it caters to a wider crowd, like product designers and growth hackers.
In a previous post, I elaborated on the differences between Kaggle and Startcrowd. In short, Kaggle is just a game, while Startcrowd aims at being more productive.
Of course, there are many more competitors, I made a longer list here.
Conclusion
A Startcrowd MVP is already launched, where everyone can post and contribute to AI projects. To get the full Startcrowd platform, a blockchain feature remains to be added. This will be done when enough users and crypto-investors will be involved. Startcrowd tokens can be pre-ordered here.
Questions can be asked in the comment section of this post, or on the Startcrowd chat.
