Forum Replies Created
-
Posts
-
I’ve been focused on the data collection process at this point. I have built some Power BI reports using the data, and I’m planning on integrating the various reports and visuals into a webpage interface. A single homepage view of all the data categories (my calendar, motivational quote of the day, tasks list, KPIs, a chat interface to an AI assistant(s), etc.). It sounds simple, but the biggest hassle has been wrangling the data from external sources. The companies that we interact with that have this data don’t have much interest in our needs, use our data for their purposes, and our needs are not part of their business model.
The point of pulling all your data into a central repository that you control is that you can use advanced AI tools to improve your life. The AI agents work directly for you rather than someone else. I think this is very powerful and a big paradigm change for society.
On a technical note, I’ve started using the ChatwithRTX local AI tool from Nvidia to start connecting to the data I have (docs and YouTube content at this point). This currently uses Mistral 7B and Llama 2 13B models. In its current iteration, like many current LLM tools, this has been a bit disappointing so far. I think that producing a useful and reasonably accurate chat interface will require much more testing and likely developing more customized tools.
@St.Clair4 – The energy company I work full time for is just starting to gear up. They’ve created an AI department and started to brainstorm ideas for AI implementation. I am lucky to be involved at this point but don’t yet have a clear picture of the direction they want to go (much like many companies, I think). I’ve also been actively involved with EPRI and their digital transformation group for the past year and that group is also just starting to work on use-cases. I strongly recommend this EPRI group (and/or any others like this) to make connections and find out what other companies are working on. I really think this is vital for any company at this stage.
Generally, I have been experimenting with the commercial models (Grok, ChatGPT, and the Microsoft and Google offerings) to get familiar with the nuances of interacting with LLMs. I have learned the hard way how to effective leverage LLMs to help me with my business (developing strategy, collecting and summarizing information, troubleshooting code, etc.). I think at this point I’ve really improved my Prompt Eng skills to help me leverage the strengths of LLMs and understand their weaknesses. I think learning prompt eng basics and developing experience in a wide range of tasks is critical and way under rated.
J
Hey Y’all my name is Jay.
Background in mechanical engineering, robotics, and some data science. I spend most of my time in the Microsoft Power Platform ecosystem. I work for a fortune 500 energy company here in Texas. I’ve been doing some freelance work building apps and custom reports in the last year or so. The corporate AI race has just started and already getting deeply involved in this field. I have spent the last few years studying and trying to learn as much about AI as I can. Still much to learn but one day I hope to achieve AI superpowers:)
J
