Meta built its AI reputation on openness -- that may be changing | TechCrunch
Briefly

Meta's Superintelligence Lab is contemplating a shift from their open-source AI model, Behemoth, towards a closed model after its training yielded underwhelming results. Although discussions are ongoing, final approval from CEO Mark Zuckerberg would be necessary for any changes. A Meta spokesperson reaffirmed the company's commitment to releasing leading open-source models despite potential moves towards closed models. Additionally, internal competition from advanced closed-source models has intensified as Meta aims to monetize AI investments while remaining competitive in a landscape dominated by firms like OpenAI and Google DeepMind.
Meta's new Superintelligence Lab is considering a pivot from their open-source AI model, Behemoth, to developing a closed model due to underwhelming performance.
The spokesperson emphasized that Meta's commitment to open source remains, stating, 'We plan to continue releasing leading open source models' even as discussions around closed models unfold.
Read at TechCrunch
[
|
]