FAQ

Do I use AI?

Articificial intelligence, in particlar LLMs (Large Language Models), are both a huge boon and risk for my business, so I use them conservatively and with caution.

As an example of risk, a core product for AI Companies is your data.
Not only are frontier LLMs trained on data with questionable copyright, but most model providers explicitly use the data provider to their LLMs for further training unless you pay for a much more expensive plan.

My on-site laptop has an isolated virtualized environment for the agent tools I use via QubesOS, which has no access to user data.

I do not allow frontier models access to user data unless permission is granted.
In some cases of recovery and software configuration it may be useful for AI to review or provide assistance.

When open-source locally-run models are more viable, I will be more open to their general usage.