Writing & Videos — Page 5
Page 5 of 14: older articles, tutorials, and videos.
June 2026
Jumprope Tracker - Demo & Case Study
I built a minimal jump rope tracker app with an AI coach that reads your workout data, goals and streaks through tool calls to give personalized insights and automatic weekly digests. I walk through the features —…
"Who cares what the code looks like?" is dumb
I argue that “if it works, who cares what the code looks like” is a dangerous standard for agentic coding because it assumes the agent perfectly understands the codebase, has complete context, and won’t hallucinate bad…
Build deterministic protections into your workflows
Without deterministic protections, what you have isn't an AI agent — it's a ticking time bomb that goes off the moment the LLM hallucinates. I define deterministic protections as plain code or scripts you write…
Not knowing how to code is not an advantage
I argue that not knowing how to code is not an advantage, even when AI agents can write code for you. I compare it to filmmaking: better tools like a smartphone and editing software don’t remove the need to understand…
LLM's $3 Trillion Dollar Problem
LLMs are becoming a commodity, and I argue that the harness around the model — its UX, the context it gets, and the tools it can use — matters far more than which model you pick. I show why benchmarks are misleading…
I reviewed this ecommerce store and gave them some advice (Karatas)
I review the Karatas ecommerce store and show why its biggest conversion problem is that customers have no way to buy anything: products have no prices, no add-to-cart or buy-now option, and no checkout, only a…
AI is destroying your skills
I argue that heavy AI use atrophies your critical thinking and coding skills through what I call cognitive surrender: you stop thinking through problems and let the model decide for you. I explain why human review…
Tokenmaxxing is a failed experiment
Token maxxing is a failed experiment: ranking engineers by how many LLM tokens they burn instead of what they actually ship destroys budgets and produces nothing. I walk through the numbers — Uber blowing its entire…
Antigravity 2.0 is trash
Antigravity 2.0's agent harness is a bare chat box with an LLM behind it, and no real developer can ship production-grade software in it. I go through what's gone: you only see files changed in the current chat rather…
May 2026
AI layoffs are here. This is how you survive.
I break down the absurd survival tactics developers are using as AI layoffs hit—voice-dictating all your code with WhisperFlow, automating a coworker's entire job so you look indispensable, and peppering managers with…
I just proved why AI can't automate anyone
AI only 'automates' experts, because it takes an expert to prompt it correctly, supply the right context and steer it to a finished result — and if a human expert still has to run the agent, nothing has been automated…
Tokenmaxxing needs to stop
Token-usage leaderboards are a terrible way to measure employee AI productivity, and I argue companies should stop using them. I point to Meta's internal token leaderboard, where employees burned throwaway prompts to…
Dude, you’re at the bottom of my landing page.
Just book a call already if you’re that interested.
You scrolled all the way here.