Meta bets billions on ‘muse,’ shifting ai strategy from open source to closed
Meta’s latest artificial intelligence model, dubbed ‘Muse,’ marks a dramatic pivot away from its previous open-source approach and a significant investment – estimated to run into the billions – designed to finally challenge the likes of OpenAI and Google in the rapidly escalating AI race.
A closed-door gamble
The unveiling of Muse, spearheaded by Meta’s Superintelligence Labs and a newly assembled team of high-level AI researchers, underscores a clear strategic shift. Unlike previous attempts, Muse is a proprietary model, effectively locking its design and code behind closed doors. This wasn't a calculated move; it was a reactive one, born from the pressure to keep pace with competitors aggressively pursuing market dominance.
Following the announcement, Meta shares surged 6% in New York trading – a notable, albeit likely short-lived, boost. However, the underlying question persists: is this a genuine resurgence, or merely a desperate attempt to stem the tide of what some analysts are describing as a ‘tabula rasa’ moment for the company’s tech standing?
The spark is… dim
While the fanfare surrounding Muse is considerable, the reality on the ground is arguably more nuanced. Initial assessments suggest that Muse Spark, the initial iteration, doesn’t quite match the performance of leading models like OpenAI’s ChatGPT, Anthropic’s Claude, or Google’s Gemini. A member of the development team acknowledged as much, stating it represents “a first benchmark on our trajectory.” But Meta remains steadfast, hinting at even larger models currently in development – internally referred to as ‘Avocado’ during its creation – a deliberate departure from its previous focus on Llama models.
The company’s leadership, under Mark Zuckerberg's renewed investment, is prioritizing agility. They’ve reportedly granted significant autonomy to researchers, dismantling the traditional, hierarchical structures that have previously hampered innovation. However, this strategic shift hasn't been without consequence: hundreds of jobs have been eliminated across Meta’s AI division, a stark illustration of the ruthless prioritization driving this new direction.
Beyond the chatbot
The immediate goal of Muse is to enhance Meta’s existing suite of applications – Instagram, Facebook, and WhatsApp – but the long-term vision stretches further. Meta is exploring the potential for subscription-based access in the future, a move that could fundamentally alter the economics of AI interaction. The core chatbot functionality, however, will remain free for users, a calculated risk to broaden adoption – a tactic reminiscent of early social media strategies.
Despite the substantial investment and operational adjustments, Muse remains in its nascent stages. The $70 billion allocated to this initiative, coupled with the workforce reductions, paints a picture of a company aggressively staking its claim in the AI landscape. But whether this gamble pays off remains to be seen. The race to dominate AI is far from over, and Meta’s move – however bold – may ultimately prove to be a fleeting moment of momentum.