Igor Pogany introduces Claude Fable 5.2 and discusses reported benchmark results alongside his own assessment of its creative output. He presents the comparisons as observations and examples rather than a controlled evaluationEvaluation is the systematic assessment of a system's behaviour or results against explicit criteria..
Pogany shows a first-person game, a presentation created for his own teaching and interactive learning examples. He highlights visual consistency in the slides and describes how longer-running generation can produce playable interactive experiences.
He also discusses a video made with separate music and video generation tools and edited with Claude. Pogany notes the higher cost of one visual comparison and cites reported progress on remote work benchmarks while acknowledging that some tasks remain difficult.
Watch on YouTube
