Suggestions
David E. Best
Helping people solve higher quality problems.
David E. Best is a Principal Software Engineer at Mile Two, where he engages in software development, event planning, mentorship, and team leadership. His professional background includes a diverse range of roles that emphasize collaboration and technical expertise. He is based in Miamisburg and has accumulated over 500 connections on LinkedIn, indicating a strong professional network.
In addition to his engineering role, David has experience in facilitating and mentoring within the tech community, showcasing his commitment to fostering talent and innovation. His educational background includes studies at buildspace, which likely contributes to his skills in software development and technology.
For more detailed insights into his career and contributions, you can view his LinkedIn profile here .1
Highlights
I've got a potential tech volunteer opportunity with a Dayton Area nonprofit that is in need of a web site refresh. It's an organization with a religious affiliation that does retreats and conferences.
If you've got a tech-oriented high schooler or scout that needs service hours, we could probably work something out. :)
Let me know if you'd like more info!
I was talking to a buddy about that Eric Schmidt interview about LLMs, and it got me thinking.
What will the role of "context engineering" be in the near future? Prompt engineering is a thing, and we recognize that the large models are trained on "all the data humanity has produced" (or something to that effect), so the value of context curation will get even more important. RAG is a good first attempt at context curation, but I expect we'll hit diminishing returns with larger and larger windows.
Google's own Gemini in search is an example of poorly curated context. Gemini seems to implicitly trust the search results that it gets and creates a summary of the output. But search results are an SEO'd dumpster fire. Better prompting (and dynamic context generation) could lead to immensely more useful search AI. Imagine if it didn't take product feature descriptions on corporate sites at face value, sent everything from obvious content farms straight to /dev/null, or could better understand how different types of data have different 'rot rates'? It might actually be useful.
