Other similar profiles
Explore more posts
-
Rajesh Krishnan
Interesting approach towards addressing data inconsistency as company starts scaling and venturing into newer products - Netflix’s Unified Data Architecture (UDA) provides a central knowledge graph and a single source of truth for modeling business entities across teams. This in turn, enables - consistent and discoverable domain definitions across teams. - Enhances automation in schema generation, data movement, and governance. - Provides semantic interoperability and lineage via a central knowledge graph. https://lnkd.in/eywgxj4r
1 Comment -
Naveen Reddy
Sharing my personal takeaways from the recent Dwarkesh-Karpathy podcast -- two technical patterns that changed how I think about AI timelines and economic impact. First: entropy collapse in LLMs--coding agents excel at boilerplate but struggle with novel problems, revealing a fundamental tension between exploitation (80% of tasks, predictable outputs) and exploration (20% of tasks, high-value innovation). Second: Karpathy's "march of nines" observation (each 9 requires equal work: 90%→99%→99.9%) helps explain why shiny demos take a decade to reach deployment, and why AI's GDP impact will likely roll into the steady 2% growth trend rather than create a discontinuous jump. If you're thinking about AI adoption timelines or where coding agents actually add value today, this might be useful context. https://lnkd.in/gsFqexEE
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content