Optuna Meetup #3 talk preview connects LLM inference-time scaling with hyperparameter tuning

For a planned October 16, 2026, Optuna Meetup #3 talk, a speaker previewed slides asking what Bayesian optimization and hyperparameter optimization practitioners can do in an era they described as costing ¥300 million to choose two hyperparameters. They also said their research into LLM inference-time scaling felt surprisingly similar to Optuna, with more explanation reserved for the meetup.