Instagram’s co-founders launch Artifact, a kind of TikTok for text

Kevin Systrom and Mike Krieger are back.

The Instagram co-founder, who left Facebook in 2018 amid tensions with its parent company, has launched a new venture to explore ideas for the next generation of social apps. Their first product, Artifact, is a personalized news feed that uses machine learning to understand your interests so you can quickly discuss those stories with your friends.

Artifact — whose name stands for the amalgamation of articles, facts, and artificial intelligence — is rolling out its waiting list today. Systrom said the company plans to start accepting users soon.you can sign up here; The app is available for both Android and iOS.

The easiest way to understand Artifact is like TikTok for text, but you could also call it Google Reader reborn as a mobile app, or a surprise attack on Twitter.The app opens to a feed of popular articles selected from a curated list of publishers, ranging from major news outlets such as new york times For smaller blogs on niche topics. Tap on an article that interests you and Artifact will serve you similar posts and stories in the future, much like watching a video on TikTok’s For You page adjusts its algorithm over time.

Users coming in from the waitlist today will only see the center-ranked feed. However, an Artifact beta user is currently testing two features he expects Systrom to be at the core of the app. One is a feed that displays articles posted by users you choose to follow and comments on those posts. (You can’t post raw text without a link, at least for now.) The second is Direct Her Messages inbox, where you can privately discuss posts you’ve read with your friends.

In some ways, Artifact can feel like a throwback. Inspired by TikTok’s success, major social platforms have been chasing short-form video products and accompanying advertising revenue for the past few years.

On the other hand, like social networks in the late 2000s, Artifact has a tight eye on text. But the founders hope that the lessons learned over a decade, along with recent advances in artificial intelligence, will help their app reach a wider audience.

He told me that Systrom and Krieger began discussing ideas for what became Artifact some years ago. Systrom was once skeptical of machine learning systems’ ability to improve recommendations, but his experience with Instagram has turned him into a true believer.

“Over the years, we have seen things improve really fast every time we use machine learning to improve the consumer experience,” he said.

So why come back now? Technically, this isn’t the duo’s first project since Instagram. 2020, They teamed up to create the website Rt.live Track the spread of covid.

But Systrom told me he didn’t want to start a new company until three things happened. The second is how to connect that wave to social his technology. He and Krieger continue to feel emotionally invested. And his third is the idea of ​​how a product can solve a problem. what kind of work you can do for that customer.

The breakthrough that made Artifact possible was Transformers. Invented by Google in 2017This provides a mechanism for the system to understand the language with far less typing than before.

Transformers have helped improve machine learning systems at a much faster pace, directly leading to the release of ChatGPT last year and the accompanying increased interest in AI. (Transformer is the “T” in ChatGPT.)

It also brought new possibilities to social networks. Initially, the social network showed us what our friends found interesting: the Facebook model. Then it started showing you something based on who you chose to follow, whether you were friends or not — the Twitter model.

TikTok’s innovation was to use only algorithmic predictions to display content regardless of who your friends are or who you follow. It quickly became the most downloaded app in the world.

Artifact represents an effort to do the same thing but with text.

“I saw the change and thought, ‘Otime, that is The future of social,” Systrom said. “These disconnected graphs. Why isn’t there?” Why is Twitter still primarily follow-based? Why Facebook? “

The question is whether personalized recommendations for news articles and blog posts can give Artifact the same viral success that videos brought to TikTok. This is not a slam dunk. In 2014, a wave of personalized news apps with names like Zite and Pulse came and went, plagued by their inability to create deep habits for their users. And earlier this month, Tokyo-based SmartNews, which uses similar AI technology to personalize recommendations, Laid off 40% of its U.S. and Chinese employees amid a declining user base and tough advertising market..

Like most startups at this stage, Artifact has yet to commit to a business model. Advertising is clearly a good fit, Systrom said. He’s also interested in thinking about revenue-sharing arrangements with publishers. As Artifact grows, it may help readers discover new publications and encourage them to subscribe. It might make sense for the artifact to try to cut.

Systrom also said Artifact takes its job of providing readers with high-quality news and information seriously. told me. For now, the company won’t disclose all publishers in its system, but you can search for individual outlets within the app.

Both left-wing and right-wing publishers were included. For example, there is Fox News. But Systrom isn’t shy about the fact that the company makes its own decisions about who belongs and who doesn’t.

“One of the problems with technology these days is that many of these companies are unwilling to make subjective judgments in the name of quality and human progress,” he says. “Right? Just make the tough decisions.”

Artifact will also remove individual posts that promote falsehood, he said. Also, its machine learning system is primarily optimized to measure how much time you spend reading about different subjects. For example, to reward materials that engage more deeply as opposed to those that generate the most clicks and comments.

For now, Systrom and Krieger are funding Artifact itself, but we expect investors to break the road to the door soon. Currently, from 2016 he has a team of seven working on the app, including Robby Stein, who was Instagram’s Chief Product Officer until 2021.

Systrom and Krieger, who sold Instagram to Facebook for $715 million, didn’t need to get a job. So what is driving them now?

“We basically love to build,” says Systrom. “There is no place in the world I would rather spend my time writing code and building products people enjoy.

Advances in AI have also captured their imagination, he said.

“I think machine learning is definitely the coolest thing we’re working on right now,” he said. “Not because you’re trendy, but when you recognize and fully understand that you’re interested in a particular topic, and you’re like, ‘How can you do that when you’ve just multiplied the numbers? The OpenAI CTO said that after months of basically not working, machine learning suddenly worked and worked horribly well, and I resonate with that.”

We’ve only been using Artifact for a few hours, but many of the features the company plans to build are still in the planning stages. As you’d expect from Systrom and Krieger, the app is already pretty polished. Read an article within the app and return to your feed for similar articles suggested in a handsome carousel. The app automatically switches to dark mode at night. Also, when posting a link, you can choose to allow everyone to comment, limit comments to people you follow, or block comments entirely.

In many ways, I think the time is ripe for this type of product. AI is really enabling new things in consumer apps. The collapse of his Twitter under Elon Musk created an opportunity for a team with real expertise in the field to once again venture into text-based social networking.

To be a massive success, I think Artifact needs to do more than just display an interesting collection of links.but The current depressed state of digital publishing, the web remains rife with interesting articles, as anyone who has glanced at the list of clickbait headlines under a recent Google search box can attest. Few people spend a lot of time complaining that they can’t find anything good to read on the Internet.

Yes, AI is a big part of TikTok’s success. But like Twitter before it, TikTok thrived on its capture methods. conversation About the core feed — often commented on TikTok Better than the video itselfSimilarly, one of the reasons Twitter remains a primary source of breaking news is that it’s where elites discuss news in public.

That aspect of the artifact is still under construction. But if Systrom and Krieger can bring the same craftsmanship that Systrom and Krieger brought to his Instagram to that part of the product, it may not be long before Mastodon makes us forget to log in again.