Nick Saraev and Jack Roberts compare Qwen 3.8 Max with Fable 5 using reported benchmark chartsA benchmark is a defined test or collection of tasks used to compare a system's performance under specified conditions. and emphasize that a few days of actual use would say more than chart positions alone. They discuss model size, context length and API pricingToken pricing is the rate an AI provider charges for processing input tokens, generating output tokens, or reading cached tokens. 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 releaseAn open-weight model makes its trained parameter values available so others can run or adapt it under the release's licence. 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 tasksModel selection chooses the AI model whose capability, quality, cost, speed, safety, and operating constraints best fit a task., 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 outputsFront-end code generation uses AI to create or modify the user-interface code of websites and applications from instructions or examples. 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
