Hi HN. I heard you like dev tools and AI, so we wanted to share our project that we’ve been working on. We’re working on Horizon [1] - a higher level abstraction for LLMs so that developers can spend less time trying to grapple with LLMs to make them work and more time with users.
This is the starting feature set which takes an auto-ML approach to identify the optimal LLM model, hyperparameters, and prompt - instead of just giving you the tooling to figure it out yourself. You can read more about it in our documentations.
Our view is that as LLMs become increasingly commoditized and prompts become easier to develop all the while fine-tuning is not realistic for the many organization out there - the configuration, optimization, and management will become much more difficult at-scale (consist with Andrej’s recent post [3]). The idea came as we were trying to build LLM apps (think sales outreach, ppt generation, support, etc. etc.) last fall and realized with this tool we could easily build many of the applications with Horizon in minutes versus that days it took us before.
This is the starting feature set which takes an auto-ML approach to identify the optimal LLM model, hyperparameters, and prompt - instead of just giving you the tooling to figure it out yourself. You can read more about it in our documentations.
Our view is that as LLMs become increasingly commoditized and prompts become easier to develop all the while fine-tuning is not realistic for the many organization out there - the configuration, optimization, and management will become much more difficult at-scale (consist with Andrej’s recent post [3]). The idea came as we were trying to build LLM apps (think sales outreach, ppt generation, support, etc. etc.) last fall and realized with this tool we could easily build many of the applications with Horizon in minutes versus that days it took us before.
Let us know what you think!
[1]https://www.gethorizon.ai [2]https://docs.gethorizon.ai/quickstart [3]https://twitter.com/karpathy/status/1655994367033884672?s=20