Social media is the new nicotine. The Surgeon General already declared as much for youth in May, and as of yesterday is being urged by 42 state attorneys general to do so for adults as well. I see it’s toxic effect every day on my family, including myself, my kids, and my wife. We are now literally unplugging our internet router when the work day is over to save ourselves from the ubiquitious, addictive draw of useless stimulation.
At the same time, social media is the most efficient communication and knowledge sharing technology ever invented. News spreads at lightning speed with hyper targeting. Unfortunately, this efficiency is achieved by a workflow that inevitably leads to addiction. But I believe AI can offer an alternative workflow that preserves today’s efficiencies while 1) eliminating addiction and 2) promoting the spread of deeper insights.
Infinite Scroll of Doom
The Workflow
Today’s social media operates with the following workflow.
- Users follow other users.
- Users see a news feed of content published by those they follow.
- Users engage with content.
- Algorithms show users more content similar to what they previously engaged with.
- Rinse and repeat.
This is designed to lead to more and more engagement over time.
Why a News Feed?
So why did every social media platform build this same workflow? Because a ranked list of content was the only way, with technologies available at the time, to deliver relevant information to users.
But news feeds actually aren’t that efficient at delivering relevant information to users. The invention of infinite scrolling is proof in point – people never really find the information they’re looking, all they find are little dopamine hits from never ending distraction.
Breaking News Feeds
AI will break the news feed.
I don’t want to scroll forever through content that some numeric optimizer decided was similar to things I spent the most time looking at in the past. And I don’t think anyone else does either.
What I do want is to discover new information that helps me achieve my goals. Whether those goals are catching up with old friends or learning about the latest AI breakthroughs, I want my information provider to optimize for them.
In 2006 when Facebook invented the Feed, the state of AI was not capable of anything more than a ranked list of information optimized for click rates.
But in 2024, AI is capable of filtering, sorting, summarizing, and synthesizing information to meet arbitrary objectives defined by users in natural language.
AI is also capable of conversing with the user about the content they see and helping to craft comments and original posts that add to the collective conversation. AI can be a thought companion helping users process large volumes of data and make sense of it to meet their objectives.
AI can even learn a user’s objectives naturally through these interactions, without the user needing to explicitly explain or configure some complex set of goals.
More Utility, Less Addiction
A social media platform that optimizes for each user’s individual goals will provide more value to each user with dramatically less screen time and addiction. And building such a platform is feasible with today’s technology.
But new business models are needed to capitalize on this opportunity. Less time on platform means fewer ads. But more value for users means increased willingness to pay directly for the service (subscription revenue).
So who will be the first to build the next generation of social media?
