Meta Unveils Muse Spark, First Model From Wang's MSL
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Meta released Muse Spark from Meta Superintelligence Labs (MSL), the unit led by Alexandr Wang, with FT reporting the model is "purpose-built" for social media apps and designed to "power a smarter and faster" Meta AI across the company's products.
- Muse Spark brings new capabilities to Meta's apps, with Engadget highlighting "reasoning capabilities" in the Meta AI app and TestingCatalog/iClarified pointing to a new "contemplating mode" rolling out to Facebook and Instagram.
- The launch marks a break from Meta's open-source Llama era — VentureBeat headlined "Goodbye Llama" while Wired examined the open-source vs. closed-source implications of the proprietary shift.
- The price tag on MSL is steep: Implicator.ai pegged the superintelligence lab at $14.3 billion, and Times of India framed the model as the "first-ever" output from one of Meta's highest-paid teams after billions in spending.
- Benchmark reality tempers the hype — Decrypt reported "Gemini Pro still leads" overall, and Implicator noted Muse Spark "can't beat Claude at coding," with FT flagging that investors are questioning the huge AI investment.
- Shares of Meta rose on the news, per Sherwood News, even as outlets including Wall Street Journal framed the launch as a "major test of AI ambitions" for Zuckerberg.
Why it matters: Meta's $14.3 billion superintelligence bet now has a public scorecard, and the early returns are mixed: the stock popped on launch yet the model trails Gemini Pro overall and Claude on coding, per Decrypt and Implicator. The strategic pivot is the "purpose-built for social media" angle — if integration across Facebook and Instagram wins users before the benchmark gap closes, Wang's team validates the spending; if not, investor skepticism flagged by FT becomes the dominant narrative.
Ask SkimNews
