No Priors Ep. 113 | With OpenAI's Eric Mitchell and Brandon McKinzie

Reasoning models benefit from ‘thinking time’ and dynamic compute allocation.

O3 can spend more computation at inference to reason step-by-step, and empirical curves show that letting it think longer typically yields higher accuracy, especially on hard problems.

Tool use turns language models into higher-level agents rather than text generators.

By browsing, writing and running code, and iterating on results, O3 can autonomously decompose tasks like due diligence or research into sequences of tool calls, making its ‘thinking tokens’ far more productive.

Reinforcement learning on difficult, tool-based tasks is central to O3’s training.

Instead of only next-token prediction, O3 is optimized via RL to solve challenging, long-horizon tasks, learn when to call tools, and manage uncertainty and multi-step workflows.

Model steerability will matter as much as raw capability.

Users and API developers need to specify constraints like latency, cost, and depth of analysis; the vision is models that understand context (e. ...

Browsing and vision tools sharply improve test-time scaling and reliability.

For images and current information, O3 can recognize its own uncertainty and then act (e. ...

High-quality evals are as strategically important as high-quality training data.

As frontier models ‘solve’ many existing benchmarks, uncontaminated, well-designed evaluations become critical to measure real progress and avoid optimizing blindly against noisy or saturated metrics.

AI is reaching an inflection point as a force multiplier for AI research and coding.

The guests already use O-series models multiple times per day for navigating large codebases and research tasks, and they see a near-term loop where models materially accelerate the development of their successors.