Why Post-Training Is the New Pre-Training in AI
_Personal notes on data labeling, RLHF, and where the real gains are happening now._ --- I used to think model quality was mostly a function of scale—more tokens: more parameters, more compute, etc....
Exploring digital platforms, markets, and emerging systems through the lens of content strategy, UX clarity, and real-world execution—written with editorial discipline and hands-on experience.
_Personal notes on data labeling, RLHF, and where the real gains are happening now._ --- I used to think model quality was mostly a function of scale—more tokens: more parameters, more compute, etc....
On January 30, 2026, gold and silver suffered historic volatility spikes after parabolic rallies, with silver plunging up to 36% and gold dropping over 12% intraday. Explore the triggers—including Trump's Fed chair nomination—and what this extreme reversal means for markets and investors.
Boost your productivity with a lean AI stack: o3 for strategy, Claude for writing, Gemini for prototyping, plus NotebookLM and Deep Research—powered by context engineering.
Learn how context engineering—using RAG, memory, and clean chat history—can boost AI accuracy and productivity by 10×. Slow down to speed up your workflow.
Discover how ScoreGenius uses machine learning to transform NBA statistics into powerful pre-game predictions. Our lightweight PWA delivers professional-grade analytics with zero friction—no accounts, no ads, just pure basketball intelligence.
Google Gemini 2.5 Pro brings game-changing reasoning abilities and a 1M token context window. See how it stacks up against competitors and whether it deserves a spot in your AI workflow.
Discover how strategic discomfort accelerates career growth while avoiding burnout. Learn the neuroscience behind professional development, including how dopamine, norepinephrine, and BDNF respond to challenges—and why today's comfort zone becomes tomorrow's limitation. Find your optimal growth zone.