In a recent episode of the Super Data Science Podcast, Jon Krohn breaks down two increasingly important choices that shape the performance of large language models: model size and effort level. Krohn explains that model size largely determines the underlying capability and knowledge available to tackle a request, while effort level influences how thoroughly the model works through a task before considering it complete. Rather than simply choosing the largest, most expensive AI model for every job, users can get better value by matching both settings to the complexity of the task.
Krohn also offers a practical way to troubleshoot disappointing AI results: first make sure the prompt provides enough context, then determine whether the model didn’t try hard enough or didn’t know enough. The former may call for greater reasoning effort, while the latter may require a more capable model. He notes that this distinction is becoming increasingly relevant across major AI platforms and emphasizes that the cheapest model per token isn’t necessarily the cheapest option for completing a complex task successfully.
As an accomplished AI expert, educator, and host of the Super Data Science Podcast, Jon Krohn brings complex developments in artificial intelligence to life in an accessible and practical way. His ability to connect emerging technology with real-world applications makes him an engaging choice for organizations looking to better understand and capitalize on the rapidly evolving AI landscape. To learn more about hosting Jon Krohn at your next event, contact WWSG.
Anderson Cooper 360º: American spacecraft has landed on the moon for the first time in 50+ years
The lunar lander Odysseus, created by Houston-based company Intuitive Machines, successfully landed on the moon’s surface at 6:24p ET Thursday and has sent back its…
One reason so many are quitting: We want control over our lives again
The pandemic, and the challenges of balancing life and work during it, have stripped us of agency. Resigning is one way of regaining a sense…
Thought Leader: Amy Cuddy
Scott Gottlieb: How well can AI chatbots mimic doctors in a treatment setting?
This is an Op-ed by WWSG exclusive thought leader, Dr. Scott Gottlieb. Many consumers and medical providers are turning to chatbots, powered by large language…
Thought Leader: Scott Gottlieb
Jon Krohn: What Gets You Hired in the AI Era
Thought Leader: Jon Krohn
Jon Krohn: Top AI Insights from March
Thought Leader: Jon Krohn
Jon Krohn: In Case You Missed It in January 2026
Thought Leader: Jon Krohn
