🎉Claude just enabled something I've wanted to do for years. 😎I collect quotes and previously have used notion to use one in a conversation. ❗️The problem with notion is that you have to remember something about the quote to find it and there is no global search (at least when

🚀 Patterns in using AI ML to build differentiated SaaS products. ❓ Why marry SaaS and GenAI? 🪄 You get a magic mix of targeted tenant experience + SaaS margins. Baseline case: 🌍 Consumers execute requests against a shared, common experience. (Bedrock is stateless). This is

Marc Andreessen at the Allscale conference: 🚀Recommends a book called Rise of the Machines ❗️Should a computer be a van neuron calculator or closer to a model of the human brain? 🤖Many PhDs spent their lives on AI with zero results. The tech wasn’t ready. 🎉Internet is not

How to use GenAI to create custom UX with SaaS margins. The differentiator is the data is for fine tuning and RAG. The best practice is to isolate knowledge bases (security). Something (like Bedrock) will route different tenant requests (unless you love making minute infra

Downloaded my twitter archive data this week and used it as a knowledge base to have Anthropic generate tweets that sounds like me.

The hard part about creating custom UX with GenAI. Where is the context coming from? 3rd party API? Custom data deals with Reddit? Is it part of your application data (cameras on site?) Is your team a part of consortium (energy companies for climate change)? Do you pay

Tiered experiences to match GenAI infra to pricing. How one tiers makes a huge difference. Throttling? Limiting noisy neighbors? Align cost with consumption. Avoid service saturation. Nightmare scenario is a basic tier user who saturates the entire service so the highest

How does GenAI influence SaaS pricing? Do you want to align pricing to value? How much customization does a basic tier customer get? Best practice: An input (token) evaluation services to check for compliance to token limits (tiering). Also assess token complexity and

Satya Nadella, at Microsoft Ignite 🚀Moors law doubling in performance every 18 months, in AI we are seeing that every 6 months. ✅The new universal interface is multi modal 🏗️ We have new neural algebra full planning and executing capability 😎Capacity for longer term memory

New Yorker compares a 30 year-year-old today with one from 1884: Today's SF/Brooklyn used to be rural Pennsylvania. Where the donkeys were cheap and you could mine bituminous coal. newyorker.com/humor/shouts-m…