I mean you can't social engineer a human using poetry? Why does it work for LLMs? Is it an artefact of their architecture or how these guardrails are implemented?
It's sad that they only compare to open weight models. I feel most users don't care much about OSS/not OSS. The value proposition is the quality of the generation for some use case.
I guess it says a bit about the state of European AI
I think the future of videos as a source of truth is bleak.
You can't trust anything that you can see in a video. And you can deny any video.
I think this will make an interesting post-truth world. What will happen when we can't almost prove anything. Eye witnesses are untrust worthy. Videos ditto. Cryptographic proof? Will politician sign their videos and messages with their public keys?
I don't want to make big generalizations. But one thing I noticed with chinese models, especially Kimi, is that it does very well on benchmarks, but fails on vibe testing. It feels a little bit over-fitting to the benchmark and less to the use cases.