Episode Details

The First Open Max: Alibaba's 2.4 Trillion Parameter Qwen3.8 and the Open Weight Balance of Power Alibaba released Qwen3.8-Max on August 3, a 2.4 trillion parameter sparse mixture of experts model, and then announced that the weights themselves land during the week of August 10, the first time a Max class flagship has ever been slated for open release. Chris and Laura dig into the benchmark claims, the still unconfirmed license, and the uncomfortable objection that almost nobody can actually run a 2.4 trillion parameter model. They also unpack the larger shift underneath the headline: Chinese open weight models now account for roughly 61 percent of tokens consumed on the largest neutral routing layer, and Meta's retreat from open Llama was a voluntary decision, not a competitive defeat. Hosted by Chris and Laura. The DX Today Podcast brings you daily deep dives into the most consequential stories in the AI ecosystem. #AI #OpenSource #Qwen #Alibaba #OpenWeights