A short blog post draws a parallel between Slate Auto‘s unveiling of a $24,950 bare-bones electric pickup truck (no stereo, touchscreen, or paint) and Thinking Machines Lab‘s release of Inkling, a 975B-parameter open-source model trained from scratch on 45 trillion tokens.

Both products are described as deliberately unremarkable general-purpose bases—a blank slate and a gray model—that invite customers to add value via customization.

The business model behind each is the same: ship a cheap, generic foundation and monetize the customization layer. For Inkling, that means offering the weights for free under Apache-2.

0 while charging for fine-tuning through the company’s Tinker platform.

The article frames this as a sustainable way to fund open-source AI development, positioning models like Inkling as the ‘pickup truck of American AI’: general-purpose, rugged, customizable, and a foundation for U.

S. open-source AI.