The Improbability of Wikipedia.

We take it for granted today, but Wikipedia is the most ridiculous idea to have succeeded and represents the best of what humanity has to offer.

Every so often, I visit this Wikipedia page and marvel at it. The level of detail, from the breakdown of the idea behind the laws and the table covering each US state to the overview of current thoughts on the issue, is lovely and informative. I’ll grant you that it’s not a particularly remarkable page and is full of dry information. But it means a great deal to me because when I was in high school, I had to write a paper regarding Shield Laws in the United States. I was frustrated at having to spend one of my weekends working to finish it because I was a horrible procrastinator (and still am). I spent my weekend in a library researching the topic via physical books like some wild animal. In its early stages, Wikipedia did not possess this information while struggling to gain critical mass. While waiting to print my freshly written paper, I created a new page on Wikipedia and pasted the first sentence from my report.

I walked away, having forgotten about it for years. Looking at the organic growth that has happened since then fills me with wonder.

Let’s pretend we’re taking a trip back to the Internet of 2001 to convince a room full of venture capitalists to throw their money at an online encyclopedia concept. It’s called Wikipedia, a free, online, multi-lingual encyclopedia, and its content? Created, edited, and verified by the general public. Everyone and anyone with an internet connection and a browser.

Yes, that’s right. Anonymous people on the internet. Your Aunt Sally who thinks the microwave is spying on her could contribute. That guy on the bus who once tried to convince you that the birds aren’t real? He can chip in too.

You’d be laughed out of the room before you even got to the part about how it wouldn’t carry any advertisements and would survive purely on donations.

And yet, here we are in 2023, more than two decades later, and Wikipedia hasn’t merely survived; it has flourished. It’s the repository of human knowledge, the go-to for middle schoolers doing their homework, the cheat sheet for trivia buffs, and the quiet cornerstone of countless arguments settled at bars every night around the world.

The concept of Wikipedia operates at the edge of improbability and utter chaos. In the grand symphony of the internet, Wikipedia feels like the improvisational jazz number that should have spiraled into discord yet inexplicably came together in a fluid and cohesive masterpiece as a wild success.

There’s a potent lesson in Wikipedia’s success. The improbable happens when we let go of the rigid norms that constrain our imagination, opening ourselves up to the potential of distributed intelligence and the anarchic but productive power of the crowd. It lacks the structure and safety of a proven business model; it dances with the absurdity of relying on sheer altruism for its survival and revels in the chaos of crowdsourced intelligence.

Wikipedia has turned that idea of chaos into a radical democratization of knowledge. The contributors, a motley crew of curious individuals armed with knowledge and internet connections, play the role of both the orchestra and the conductor. It’s quite possible that Aunt Sally’s post about the history of microwaves is fact-checked by a physics professor halfway across the world. The cacophony of voices has not resulted in a deafening noise but instead a dynamic and evolving melody.

The idea of decentralization and the wisdom of the crowd is nothing new. We see it in the financial markets, where thousands of individuals participate in price discovery, their collective decisions often predicting outcomes better than the experts. It’s in a ‘free market of ideas’ where better thoughts and concepts should, in theory, rise to the top.

But Wikipedia brings its own spin to this. It’s not just an information market but an information commune, where those with knowledge freely give to those without. It’s as much about collective action as it is about collective wisdom and the urge to share knowledge as it is about seeking it.

The future, if the success of Wikipedia is any indication, will further question our traditional understandings of authority and expertise. In comes the looming future of artificial intelligence (AI) and open-source large language models – a conversation that feels as perplexing and ridiculous as that room of venture capitalists in 2001 might have found the concept of Wikipedia. Imagine walking back into that room, this time pitching not an open-source encyclopedia but a colossal open-source AI model, capable of understanding and generating human language with uncanny precision, accessible to anyone with an internet connection.

“A tool,” you might say, “that learns from the collective text of the internet and can generate human-like text based on that learning. An AI that can write essays, code, poems, even jokes. It could augment human capabilities, aid in research, education, and so much more.”

Eyebrows raise. Doubts cloud faces. The venture capitalists lean in, clearly intrigued but equally skeptical. “And who will train and control this AI?” one of them asks.

“Everyone,” you say. “It’s open-source. Anyone with the time and inclination can contribute, shaping the model, refining its abilities.”

Pandemonium breaks out. Cries of “Impossible!”, “Too risky!” and “Utter chaos!” fill the room.

Like Wikipedia, the idea of open-source AI and large language models rides on the thin line between chaotic madness and ingenious breakthrough. The concerns are real, from the danger of misuse to questions of quality control and accuracy. Here’s where the Wikipedia model whispers a gentle counterpoint: in the collective chaos, a symphony might be waiting to be written. The stakes are higher, and the challenges are more complex. Yet, the potential is breathtaking.

The venture capitalist room of 2001 might have laughed at Wikipedia. The room of 2023 may well balk at the idea of open-source AI. However, two decades from now, we might be sitting here, marveling at how a collective of anonymous internet users trained an AI model that not only understands human language but also respects its nuances and complexities.

AI introduces a sense of immediacy and personalization. Unlike a static Wikipedia page, AI can interact, clarify, and tailor responses to individual queries. It does not simply regurgitate knowledge like a search engine. It reframes, re-contextualizes, and responds in a way that aligns with the user’s query.

As we enter this brave new world, it is essential to remember the lessons of Wikipedia. Its success came not solely from the information it housed but from its principles: decentralization, collaboration, transparency, and adaptability. As we design AI systems, we must encode these principles into their DNA.

We must guard against the concentration of power and knowledge. AI models should not belong to only a handful of tech giants.

In an era of information overload, one should not underestimate the power of collective wisdom in shaping our AI tools. Here, in this space of radical transparency, unpredictability, and open access, we might discover a new symphony of intellect within open-source AI.

It might sound as chaotic and improbable as Wikipedia two decades ago. If there’s one thing Wikipedia has shown us, it’s that sometimes, chaos and improbability are what innovation needs to thrive.

In a sense, Wikipedia’s existence is less a triumph of technology and more a triumph of humanity. Suppose a website populated by the general public can navigate the swirling currents of the information age and come out on top. What other improbable successes might we overlook with our casual acceptance of AI models that are locked behind a handful of tech oligarchs?

Wikipedia’s story is a beautiful reminder that sometimes, the most powerful ideas are also the ones that seem the most ridiculous and dangerous. It’s not about the improbability of Wikipedia; it’s about the improbability of us. The future is as unwritten as that Wikipedia page on Shield Laws started many years ago.

I want us to write the future together.