Greetings!
Guess what: Meta Platforms is still in the AI model game. You might have forgotten that fact, given that it’s been a little while since we’ve seen much new on the AI front from the Facebook owner. But CEO Mark Zuckerberg took to Threads on Wednesday to reveal the new model family, known as Muse, and its first member, Spark.
Meta stock jumped 6.5% on the news, although the market was up strongly, so not all of that was a reaction to the new model. Still, the release is good news. The New York Times had reported last month that Meta had put off the release of its new models because they fell short of rival models’ performance. Given that Meta spent many billions building a new team of engineers, led by former Scale AI CEO Alexandr Wang, any meaningful delay would have been a disaster.
So will Spark light enough of a fire in the AI market to change Meta’s trajectory? Maybe. Meta seems to be aiming at the consumer market—Zuckerberg noted that the new Meta AI would be “particularly strong” in areas such as “health, social content, shopping, games.” While many consumers are likely already set on their favorite chatbot—most likely OpenAI’s ChatGPT or Google’s Gemini—there may be quite a few still undecided. Meta’s AI chatbot is installed on its social media apps, which reach about 3.5 billion people. That should give it a leg up if it can get some positive buzz for the new AI models.
That’s a big if, of course. What makes this interesting is what Meta’s success could mean for OpenAI. Right now, the ChatGPT firm’s biggest hope in driving revenue growth may be advertising, seeing as Anthropic seems to have the enterprise market sewn up. But the two companies that know how to sell digital advertising best are Meta and Google. Both are already using their AI tech to turbocharge their ad businesses. If Meta can make progress on the chatbot front, OpenAI will have an even harder time building a decent ad business in ChatGPT.