🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

My second brain stack: - Google Keep: quick notes + syncing across devices. - Obsidian: powerful tool for organizing thoughts with backlinking and graph views - Make + Zapier: workflow automation - Blog: Hugo + GitHub + Hostinger: robust and reliable publishing

Exhaustive unit tests can be limited by computational or financial constraints. Initially, tests can be guided and informed by user testing (if available). These tests can then be extended to cover edge cases beyond basic functionality.

I got these presents. I am excited to read them :) Generative AI System Design Interview by Ali Aminian & Hao Sheng and The Notebook. A history of thinking on paper by Roland Allen.

Surprised about perplexity being in second over chatgpt. Wondering how long until the second place advantage over third becomes large

sometimes the decision making is not individual

Any ideas about how to measure the individual contribution to the changes in position and direction to the big object?

this should be an apple watch app with tags

Google AI's LearnLM tailors responses to teaching styles, enhancing education by fostering critical thinking and personalized learning. More info: services.google.com/fh/files/misc/…

doesn’t get old …

agree this book by Sebastian Raschka, LLM engineers handbook by Maxime Labonne and Paul Iusztin, building LLM for production by Louis-François Bouchard 🎥🤖 and Louie Peters are such great books.

Bye 2024. You brought so many things. Welcome 2025. Looking forward to see what you will bring.

Co-editing: "AI & Natural Learning Systems: Towards a Unified Understanding of Intelligence." We’re calling for papers on AI, neuroscience, & cognition. Join us! Details: tinyurl.com/4ty58app #AI #Neuroscience #CognitiveScience

There's a new kind of coding I call "vibe coding", where you fully give in to the vibes, embrace exponentials, and forget that the code even exists. It's possible because the LLMs (e.g. Cursor Composer w Sonnet) are getting too good. Also I just talk to Composer with SuperWhisper

In RAG systems, like in life, how you measure anything is how you improve everything. Be the person who measures recall before generation Not the person who tweaks prompts endlessly Be the person who generates synthetic data first Not the person who waits for production

|￣￣￣￣￣￣￣￣￣￣￣￣￣￣￣￣￣| |Don't Push To Production On Friday| |＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿| \ (•◡•) / \ / —— | | |_ |_