Wikipedia is one of the world's greatest examples of human collaboration. But it's not without its flaws. By giving contributors ownership of their contributions, Golden lets them profit from their insight and wants to offer knowledge on every topic that world needs. Ekin Genç speaks to Jude Gomila about how he's building it.
Our online experience would be severely lacking without the ability to quickly look up information on Wikipedia. With over 6 million articles comprising 4 billion words on the English version alone, the platform is the best way to get an overview on almost anything. It also features many components of the web3 ethos: it is community-driven, highly collaborative, and debates are out in the open for everyone to see, and indeed, participate in.
But it’s not without its problems. The world’s encyclopedia is known for a gender bias against articles about women, and though wildly successful in sourcing information for free, leaves some important areas untapped, like emerging technologies, new organisations, or particularly niche or complex studies. There just aren’t many people who can write these articles, and those who can are often too busy to do so.
Indeed, Wikipedia is refreshingly transparent about its limitations. But beyond transparency, Wikipedia lacks other web3 components that reward contributions, help accrue value back to its user base, and could increase its value further,
Jude Gomila thinks there is a better model. Previously the co-founder of Heyzap, a mobile advertising platform that was acquired for $45m in 2016, Jude's answer is Golden, the decentralised knowledge protocol he founded the following year. He believes a blockchain-native and protocol-first design could improve content by introducing new data models, greater citation depth, and more sophisticated validation processes. Ultimately, Golden represents his thesis that Wikipedia would be even better if there was a monetary mechanism to incentivise entries, allocating research skills and expertise into neglected areas.
“The world is lacking a decentralised graph of canonical knowledge,” the company's website explains, adding that what is needed is one “that is open, free, permissionless, and incentivises agents to enter data into the graph.” Building exactly this is Golden's manifesto.
It’s a project that has already won custom from some of the most world's most established organisations. Formed in 1836, the U.S. Patent and Trademark Office is one of them, relying on Golden for 15 years of patent entries – 74,321 data items, specifically.
“Golden goes beyond 'web3 Wikipedia’ but it's a reasonable analogy,” Jude tells me. On Golden, contributors can earn protocol tokens and revenue share rights when the data is used commercially (for example, as part of an advertisement or technology development). Companies can pay (with fiat) for tools that help them access Golden’s information in more powerful ways.
Golden uses a burn-and-mint model that lets companies buy Golden tokens (with fiat) and convert them into credits that can be spent retrieving data from the platform at scale. Companies never have to think about any aspect of blockchain, including cryptocurrencies or digital wallets – normally a barrier to entry, particularly for enterprise clients.
Unlike Wikipedia, Golden has no notability requirement. This means that projects, people, and events which do not meet high standards for inclusion can be included, justified on the notion that these things are still part of human knowledge and experience. And unlike Wikipedia, Golden’s data – once accepted as true by validators – is immutably stored onchain: Wikipedia could be deleted, Golden can't be. Jude's platform also adds new ways of displaying the data, such as through timelines and how it connects to other topics, whilst debates about the topic and contributors to it are easily viewable.
With its core function of leveraging and compensating distributed contributions, Golden seems like an obvious blockchain use case, but that is not how things started in the first place. Instead, Golden started as a web2 experiment in Silicon Valley in 2017, raising $20m from top investors like Andreessen Horowitz, Balaji Srinivasan, and founders from Square, Scribd, and Airtable by 2020.
Things took a radical shift for Golden when, in the autumn of 2021, Golden identified blockchain as a tool that could help the company achieve its vision. As a result of the web3 shift, “a fair amount has changed in terms of approach to building the product,” says Jude. “For 3 years, Golden was building the web2 attempt at superseding Wikipedia and only a small community was around to build it [in the hundreds],” he tweeted in May 2021. “We switched to [a] web3 protocol approach and there are now 12,000 people (and growing) entering data into the system as we speak.”
Today, Golden has a sizable web3 community that provides frequent feedback and contributions, allowing for rapid iteration. “[Our] timelines now have also external pressure to move very quickly, as the community have great expectations for us,” he explains, opening the door on just how much had to change to underlie his company's web3 pivot. “We have had to upskill in web3, hire in, and tune our thinking in architectural design. Security has also become a premium drive.” They raised a further $40m last October to see the shift through.
Knowledge graphs may sound like an alien concept to most people. But in practice, they are not too complex. Wikipedia and Golden are libraries of millions of articles, and every concept contained in each article is a tag to yet another article about that concept. For example, Van Gogh’s Wikipedia article has tags to ‘Western Art’, ‘oil paintings’, and ‘sunflowers’, each linking to an article explaining that topic. If you imagine how you might draw this ecosystem, you would draw millions of articles, each connecting to each other, to form a knowledge graph.
This is all a knowledge graph is: a network of how different concepts relate to each other. Van Gogh’s page on Golden illustrates some of these relationships: a person, a painter, who died at Auberge Ravoux. Knowledge graphs are created from these ‘triples’: [Van Gogh] [is] [a painter], or [Van Gogh] [painted] [Sunflowers].
But who checks the validity of these triples? What stops someone from adding [Van Gogh] [is country] [Africa], which is patently incorrect? On Wikipedia, anybody can make an edit, though some edits may not be accepted by an editor. With claims stored immutably onchain, Golden aims to be more accurate from the very start, with a network of multiple editors (known as validators) who verify the accuracy of triples by casting votes before the edit can be displayed to others. When approved, the contributed triple is accepted as a fact and stored onchain forever.
The core mechanism through which Golden aims to incentivise contributions is though NFTs. Contributors have a form of ownership of the triples that they supply – or other assets, like images or references. These contributions will be free for browsing consumers, but contributors will profit whenever their triple is retrieved by Golden's corporate clients.
“The world is lacking a decentralised graph of knowledge that is open, free, permissionless, and incentivises agents to enter data.”
— Golden's manifesto
The model bears similarities to the Wikimedia Foundation’s commercial arm, which has raised concerns in the Wikipedia community regarding a centralised bureaucracy profiting from volunteer labour. It's a situation that Golden plans to avoid with NFTs underpinning a recurring royalty structure to compensate contributors when commercial enterprises access their contributions.
Several attempts at Wikipedia alternatives have been made, of course, many of which have leant into similar blockchain tools. The most notable is Everipedia, though the for-profit company has struggled to gain traction. Unlike Everipedia, Golden's protocol-first approach opens up contributing and validating to everyone, with contributors earning tokens to govern and share the profits of the platform.
“Protocol governance ... allows for less shadowy decisions,” Jude explains, arguing that “there seems to be scope for higher longevity and precision in web3.” His company is a bet on realising that opportunity, ultimately creating a community-owned encyclopedia economy, and a complete map of human knowledge.
The board of ChatGPT may be firing its CEO, but that doesn't mean this revolution is slowing down. A whistle-stop tour from the dawn of time with Stanley Kubrick, to Gutenberg in 1453, and up to now, we look at the economics of innovation. With AI are we now facing catastrophe or a creative surge? Here's our bullet-point summary of the podcast between David McWilliams and Eric Lonergan.
When a Harvard Senior Fellow told Mark Fielding that every university dean was a fan of the metaverse, he had to find out how they're using extended reality to reinvent university education. What he learnt was how classrooms without walls can transform the university experience.
The metaverse is the future of the internet and the brands that don’t embrace it will be left behind. Providing experiences and opportunities that are more unique and exciting than those in the real world, it is unsurprising that major brands are getting involved.