Elon Musk's Twitter Solutions Examined by Former Twitter Head of Engineering

Former Twitter head of engineering, Alex Roetter, went on Alex Kantrowitz’s Big Technology Podcast on April 27 to discuss Elon Musk’s potential proposals to save Twitter, as he comes close to officially finalizing the deal to acquire the giant social media platform. 

Barring a last second catastrophe that would result in the loss of $1 billion for either Twitter or Musk, the Tesla CEO is set to buy Twitter for a total of $44 billion at $54.20 per share. Musk secured the funding through $25.5 billion of fully committed debt, which included $12.5 billion in loans against his Tesla stock, along with selling 9.6 million of his Tesla shares valued at $8.5 billion. 

While the SpaceX founder and CEO hasn’t released a formal plan for his proceedings post Twitter acquisition, and not that he has to, Roetter discussed some of the big ideas Musk has already talked about in the previous few weeks, including authentication humans, defeating the “spam bots,” free speech, paid Twitter subscriptions, open sourcing the algorithm and creating editable tweets. 

Out of all of the propositions discussed Roetter was extremely high on authenticating all humans in an effort to cut down on pure anonymity, “foster[ing] the worst parts of speech online.” Twitter most likely has not done this in the past to keep Wall Street happy, as they knew that if they authenticated all humans, the total account number would plummet. 

In general, Roetter explained how a lot of these issues are interconnected and the nature of that interconnectedness could pose a challenge to accomplish them. For example, in order to promote free speech, which is probably the primary reason for the Musk takeover, you would have to be lenient in your effort to defeat the spambots. 

A spambot is an autonomous program on the internet that sends spam to a large number of users or posts spam on online forums. If you want to defeat them, Roetter said a classifier that looks for characteristics intrinsic to bots must be created and tuned to Twitter’s liking. The problem is if you tune it to an aggressive setting, you will also flag and potentially ban real human accounts, which would of course go against Musk’s free speech campaign. On the contrary, if the classifier is tuned to be lenient, the spambots will not fully be eliminated. 

Not eliminating spambots might go against Musk’s April 21 tweet, “we will defeat the spam bots or die trying!”

Roetter also liked the idea of paid Twitter subscriptions for certain premium features. Not only will this create additional revenue, but it will also reduce spam. 

“It’s a really interesting idea,” Roetter said. “If you don’t want a bunch of what you think of as low-value activity to happen, if you charge more than the value that you think people are extracting from it, it should go away.”

The one potential proposal Roetter was skeptical on was making the algorithms open source. “This one is a head-scratcher to me,” Roetter said. 

He explained the algorithm which essentially looks at billions of examples of content and tries to predict which ones each user will react to the best in order to optimize the experience for each user. There’s nothing other than that, that the algorithm shows. However, open sourcing the algorithms would mitigate the amount of conspiracy theories out there claiming that if you are of a certain political party you are more likely to get banned. 

Author: Tyler Irvin

Related Articles