Yeah agreed, especially for legal and healthcare models.

[edit] I'd go one step further too and say the training data should be explicitly requestable too so third parties can verify (freedom of information act styles).

Training data being discoverable is much more interesting than open weights. Also a PITA so nobody does it.

Not in the West, but they will be in China and the rest of the world.

bold claim with "the rest of the world". more like rest of the 3rd world

Such a racist point of view. Everywhere outside the Western world, I said, which is clear enough, but you had to add a racist undertone to downplay the rest of the world.

Most people don't even know what 3rd world means, and just use it as a pejorative for poor countries

Don't underestimate the bang for the buck benchmark. I sure would be interested in any model that performed 10x above open models for half the cost.

If you are not self-hosting on your hardware inside your network, why do you care whether the model is open or not?

This model would be self hostable quite easily (~$14k hardware investment) if it’s released as open weights, since the base model is GLM 5.3-Flash. If not, it’s completely DOA for legal use.

because vendors stop serving models and it is good to be able to choose to self host for any reason at any time

we can start out with the hosted version for evaluation and then self host for production and real customers

In theory if your provider raised rates you could switch providers with an open model. So price competition would be one reason.

Because I don't trust z.ai and their servers and APIs, but I do trust AWS Bedrock ZDR in-region inference.