Personally I like the idea of something like chatgpt replacing google search.
But, and a huge but, I miss the sources its trained on. Are these copyrighted sources, who owns the rights to the content it spews out? I like to know whos content I am reading and or using (when opensource) to reference them when needed.
Use UFW with fail2ban and set a whitelist for IPs that can connect to ssh. Stopping connection attempts is like trying to stop a toddler from touching everything. IP and Portscans run continously day in day out. The servers I manage are scanned all day long and as safe as the security you set.
Self host in the cloud or self host locally? I run a bitnami vm locally accessable by token based ssh or key based openvpn. So in essence its in the cloud