The company that claims it "may never be profitable" just signed a deal to take even more of your tax dollars, is what I'm reading.
GSA should be working to host its own inference and models for federal agencies, not selling out to unprofitable private infrastructure.
Read their SEC filing, then decide if the "discount" is sustainable as a multi-agency dependency for decades to come. Or just blindly accept this as a positive and downvote away.
This isn't the government subsidizing a private venture, it's a private venture offering a discount to the government. What does profitability have to do with whether or not the government should purchase their services?
...no? Whatever value they got from inference they already did is still there? Of course the government still needs to worry about vendor lock-in, but they'd have to do that no matter what, and also general-purpose LLMs are relatively replaceable for most use cases.
I'm of the mindset that "GovCloud" and related products that make government dependent on private companies for vast government services is a bad idea.
If it's critical infrastructure, it should be publicly owned, funded, and protected by the federal government directly. Not rented from someone else under contracts that can be violated.
There is a whole side of sci-fi that has explored the idea that AI government could finally cure us so much more than just one class of corruption. There are, however, new downsides.
GSA should be working to host its own inference and models for federal agencies, not selling out to unprofitable private infrastructure.
Read their SEC filing, then decide if the "discount" is sustainable as a multi-agency dependency for decades to come. Or just blindly accept this as a positive and downvote away.
If it's critical infrastructure, it should be publicly owned, funded, and protected by the federal government directly. Not rented from someone else under contracts that can be violated.