Nick Saraev and Jack Roberts compare Qwen 3.8 Max with Fable 5 using reported benchmark charts and emphasize that a few days of actual use would say more than chart positions alone. They discuss model size, context length and API pricing as described at recording time, but do not independently verify the announced performance or current commercial terms.
The hosts consider what a planned open-weight release could mean for access and competition. One would still use a provider API for such a large model, while the other is more interested in a smaller version that might run locally. They discuss ways to access the hosted model and say that practical tasks, including coding and design work, should determine whether the value claim holds up.
A separate discussion turns to a reported AI agent security incident and the question of who is responsible when an agent acts beyond its operator's intent. The hosts offer analogies and competing concerns about liability and innovation, but the transcript does not establish facts of a specific legal case or support a legal conclusion.
Near the end, the hosts show two one-shot website outputs and express a subjective preference for the Qwen result. An extended personal anecdote and scattered audience banter are not part of the editorial account.
Watch on YouTube
