Twitter calls Elon’s bluff
Elon Musk doesn't really want to buy Twitter. And he thought Twitter's refusal to provide user data was his way out. But Twitter just agreed to share the data. What now, Elon?
Twitter just reversed itself. Now Twitter says it will comply with prospective owner Elon Musk and his demand to see user data.
Musk demanded the data weeks ago, and Twitter refused. Now, in what sounds like passive-aggressive compliance, Twitter will give Musk the “full firehose” — more than 500 million tweets a day. That “firehose” may be provided as early as this week.
To me, this feels like a game of chicken, where Musk really doesn’t want to buy Twitter for a variety of reasons (listed below), and Twitter is basically daring him to.
Mike’s List of Brilliantly Bad Ideas
Ford trolls Tesla with Tesla-charging accessory
Ford wants to promote the ability of its electric F-150 truck to send electrical power in the other direction and charge things. So to drive that point home, Ford is apparently including a third-party accessory from a company called Lentz for charging a Tesla from an F-150.
Now, it’s only a matter of time before an F-150 truck own charges a Tesla in public, and the resulting photo will go crazy viral. This is a next-level marketing stunt.
What the world needs now: shitposting antisocial AI
Georgia Tech scientists are building a better graffiti robot. Called GTGraffiti, the wall-defacing machine not only writes graffiti, but also mimics the body movements of human graffiti writers as they’re tagging.
It does this by using motion capture equipment on graffiti writers, which is then applied to a robotic cable and pulley based robot.
The research aims to be able to record graffiti as it’s being created, and eventually to expand the capabilities of graffiti artists — for example, to enable to tag an entire building with magnified versions.
Meanwhile, An AI researcher and YouTube influencer named Yannic Kilcher trained a chatbot using three years of toxic trolling posted on the Politically Incorrect 4chan message board. And then he figured out how to enable his bad bot to post on that very board. The bot, called GPT-4chan, perfectly captured 4chan’s “mix of offensiveness, nihilism, trolling, and deep distrust of any information whatsoever.”
The bot posted roughly 15,000 times in 24 hours.
Some AI researchers criticized Kilcher’s project. But he countered that nobody on 4chan would be shocked by what his bot posted. It’s really just holding a mirror up to 4chan’s “Politically Incorrect” board. He also pointed out that he doesn’t allow the bot to be downloaded and used for targeted campaigns and that, in any event, his bot is extremely hard to control.
I’m probably in the minority, but I’d love to see these projects combined into a single robot that spends its days shitposting on 4chan and its nights defacing public property all over the city. If anyone is going to be replaced by a robot, it might as well be the sociopaths.
Mike’s List of Shameless Self Promotions
Why remote work will lead to greater success (Subscribe to my Future of Work newsletter!)
What’s so great about Google’s ‘translation glasses’?
Malicious reconnaissance: What is it and how to stop it