Maybe because it’s a badly coordinated mishmash of Project 2025 and the random political revolution theories of Peter Thiel’s favorite Silicon Valley “philosopher” Curtis Yarvin 🤷‍♂️

All that self-debasement just to get someone’s systems permissions 🙄

I can't wait to hear their take on source material for requirements. If these guys are pushing past civil workers who would typically spell out processes that need to be replicated, they're probably just feeding available random documents into an AI and asking it what the requirements are. 😬

So glad you’re on this angle. Having worked with gov’t data as well as AI, the requirements of this endeavor are enormous, and the process/product is rife with anti-competition and conflicts of interest, not to mention resulting power. Already reached out to my Rep/Senators, if they’re listening

There is no "out-of-the-box" data permissions/security management in LLMs, so the coders will have to build it from scratch. The risk of data mixing is very high here. This should never touch real-life data until after extensive testing. These engineers aren't smarter than reality.

The test cycles needed for a tool like the reported GSAi chatbot MUST be extensive due to confidential/private data being analyzed/synthesized for response. You would need human SMEs to craft and evaluate prompts and responses related to every process in the government and catch errors.

This isn’t internet research, I lived it. It was like dating a family to see if they’d want to teach you. Lots of dead ends and half promises. I finally applied to college under everyone’s backs and told my parents right before my first semester. I’ve never looked back. It was shady boots.