Exactly one year has passed since Elon Musk, fresh off a months-long legal battle that forced him to buy the company, strolled into Twitter headquarters carrying a sink.

At the time, we weren’t entirely sure what to expect. But there were no shortage of predictions — including from us at Engadget — about just how messy and chaotic Twitter might become under Musk’s leadership. I’ve spent the last week revisiting many of those stories, and I’ve been struck by how, for a famously erratic CEO, just how predictable many aspects of his takeover have been.

Before the acquisition closed, Musk spent months talking, tweeting and texting about his plans for the platform. Many of those early statements, like promises to relax moderation rules and to create an edit button, have actually happened.

But as is so often the case with Musk, even the most predictable of his decisions have played out in unexpected ways.

Amnesty for Donald Trump and other banned accounts

Of all of Musk’s plans for Twitter, one that drew the most headlines was his intention to restore Donald Trump’s account. (He went as far as calling the original ban a “morally bad decision.”) While Musk’s biographer claims he had some second thoughts about the matter, he opted to reinstate the former president following the result of a poll from his Twitter account.

What many may not have fully anticipated was just how many former offenders Musk was willing to allow back on the platform. Despite his initial promise that Twitter “cannot become a free-for-all hellscape,” Musk decided to offer “general amnesty” to more than 12,000 previously-banned accounts, including a number of neo-Nazis.

The slow death of content moderation (and Twitter’s ad business)

It was no secret that Musk wanted to loosen Twitter’s content moderation rules. Before his takeover, he suggested that he was in favor of allowing all speech that was legal. While many pundits predicted advertisers could be wary of Musk’s more permissive approach, it’s hard to overstate just how dire the company’s ad business has become over the last year.

A majority of major advertisers have stopped buying ads on the platform, despite CEO Linda Yaccarino’s rosier (and misleading) suggestions otherwise. Musk himself has admitted ad revenue has dropped at least 50 percent. And The Wall Street Journal reported this week that the banks that financed Musk’s Twitter deal are expecting to lose hundreds of millions of dollars as it’s become nearly impossible for them to offload the debt.

More “transparency”

One area that gave even some Musk critics a little bit of optimism was his insistence, prior to the acquisition, that he would bring a new level of transparency to the platform. He vowed to open-source the company’s recommendation algorithm, and pull back the curtain on the dreaded “shadowban.”

In some respects, he’s followed through. Code for the company’s main recommendation algorithm is on Github. X has also previewed alerts that will notify users when their accounts have been restricted from search and other areas of the service.

But those efforts may also be more shallow than what some had hoped for. Publishing “the algorithm” didn’t actually reveal much about the inner workings of the platform, according to those who have studied it. Users still have very little insight into how posts are prioritized or how accounts not belonging to Elon Musk can expand their reach.

At the same time, Musk has taken a number of steps that have dramatically reduced outsiders’ ability to understand how information spreads on X. Musk dismantled the company’s previously open and accessible APIs in favor of tools that now cost tens of thousands of dollars a month for much more limited insights. Paywalling has had a devastating impact on researchers, the vast majority of whom can no longer afford to access the limited data that is available via X’s API.

“But wait,” I can already hear a few people screaming from the comments, “wHaT aBoUt ThE tWiTteR fiLeS? Surely, that is transparency?”

While Musk’s decision to selectively leak the messages of former employees was unprecedented and potentially illegal, the truth is that the so-called “Twitter Files” didn’t actually reveal all that much about how Twitter operated. And the company’s own lawyers have refuted, in court, that the details within them are proof of any kind of government censorship or overreach.

Moreover, the disclosures themselves weren’t all that… transparent. Partial documents were only provided to a couple of writers — handpicked by Musk — who only published snippets of Slack messages, emails and screenshots from Twitter’s internal tools. The underlying documents have still not been released in their entirety, or provided to other media outlets for dissemination. Even Jack Dorsey said it would have been better, and more transparent, to release everything “Wikileaks style” for all to see, rather than selectively tweeting out bits and pieces.

X and the “everything app”

Shortly before his takeover, Musk tweeted that “buying Twitter is an accelerant to creating X, the everything app.” At the time, many people assumed Musk was talking about emulating WeChat, which in China is used for almost all facets of daily life from shopping to messaging to banking.

However, one year later, it’s still not entirely clear what he means or how the service formerly known as Twitter will transform itself into something resembling an “everything app.” There are some signs of change: He and CEO Linda Yaccarino have said they want to add banking and other financial services to X. The platform has also introduced new features like video calling and is planning on adding live shopping.

The WeChat comparison breaks down, however, when you consider that, while WeChat is unquestionably the most dominant app in China, Twitter is, at best, the sixth-most popular social media app in the US. And the app’s usage has been sharply declining for the last year. There may be room to expand the types of features offered on X, but it’s not clear they’ll result in a significant boost in popularity or engagement.

The existential threats to Twitter

If you had told me a year ago that Musk’s decisions would be so unpopular it would lead to a mass exodus from the platform, I wouldn’t have been surprised. Threatening to rage-quit Twitter over even the most minor of changes is a time honored tradition.

But if you had told me that in less than a year, there would be a surge in upstart alternatives that are starting to feel like viable challengers, I would have been less credulous. And if you had told me that, a year later, the strongest-looking challenger was a fediverse-compatible platform created by Meta, I definitely wouldn’t have believed it.

But, with a year of hindsight, it strikes me that perhaps the most significant part of Musk’s legacy so far is not what he’s done to the service previously known as Twitter, but the wave of new platforms inspired by his actions.

Bluesky, previously a backwater Twitter offshoot, has surged to a million users, despite still being an invitation-only network. Mastodon and the fediverse, both of which predate Musk’s takeover, are more popular than ever. And Meta, which has a previously dismal track record at building its own apps (not called Facebook) that people like, has managed to make Threads into a viable (if flawed) alternative.

None of these are perfect replacements for what old Twitter was, at its best. And they may not be able to sustain their momentum for years to come, but Musk’s chaotic takeover of Twitter has ushered in what many believe to be a new era of social media. It feels increasingly possible that we may look back at the end of Twitter and the rise of X as a boon for a better version of social media.