Ep 865: Open Source AI 101: Why Local Models, Cheap APIs, and AI Agents Change Everything (Start Here Series Vol 24)
Ep 865: Open Source AI 101: Why Local Models, Cheap APIs, and AI Agents Change Everything (Start Here Series Vol 24)
Ep 865: Open Source AI 101: Why Local Models, Cheap APIs, and AI Agents Change Everything (Start Here Series Vol 24)
Open-source AI has moved from a niche experiment to a serious enterprise alternative. This episode explains why shrinking performance gaps, cheaper APIs, local models, and always-on agents are changing how organizations should choose and deploy AI.
Jordan Wilson presents an Open Source AI 101 guide focused on the practical tradeoffs between open and closed models. The episode covers model distillation, falling API prices, local and self-hosted systems, Gemma 4, agentic workflows, enterprise legal exposure, and task-based model selection. The central recommendation is to triage workloads: use inexpensive open models or local systems for routine, private, and high-volume work, while reserving premium closed models for difficult reasoning, regulated use, and customer-facing tasks that need stronger support and protections.
02:25
02:25
The open-source versus closed-source decision is now a boardroom issue
09:43
09:43
The ELO gap has narrowed by about 90%.
15:49
15:49
Frontier AI can now run locally for free
20:30
20:30
Open models are crashing the price of intelligence
24:30
24:30
A hundred agents can become dramatically cheaper
29:49
29:49
Do not use the neurosurgeon for every task
32:53
32:53
Self-hosted models enable full control
35:33
35:33
Specialized models are the next wave
