Fun thing is I'm actually onboard with that if all said data stays local and in my control. I can understand why it's beneficial to providing a more useful experience, but also very concerned about the potential negative externalities.
Sadly that doesn't seem to be the direction we're headed in, but I could happily get behind proprietary models that run on my machine/infrastructure - I'd always prefer OSS but the key issue with generative AI as it stands today (for me) is data privacy/sovereignty.
I want to use these tools to accelerate my progress, I don't want to hand over my business ideas, or personal life to the void to be seen by quality checkers and so forth.
The balance I'm striking at the moment is experimenting heavily with them for my OSS work, since it's in the public domain regardless, and being very cautious around anything more commercial.
Sadly that doesn't seem to be the direction we're headed in, but I could happily get behind proprietary models that run on my machine/infrastructure - I'd always prefer OSS but the key issue with generative AI as it stands today (for me) is data privacy/sovereignty.
I want to use these tools to accelerate my progress, I don't want to hand over my business ideas, or personal life to the void to be seen by quality checkers and so forth.
The balance I'm striking at the moment is experimenting heavily with them for my OSS work, since it's in the public domain regardless, and being very cautious around anything more commercial.