The presenter compares the public Fable release with its restricted Mythos counterpart, explaining reported safety classifiersAI safety is the field and practice of reducing harmful failures, misuse, loss of control, and unintended consequences from AI systems., routingAI model routing sends each request to a model chosen for that request's complexity, cost, speed, privacy, or other requirements. and the distinction between model evaluationsEvaluation is the systematic assessment of a system's behaviour or results against explicit criteria. and everyday use. The discussion covers coding, cybersecurity, science and alignment assessments as reported by the model provider.
In hands-on experiments, the presenter describes useful outputs alongside flaws in simulations and visual tasks. They report high usage costsAI inference economics examines the revenue, cost, utilization, pricing, and margins involved in serving trained models. and conclude that the model did not offer a clear advantage for their own routine workflow.
Watch on YouTube
