Talking about the practical aspects. In 2024, there were weeks when the major players like Google, OpenAI, Meta, xAI, Anthropic, NVIDIA, IBM, Amazon, or startups (e.g. the latest buzz with DeepSeek), would launch multiple new Large Language Models or tools, within a matter of days.
Some of these would reset prior limitations and open new opportunities.
That's fascinating progress but... how do you keep up with all the novelty if you are deploying applications within an enterprise?
At Robotic Online Intelligence (ROI), our approach has been to focus on what doesn't really change - workflow and abstracted business problems, with the flexibility to plug in (and test) new models and methods.