Some thoughts on analog thinking in a digital world — and why I still reach for a corkboard when I need to connect ideas across domains.
Ten minutes in a deep squat before touching the bar — two drills that target exactly what the snatch demands in the catch.
Breaking down the three phases of the snatch — the pull, the turnover, and the catch — with cues that actually stick.
Two lifts performed as one. Breaking down the clean's front rack and the jerk's split — the positions, the cues, and where beginners lose the lift.
Weights, loss functions, and gradient descent — explained using 1-rep maxes, WOD times, and logbooks. No prior ML knowledge required.
Three explicit iterations walked through step by step — weights, errors, and updates — using the same 1RM dataset as the linear regression post.
An animated walkthrough of how programmatic advertising works — from empty ad slot to rendered creative, all before the page finishes loading.
How variable reward loops, algorithmic amplification, and notification engineering combine to make platforms compulsive by design.
Why advertisers pay per impression rather than per sale, what targeting precision is worth, and how conversion rates justify the premium.
What ad inventory actually is, why session length is a direct revenue metric, and why every design choice that keeps you scrolling is also a financial decision.
How Amazon became the third largest ad platform, why retailer purchase data commands a premium, and what closed-loop measurement means for the future of advertising.
How two companies share analysis without sharing data — the infrastructure behind modern advertising measurement, incrementality testing, and post-cookie attribution.
Producers, partitions, offsets, and consumer groups — the three animated diagrams that explain why Kafka is a log and not a queue.