"Open"AI's Closed LLM
Corporate customers, many of whom already view Gen AI as hype, surely will not be thrilled that OpenAI decided to hide o1’s chain of thought, essentially making the model a black box.
Google’s Gemini will likely follow suit as will Anthropic. Hopefully Meta will allow its open source Llama model to remain open source, which could serve as a real competitive advantage versus the other frontier LLMs should they go the closed route.
If your firm is considering using OpenAI’s o1 model or any LLM, I would be interested to hear your thoughts on Gen AI.
What are your thoughts on the various LLMs?
Is it important that you operate a model with Nvidia’s latest chip architecture, or, would an LLM powered by an alternative such as Groq’s LPUs or Google’s TPU chipset suffice?
What are the use cases for which you are contemplating deploying Gen AI? If you have deployed Gen AI in one form or another, have you realized an ROI?
You may reach me at jmaietta@tek2day.com or info@tek2day.com