“They trained illegally using Twitter data. Lawsuit time.” So spake Elon Musk yesterday, as war broke out between the poly-CEO and Microsoft.
This has been brewing for a while. Microsoft is in a tight partnership with OpenAI, which Musk cofounded and (according to a recent Semafor report) unsuccessfully tried to take over; when rebuffed, he reportedly stomped off with his funding, at which point Microsoft stepped in. Now Musk is working on his own chatbot to rival OpenAI’s ChatGPT, and OpenAI is using Microsoft’s cash to invest in robots that will compete with Tesla’s promised Optimus. Everyone is in everyone’s face.
The final straw came yesterday when Microsoft informed users that, as of next Tuesday, they will no longer be able to use its ad platforms to access their Twitter accounts and manage their Twitter campaigns. This may be connected with Twitter’s recent decision to charge at least $42,000—and as much as $210,000 a month—for enterprise access to its API. Microsoft declined to explain its precise motivation when I asked earlier today.
Which brings us to Musk’s legal threat. When someone responded to that tweet by criticizing Twitter’s shift from open access to high API fees, he explained further: “I’m open to ideas, but ripping off the Twitter database, demonetizing it (removing ads), and then selling our data to others isn’t a winning solution.” Microsoft declined to comment on that one, too.
This isn’t just a matter of Musk being sore at Microsoft; it does at least partly seem to be about the principle of standing up to the scraping and use of its data to train A.I. models. Earlier this week, investor Jason Calacanis tweeted about Reddit’s decision to charge large companies for API access—CEO Steve Huffman said he was motivated by A.I. companies training their large language models on Reddit data—and Musk responded: “They’re right.”
But Musk sees OpenAI as having been swallowed up by Microsoft. “In effect, Microsoft has a very strong say [in], if not directly controls, OpenAI at this point,” he told Fox News’ Tucker Carlson when revealing his rather loose-sounding plans for TruthGPT, a less politically correct ChatGPT rival.
I wonder what TruthGPT is going to be trained on if Musk is so against scraping data on the sly. As it happens, I don’t disagree with him on that principle (though Twitter’s API fees are absurdly high and an insurmountable barrier for researchers). But given that he said he wants to create “a maximum truth-seeking A.I. that tries to understand the nature of the universe” and would therefore be “unlikely to annihilate humans,” I’m also mildly terrified by the prospect of the chatbot being primarily trained on Twitter data.
Anyhow, congratulations to Musk’s SpaceX for getting Starship’s full package off the launchpad today—even though not all engines fired properly, and the rocket failed to separate from its Super Heavy booster, and the whole thing blew up after launch, this must have been a valuable experience. SpaceX is probably going to be Musk’s most important legacy, easy as that is to forget in the noise of Musk’s excessively numerous other ventures. To which we will soon be able to add an omniscient yet hopefully merciful chatbot. Joy.
Want to send thoughts or suggestions to Data Sheet? Drop a line here.
David Meyer
Data Sheet’s daily news section was written and curated by Andrea Guzman.
