A Look at How AI Has Changed Reddit

Hey everyone, I'm John Coogan and today I want to tell you an interesting story about Reddit. The company has a really interesting founding story, got acquired early in pretty unusual circumstances, and now occupies an odd place in the landscape of social media sites. I think the lessons from the Reddit founding story are useful to any entrepreneur and the way the site is changing in the age of artificial intelligence is incredible. Let's dive into it. Reddit was founded in 2005 by Steve Huffman and Alexis Ohanian, who were college roommates at the time. After attending a lecture by Paul Graham, they were invited to apply to the first ever batch of Y Combinator. Their first idea failed, but they quickly pivoted to creating the front page of the Internet. The site really hasn't changed all that much since launching. The core idea remains the same. Users upload links with titles and comments, and then other Redditors can upvote or downvote posts based on what they like. Reddit is a great case study in growth, hacking and early stage social product. Even though they were able to build a functional application, there was no content, so there was still no reason to visit the site. They needed a way to kickstart the cycle of posting and voting on links, so they did it all themselves. Here's Steve Huffman explaining exactly how that worked.

But we did a lot of very important things that I think helped, one of which was in the beginning, Alexis and I submitted all of the content. If you go to Reddit's submission page now, it basically boils down to two fields, and in the early days it was just two fields. You basically had a field called URL, which was the URL of the site you were submitting, and title, which was the title of that site, and that would be the link you submitted to Reddit. If Alexis and I went to the page, we had a third field called user, and we could type in a name here, and then when we hit submit, we would automatically register this user if that name didn't exist and submit the content as that user. So you would go to Reddit in the early days, the first couple of months, and there'd be tons of content. It was, I shouldn't say fake content, but it was fake users. It was really all just Alexis and I. And that did two things, and I think this is important to any site where you want to have a community feel. The first thing it did was it set the tone. Websites have this kind of inertia and we submitted content that we would be interested in seeing and that meant the content on Reddit, at least for us and for our kind of peer group, it was good, interesting stuff. We wanted a site with the most interesting content online, and so we did our best to find it, and then we submitted that content ourselves. So when you show up to Reddit, you kind of know what the site's all about. The other thing it did is it made the site feel alive. Users like to feel a part of something. If they showed up to the website and the front page was blank, it just looks like a ghost town. You don't want to live in a ghost town. You'd rather live in New York City. Most people would. That's why New York City is so big and ghost towns, by definition aren't. That was really fortunate, and I don't know what made us do this. In hindsight, it's really easy to say, oh, this is how it happened. But at the time, I think we were just embarrassed to have an empty website. So we submitted the content and it worked. And I remember the first day a few months in when we didn't have to submit any content, and that was such a magical day for us because then we knew the site was working on its own. It was really, really cool.

I think this is a great lesson for early stage startups. It can be very depressing to release a product that you've spent weeks or months building, only to see it seemingly flopping. No matter what type of product you're building, it's critical that you figure out how you will attract early adopters. You can grow the product through a beta program and leverage your friends and personal networks, but sometimes you might need to do some of the work yourself in order to make sure you're providing value to your users on day one. The second interesting thing about Reddit is that even though it's one of the most trafficked sites in the world today, they actually sold the business to Conde NAST for $10 million just one year after launching. In today's startup environment, that seems crazy. Social networks like Clubhouse have raised valuations 10 times that while still in private beta. But the world was very different back in 2006. Facebook was still small, and it would be another five years until LinkedIn went public. A few days ago, Alexis posted a heartfelt thread on Twitter about his decision to sell the company so early. It's a really great read, so I'll link it below. And it really drives home how hard it is to say no to $10 million when you're 24 years old and have family health issues to deal with on top of the stre of running a growing business. Fortunately, current funding markets make outcomes like this much easier to avoid. If you're growing a business rapidly these days, you'll have VCs beating down your doors to lead your Series A. In 2014, after five years away from the company, the founders returned to leadership roles at Reddit. It's extremely rare for a company to be acquired and then spun out as an independent business, and it's even more rare for founders to rejoin after leaving. But Reddit is an unusual company, and it's clear that the founders have a unique vision for the site and are really the only people who can lead the company successfully. The history of Reddit is interesting, but the way the content of the site has grown is perhaps even more important because the company is independent and still closely tied to the Y Combinator community. Reddit was one of the main training sets for OpenAI. Sam Altman, the CEO of OpenAI, was actually in the same Y Combinator class with the Reddit founders and and they also work together at YC as investors as well. OpenAI took 2 billion comments from Reddit's database and used the data to train a machine learning model that could produce truly remarkable results. The resulting model, called GPT3, is capable of answering questions, translating languages, summarizing complex articles, and much more. What I find most interesting is that Redditors are now using GPT3, which was trained on Reddit comments to produce new content for Reddit. Although OpenAI has several guardrails in place to prevent abuses of the system, people still seem to be able to take GPT3 outputs and use them to farm Reddit karma. I think these looping cycles of events are really interesting. Reddit's founders start by creating fake content for the site, then leave the company, only to return and share the Reddit data with OpenAI, which results in fake content returning to the site. This isn't a criticism of the company by any means. I think it's awesome that the Reddit founders were able to return and help grow the company. And I also think these recent developments in artificial intelligence are absolutely remarkable. The question is what the future of AI on Reddit looks like. Currently, the community is actively trying to shut down accounts that use AI to generate content, but what if they fail? What if AI is better suited for collecting, aggregating, and disseminating information on Reddit? Would that be a better or worse experience for the average user? I suspect this would be deeply disturbing to most people the whole reason we log on to social media sites is to engage with other humans, and if it's increasingly difficult to tell who's human and who's a robot, people might move on to other channels that are harder for bots to infiltrate. This bodes well for networks like ClubHouse, Twitch, and YouTube after all. I doubt any of you watching this have any suspicions that this video could be generated by an AI, but maybe it's just a matter of time. If you enjoyed this video, please subscribe to the channel. And if you have any thoughts about Reddit or AI, leave them below. I always respond to the comments. Thanks for watching.