Raw LLM Responses

I like to distinguish "tools" and "products". Big closed model totally win the race in regards to productivity tooling. If your business is selling gadgets you want to sell more gadgets and will give the premium subscription to your team to any service so they can sell more gadgets. Open source is winning if you are selling an AI powered service. I don't mean here a small chatbot but volume of inference. Data-extraction and validation come to mind but anything else that say "volume" may benefit to use open source and "on prem" machines. You want a bot to mod a large community this can get very expensive with big providers API and could be done cheaply (and at a predictable price) with open source model. Requirements and results will differ for every business case but when the monthly bill is on-par with the one time cost of buying the material... This makes sense to hire someone to setup and maintain on prem machines.