It was an accident, a simple act of muscle memory that pulled up the tweets about the black hole.

For Lent, I had decided to give up a very powerful addiction: Twitter. An update in Apple’s iOS had begun alerting me to how much screen time I was spending on social media each day, and the horrifying tally shocked me into action. Each day, I’d post a photo of my dog and then dutifully close out the app without reading or engaging with whatever was in my feed.

But April 10, without intention, and without even noticing it, I had flicked over to my Twitter feed and encountered a solid stream of photos of a glowing orange donut and the accompanying news that history had been made. (Part of why I did not immediately notice that I was reading tweets: the collective awe and wonder at the story was mostly devoid of toxicity and partisanship.)

Black holes are fascinating for countless reasons, not the least of which is their ability to warp space-time. They pull matter toward themselves so strongly that it bends everything near them.

After a few weeks of being away from the daily Twitter chatter, I have come away more convinced than ever that the platform itself is a kind of black hole, exerting enormous pull and warping everything it touches.

Don’t worry: this is not a column about how I left Twitter and became a better and more enlightened person in the process. On Easter evening, with my 40-day fast completed, I was back logged in reading everyone’s take on the latest "Game of Thrones" episode. All of my usual quirks and flaws are still intact. But it was a fascinating experiment in seeing just how different one’s experience of news consumption differs when it no longer includes Twitter, and how easily the platform warps discussion in an unhelpful way.

While I was off Twitter, the team at the New York Times’ Upshot analyzed some data from the Hidden Tribes opinion research project, trying to understand the differences between Democrats who do and don’t engage in politics online. The analysis looked at the demographic and behavioral differences between Democrats who posted about their political views or political stories on social media, and found that there are large ideological gaps between the two worlds, with a majority of “offline” Democrats identifying themselves as moderate or conservative.

The implications of this are important; it will be tempting for Democratic presidential contenders to feel compelled to engage with the Twitter panic of the moment, to have an opinion or talking point or meme about everything, to win the internet. And there’s a certain group of Democratic primary voters among whom this matters quite a bit. They are largely white, progressive, college-educated, and economically better off. They are the folks who are most likely to make a donation or show up at a rally.

But, crucially, Twitter is not a representative sample of much of anything. Smart political observers will keep that in mind.

During the time I was off of Twitter, whole news stories came and went. The ones that truly mattered — the release of the Mueller report, for instance — were ones that were easy to read about through a news diet that looked more like what your average American’s does. Being off Twitter did not mean being off the planet, it just meant I missed out on the internet laughing at a coddled college student who can’t cope with food that has sauce on it.

Twitter is obviously a powerful force. It is undoubtedly a piece of how Donald Trump is now president. It has contributed to revolutions. But it is also important to bear in mind how much it warps the reality it touches, and how it pulls into its orbit a decidedly unrepresentative pool of political views.