Attention, Expertise, Trust

Daniel Tunkelang
6 min readNov 28, 2022

I’m old enough to remember when the web was new, and honest enough to admit that I was a skeptic at the time. Fortunately, I was wrong, and I owe my career to the digital revolution that the web enabled.

When social networking and media platforms came along, I was more open-minded. I joined LinkedIn in 2003, Facebook in 2006, and Twitter in 2008. I hosted my own WordPress blog and later posted content on Medium and Quora. I learned about social networking and media by experimenting on these platforms, and I derived significant personal benefit from them.

Nonetheless, I quit most of them a few years ago. Why? Partly because of where I’m at in my life and career: I don’t get as much upside from self-promotion as I used to, which makes me more sensitive to the downsides. But part of it is disillusionment. These platforms could be so much better!

In this post, I’ll reflect on three areas where I see the biggest opportunities for improvement: attention, expertise, and trust.


Herb Simon, one of the greatest minds of the 20th century, said:

What information consumes is rather obvious: it consumes the attention of recipients. Hence a wealth of information creates a poverty of attention, and a need to allocate that attention efficiently among the overabundance of information sources that might consume it.

Before you dismiss this observation as cliché, consider that he wrote those words in 1971. Perhaps he was far-seeing enough to imagine our digital future. But I suspect he was just reasoning from first principles. After all, he won a Nobel Prize, at least in part, for his work on bounded rationality.

Social networking and media platforms offer a wealth — or at least an overabundance — of content that exceeds our ability to consume it. Hence, their biggest problem is allocating consumer attention efficiently.

But I don’t believe that any of the dominant social networking and media platforms do a great job of allocating that attention. Instead, they prioritize the ad revenue they can generate by selling that attention to advertisers.

Efficiently allocating consumers’ attention to content requires learning what consumers want and like. The dominant platforms train machine learning models based on engagement. But engagement is subject to presentation bias: platforms only gather feedback for the content they present to consumers. Moreover, engagement is like junk food: it tends to prioritize urgent, limbic cravings over longer-term utility. Optimizing for engagement doesn’t necessarily optimize for longer-term retention.

We can do better than following the gradient descent of engagement signals down to the nadir of human nature.

What might a better attention allocation scheme look like in practice? Perhaps one where consumers are motivated to provide explicit signals about the content they want to consume, along with simple but powerful controls to adapt their feeds to the topics and people they prefer. Maybe a middleware layer that combines powerful controls with simple interfaces. Platforms that solicit and enable this sort of personalization will not only allocate attention more efficiently, but cultivate greater consumer loyalty.

At the bare minimum, I’d love to see social networking and media platforms treat their consumers’ attention as something more than a commodity to be auctioned off to the higher bidder.


The web was supposed to democratize knowledge by providing open access to information. Not all information on the web is openly accessible, but that isn’t the main problem. The problem is, again, one of overabundance. Without training in a topic of interest, you have to rely on fundamental media literacy and critical thinking skills to decide what to believe. Unfortunately, those skills have not kept up with social network and media. Indeed, there’s even a compelling case that media literacy has backfired.

This isn’t an abstract concern. We rely on content from social network and media platforms to make decisions about our health, careers, finances, and more. What we learn on these platforms influences our views of politicians and their policies. In aggregate, this content holds enormous power.

Historically, we’ve delegated many decisions to people we consider experts. We trust doctors to make decisions about our health care and plumbers to make decisions about our pipes. We trust lawyers to negotiate our contracts or represent us in court. For every endeavor that requires significant training, there’s a class of specialists who have trained to be experts in it.

Of course, experts don’t always agree. And an expert opinion is still an opinion that you are free to ignore. Nonetheless, it’s usually wise to rely on people who have studied and worked on a problem for a lot longer than you have. I trust my instincts and my friends, but I turn to experts when my life or livelihood are at stake.

Social networking and media platforms should be great ways to find and leverage experts. Platforms like LinkedIn, Quora, and Medium seem custom-built for this goal. Yet even these platforms fall short of the mark.

LinkedIn allows people to promote themselves as experts through their profiles and feeds, as well as to provide or solicit professional endorsements. Unfortunately, LinkedIn does not offer a robust mechanism to evaluate claims of expertise, let alone to find experts. At the end of the day, Linkedin hosts the world’s biggest resume collection. Caveat emptor.

Quora has less of an excuse, given that its stated mission is to share and grow the world’s knowledge. Unfortunately, Quora’s trajectory shows what happens when you prioritize for engagement. The most popular content focuses on entertainment rather than expertise. There’s a lot of useful expertise on Quora, but finding it there is much like finding it on the rest of the web — you need to sift through a lot of crap and then rely on your media literacy and critical thinking skills to decide what to believe.

And Medium, which has struggled to find itself for years, has become just another publishing platform. Not that I’m complaining!

How can social networking and media platforms better promote expertise? First, they’d have to commit to this goal — no small feat, given that it would require them to embract the asymmetry it entails. Most people are not experts on most subjects, and large platforms tend to prioritize scale. Second, they’d need to establish expertise through credentials (I’m glad to see LinkedIn finally leaning into professional certifications), objective evidence, peer feedback, or some other mechanism. None of these signals is perfect, but the platforms don’t need to achieve perfection. Third, and perhaps most important, the platforms need to establish incentives that motivate experts to participate and not engage in abuse. That’s hard. Still, there’s evidence that platforms like Reddit and Stack Overflow have achieved this to some extent, so I believe it’s possible and worth trying.


As Abraham Lincoln said, “Don’t believe everything you read on the internet.” Seriously, if there’s one thing we all agree on, it’s that we’re not sure what content we can trust, even if we often disagree on which content to trust and why. Trust is a fundamental challenge, especially for social networking and media platforms.

You’d think that social networking and media platforms have the advantage that you can decide what to trust based on the author. But how can you be sure of the author’s identity? It’s pretty easy to create a fake profile on most platforms, and verification is an expensive, manual, error-prone process.

Besides, even if you are certain of the content’s authorship, you don’t know the its provenance. The author may have obtained the content from a source you wouldn’t ordinarily trust. The author may have undisclosed incentives to distribute the content. Or the author may have been hacked!

All of the major social networking and media platforms have faced challenges around trust. But what can the platforms do to address these challenges? Here are a few ideas:

  • Prominently display the creation date of the author’s account. Trust comes from consistency over time, which makes time a prerequisite to establish trust. It’s not that all old accounts are trustworthy and all new accounts are untrustworthy. Still, the date tells you know how much time the author has invested in accumulating a trustworthy reputation.
  • Slow down the propagation of viral content. Viral content isn’t inherently untrustworthy. But virality can distribute content so quickly that no one has the time to examine the content critically. Slowing down the propagation of viral content gives sanity a chance to catch up.
  • Surface disconfirmatory evidence. It’s a good idea to get a second opinion, especially when something looks too good to be true. When platforms detect content that seems anomalous, they can try to present content that supports an alternative perspective. This isn’t “fact checking” so much as encouraging consumers to read more critically.

Slouching Towards Dystopia

Today’s social networking and media platforms aren’t great, and they seem to be getting worse. It often feels as though their creators read dystopian science fiction novels and are treating those books as how-to guides.

Much of what’s wrong with these platforms comes down to how they manage attention, expertise, and trust. Some of you work on those platforms; others are building the platforms that replace them. Either way, I hope these ideas help nudge us away from dystopian path. We deserve better, and we can do better.