Episode Details

Today's papers: - ReasonAlloc: Hierarchical Decoding-Time KV Cache Budget Allocation for Reasoning Models: https://arxiv.org/abs/2606.11164v1 - DIRECT: When and Where Should You Allocate Test-Time Compute in Embodied Planners?: https://arxiv.org/abs/2606.12402v1 - UPLOTS: A Unified Pretrained Language Model for Constrained Time-series Generation: https://arxiv.org/abs/2606.10466v1 - ABC-Bench: An Agentic Bio-Capabilities Benchmark for Biosecurity: https://arxiv.org/abs/2606.11150v1 - Assessing Automated Prompt Injection Attacks in Agentic Environments: https://arxiv.org/abs/2606.10525v1 This podcast is from Colin Davis (colin-davis.com) using Claude & Elevenlabs.