<?xml version="1.0" encoding="UTF-8"?><oembed><type>video</type><version>1.0</version><html>&lt;iframe src=&quot;https://www.loom.com/embed/5ac2b2bdfaa44a29861fd8af73ecd834&quot; frameborder=&quot;0&quot; width=&quot;1898&quot; height=&quot;1423&quot; webkitallowfullscreen mozallowfullscreen allowfullscreen&gt;&lt;/iframe&gt;</html><height>1423</height><width>1898</width><provider_name>Loom</provider_name><provider_url>https://www.loom.com</provider_url><thumbnail_height>1423</thumbnail_height><thumbnail_width>1898</thumbnail_width><thumbnail_url>https://cdn.loom.com/sessions/thumbnails/5ac2b2bdfaa44a29861fd8af73ecd834-3ba454d1c04a48a1.gif</thumbnail_url><duration>821.001</duration><title>Choosing the right ChatGPT model to optimize your token usage.</title><description>This Loom discusses rising AI spending costs and the resulting need for better AI model literacy. It cites an example where Uber burned its entire 2026 AI budget in just four months due to the massive amount of tokens used by Claude Code, and notes Microsoft’s CEO telling employees to use appropriate models rather than frontier models for non-frontier purposes. The video then explains that in ChatGPT you can start with an auto model for low-stakes, everyday questions, while more important tasks require being in control of which model you are using. It also says the instant model is best for simple, quick tasks that do not require much thinking.</description></oembed>