More details about the financials behind Twitter’s purchase of Magic Pony have been revealed. According to Business Insider, the deal was worth $150 million with the majority of the spend on retaining the talent at the AI technology company. This is the third AI-centered purchase Twitter has made in three years.
What does this mean for Twitter users?
The Magic Pony team will be joining the team at Cortex and add their specialty in image recognition to the growing focus on providing relevant content for Twitter users. This could essentially mean anything from creating algorithms to highlight photos or pictures that are more similar to what users are already sharing or perhaps understanding why a user shared a particular image based on their personal history.
Twitter already changed their timeline once this year to let users see “Tweets you are likely to care about most… based on accounts you interact with most.” It’s not too much of a stretch to think Twitter will move pictures you are likely to care about most to the top of users’ timelines.
Twitter has also made a push to become the customer service social media platform of choice for companies and customers alike.
Magic Pony’s image recognition focus could help here to provide companies with better insights on recurring customer service issues.
Instead of relying on keyword searches for proactive support, brands may be able to search for common images such as screenshots or locations when users don’t have their geolocation options turned on in their tweets. This may speed up new product iterations or reduce customer effort to get resolution for nagging service issues.
What it means for Twitter.
Outside of the arms race in streaming video services with other social media networks, Magic Pony adds additional armaments. Twitter has been tinkering with ways to give users more than 140 characters per tweet with rumors swirling from things such as 10,000 character limits to simply removing @names and links from the count limitations.
Twitter users often use screenshots of longer messages written in notepads or other similar apps. When users do this, Twitter is (currently) unable to cleanly convert these pictures to actionable data it can package and sell.
After all, one of core missions of Cortex is “to apply the most complex AI algorithms to our most challenging datasets, seamlessly” and Twitter’s Gnip is the gateway to the data firehose. With the new launch of #stickers available to further modify shared pictures, having the ability to track what’s behind the stickers is going to be a huge boon for Twitter’s data war chest.