4 ms·
There will be frontier models that are non-commoditized, but they'll be kept guarded and hidden away, and you'll only get the final result, so that they can't b
by CuriouslyC 4mo ago
There will be frontier models that are non-commoditized, but they'll be kept guarded and hidden away, and you'll only get the final result, so that they can't be distilled and their harness can't be reverse engineered. They'll be billed like employees, rather than like a tool.
- greenavocado 4mo agoEverything can be distilled, it will just become more painful
- naravara 4mo agoI think this will be isolated to highly specialized fields where training data will need to be selectively curated.
- hedora 4mo agoThe non-commodity network services of the early 1990’s and the non-commodity 3d graphics hardware of the mid-1990s made the same argument.
- ls612 4mo agoThey didn’t have the security state backing up their business thesis at gunpoint.
- yandie 4mo agoI doubt that. What stops the Chinese labs from figuring it out? It’s not like these models are fundamentally different from each other
- CuriouslyC 4mo agoIf all you have is the starting point and the finishing point, the lack of the path taken from one point to another limits your ability to train models that can efficiently recreate the work, and increases its cost enough that it's possible the US labs can progress capabilities faster than Chinese labs can distill that behavior.
- wahnfrieden 4mo agoThat’s already the case. Chinese ingenuity allowed them to achieve what they did without access to reasoning outputs
- throw10920 4mo agoThis has got to be satire. Everyone, especially Singaporeans, know what "Chinese ingenuity" really is.
- wahnfrieden 4mo agoIt’s merely descriptive of ingenuity required to distill models back into reasoning models without having any of the chain of thought. You underestimate the original work required because of biases
- throw10920 4mo agoIt's a uniquely Chinese cultural feature to laude finding more clever ways to cheat and steal and deceive as "ingenuity".
- sealeck 4mo ago> lack of the path taken from one point to another limits your ability to train models that can efficiently recreate the work Isn’t this the problem inference (training) a model is designed to solve :)))
- jmalicki 4mo agoIt is! And it's a hard problem. What's an easier form of training is being able to see the intermediate results and train to imitate them.
- nyrikki 4mo ago[dead]
- 4mo ago
- mingqiz 4mo agoIsn't that what they are doing already? The model is already guarded and hidden and i only get to send it what i want. Talk with it to clarify my requirements. And i can switch to a different provider for cheaper/better results.
- lacy_tinpot 4mo agoThey tried to do that with operating systems and the browser.
- throwaway85825 4mo agoThe economically useful frontier models will be fine tuned on data to make them useful for a specific project or task.