The AI labs have somewhat the same problem that publishers have. Essentially their asset is a static piece of IP: a huge file full of numbers. They’re like a publisher only letting you read their book through some scuzzy Flash reader UI because they have to protect that file at all costs. At some point the weights get out/get reproduced and then what they have a is bunch of sunk costs.

I think the theory is that it’s not purely static - you need to keep training and tuning the model (even just for general knowledge upkeep with current architecture) and so the infra/data is a contributory moat. Exfiltrating weights would get you a depreciating asset (plus we have all the lovely legal and regulatory frameworks to further protect them, which IS more like publishing)

[dead]