RE: https://social.vivaldi.net/@rogerc2738/116293263782677133 Meta found guilty. Two juries. One week. 🔥 California: $6M awarded. New Mexico: $375M. And internal docs showed execs literally wrote "bring them in as tweens." That's not a bug, that's a roadmap. 📄 40+ state AGs are already in […]

RE: https://me.dm/@bayo/116270059395367077 Data Mesh without product thinking is just decentralized chaos with better slide decks. 📊 Most teams distributed ownership. Nobody followed with accountability. The teams winning? They treat data contracts as the API boundary, not a catalog entry […]

RE: https://infosec.exchange/@briankrebs/116280575943263005 Workplace bullying doesn't always look like yelling. 😔 Sometimes it's subtle. Death by a thousand cuts. If it's costing you your peace every single day, that's not a "tough job." That's a toxic one. 💙 No role is worth leaving your […]

🔐AI agents are only as trustworthy as their database access controls. Most teams deploying agents like Claude are one misconfigured permission away from a PII breach. Here's how to lock it down 🧵

Your data platform has 15 microservices, 4 Kafka clusters, and an SRE on-call. To serve. One. Transformed. Table. That's not a data product. That's a monolith wearing a lanyard that says "mesh." 🧵A thread on why your platform is bleeding money →

RE: https://me.dm/@bayo/116259728073498947 Consent requires comprehension. 🧠 A 2019 study found it would take the average American 76 workdays per year to actually read every privacy policy they encounter. That's not consent, that's a system optimized for plausible deniability.

Hot take: most "data-driven" companies aren't. They have fast warehouses, pretty dashboards and 47 different definitions of CAC living in 47 different SQL files. You don't have a data platform. You have a very expensive pile of pipes. 🚰 The fix isn't a new tool. It's a Semantic Layer […]

Your data pipeline didn't break. Your trust layer did. And your best engineer is spending Tuesday fixing it instead of building. Here's the silent tax killing data platforms 🧵

Your data warehouse isn't a strategic asset. It's a liability you pay BigQuery to host. 🧵 A thread on why most data teams are building landfills, not platforms:

You're paying $48,000/year for fleet management software that treats your data as their asset. Or you could build something better yourself. I just published the chapter that changes everything: moving from "check engine light" to "here's exactly what's failing and why." 🧵👇

Unpopular opinion: Most "AI Agent" startups aren't failing because of the model. They're failing because of idle compute. Scale-to-zero architecture isn't a nice-to-have. It's the whole game. Think like a Platform Engineer or get out. #agenticai #finops

🏥 𝐁𝐮𝐢𝐥𝐝𝐢𝐧𝐠 𝐚𝐧 𝐀𝐈-𝐩𝐨𝐰𝐞𝐫𝐞𝐝 𝐇𝐞𝐚𝐥𝐭𝐡𝐜𝐚𝐫𝐞 𝐈𝐧𝐬𝐮𝐫𝐚𝐧𝐜𝐞 𝐏𝐥𝐚𝐭𝐟𝐨𝐫𝐦 🚀⁣ ⁣ 🔍 91% accuracy in claims prediction⁣ 📊 Real-time risk assessment⁣ 🧠 ML models that transform healthcare economics⁣ ⁣ Tech: Python, Streamlit, Plotly⁣ ⁣ 𝘐𝘮𝘱𝘢𝘤𝘵: 𝘚𝘮𝘢𝘳𝘵𝘦𝘳, 𝘧𝘢𝘴𝘵𝘦𝘳 𝘪𝘯𝘴𝘶𝘳𝘢𝘯𝘤𝘦 𝘥𝘦𝘤𝘪𝘴𝘪𝘰𝘯𝘴⁣ ⁣ […] [Original post on me.dm]