it is a regression. the user space application also would silently fail, it is a chain of oversights.
also having the encryption keys in memory does not mean you can extract them, it is more of unnecessarily letting it there indefinitely, not having it where it shouldn't be.
You are not allowed to use any model from Chinese labs at American companies. Even from other companies like Mistral it is hard to get access, since enterprises require their own instances on vertex/bedrock/azure, and to run alternatives you have to go through multiple layers of management. In practice, no employee will proactively advocate for alternative models and the oligopoly is written in the wall, specially since the investments are circular anyway.
well, I think this is so common now that it does not even get attention anymore. the worst part is that a data leak from Microsoft wouldn't really benefit anyone, since they have nothing technical of value, just ungodly amounts of money.
ok, but you had 1x token to generate , then more 1x to review locally, 1x for the agent local, 1 x for the cloud. then ???x until all bots are satisfied.
You end-up spending at least 5x the amount of tokens for maybe prediction machine to find a discontinuity?
I would say a way better approach is 1.123x to generate code + tests + passing analysis tools + human review + 1x "simplify as much as possible", than letting the snake its own tail without boundaries.
So... nvidia agreed to pull a qualcom... well, enjoy the failure. people that would be early adopters want a real operation system that would actually allow them to leverage the hw, not a pathetic web-ui-based vibe coded operation system that requires wsl to make anything useful.
a good harness is supposed to do what you are describing.
sonnet on pi.dev is pretty terrible but fast. Claude Code has ridiculous amounts of prompt engineering at system prompt level and sub session spawing combined with low temperature, to provide the predictable results people like. CC screws up and you never see, because the harness auto corrects, while on OSS you see everything, and does not comes with the level of monitoring by default.