This is way too short-sighted. The reason most companies are not using it is because they can't.
- The chatgpt API was not even public a few weeks back. Most companies don't have the infra to build a LLM so they rely on big tech or some startups for it and it takes time to launch finished products.
Google doesn't even have any APIs right now I think? Or it may just be i. Early release.
- Companies are not ready to share their data with LLMs, they don't know if their own data will get used in training and be used against them. Unless they can get some ironclad SLAs around this thing they will not use it. You don't want copilot to sniff around your codebase. Neither do you want gpt to suck up on all your internal knowledge bases.
- They are all waiting for the ecosystem to mature. Why hurry when your own cloud provider will launch their offerings soon and you can easily deploy it in your infrastructure.
- Copyright infringement. Gpt-4 paper didn't tell about the training data. That's a red flag and they might be using proprietary information for training.
