CxC Ep23: Scaling Ontologies скачать в хорошем качестве

CxC Ep23: Scaling Ontologies

CodeStrap's "Code and Connor" Episode 23 features our friend Francisco Ferreira, Founder and CEO of Datalinks, a company that is laser-focused on bringing human understanding to the core of the data and AI value chains. Prior to founding Datalinks, Francisco spent 9 years at Palantir. At Palantir, he architected and created the product and infrastructure group for Skywise, the world’s biggest aviation platform, growing it from 3 airlines to 100+. He went on to create the Developer Relations groups and co-led Channel partners. Before Palantir, he had a multi-year stint at eBay. This episode focuses on: → Francisco’s journey from Palantir to DataLinks → Large scale integrations and how to do them successfully → Why a clean ontology matters and how to build reusable components → What Datalinks is working on, including AI-generated ontologies → Enhanced search/retrieval across complex data relationships → The AI Bubble → AI model benchmarks and the pressures to align to them → Frontier models, open-source, small language models Be sure to follow us to get the latest information, and schedule a meeting with us through www.codestrap.com! Datalinks Website: https://datalinks.com/ LinkedIn:   / datasetlinks   Francisco:   / franciscomsferreira   CodeStrap Dorian:   / dorian-smiley-97a72a14   Connor:   / connordeeks   CodeStrap (X): https://x.com/CodeStrap411 CodeStrap (LI):   / codestrap   Chapter Codes 0:00 Intro 2:10 Francisco to Dorian and Palantir Developers to CodeStrap & Datalinks 6:16 The definition of ontology, how it works, and how to scale them from the Skywise Architect himself - proper data foundations, schemas, interfaces, actionability, etc. 19:15 Why a clean ontology matters for both human and AI, technology and business...and what happens when you act like it doesn't matter (e.g., Microsoft) 26:30 Leveraging AI for generating ontologies, with tariffs as an example use case 38:39 Lack of capacity in GPUs, the killer AI use case, the codepocalypse from ads, space travel, and the workarounds 43:28 The AI bubble, misunderstanding LLMs, and AI hype 52:38 Throwback - Gemini 3 release (10 years ago in the age of AI) and benchmarks vs. first principles of software and business 57:00 Fine tuning and small language models, and covering privacy and security with them 1:04:07 Timeline for mass adoption for agentic 1:05:55 Outro