I started as an intern grabbing coffee for Lil Wayne. Ten years later, I was producing for Beyoncé, Nicki Minaj, and Jason Derulo. Platinum records. Long nights. But not sustainable with a family. So I moved from the studio into the business side - major labels. A&R.

I’m convinced that AGI won’t come from a single “perfect” model, but from multiple models looping together, challenging and refining each other’s answers in real time. Think about this…half the time you spend with LLMs is reminding them to avoid assumptions, check their work,

"Born in the USA": How "Springsteen: Deliver Me from Nowhere" Captured Lightning in a Bottle Listen to the full Filmmaker Toolkit podcast on the film, in theaters now from @20thcentury https://t.co/xcuU7I3gKn

Who needs a release report

What do you obsess over?

4 Main Approaches to LLM Evaluation https://t.co/h5vYG7AGaI

turns out, senior engineers accept more agent output than juniors. this is because: - they write higher-signal prompts with tighter spec and minimal ambiguity - they decompose work into agent-compatible units - they have stronger priors for correctness, making review faster and

Day 34… Happy thanksgiving

I finally found the clearest way to explain how modern AI systems are actually built… Most people talk about “agents” like they’re magic. In reality, it’s a simple stack: 1. Prompts The smallest unit. They tell the model how to think. 2. Tools Functions the model can call to

Day 33

made in America motherfuckerrrrrrrrrs

do girls vibecode? asking for my bro

Day 33

Day 32…trains

Every time I hear the first line of Billie Ellish birds of a feather I think of Big Sean I don’t fuck with you. 😭

We really need to replace the whole music industry plumbing with stablecoins. Totally invisble to everyone. not speculative. Just instant payments, splits, transparency, attribution.

Spoiler alert it got past the error

Day 30…new phone

The start of something new

Over the past year, talking with labels, distributors, and music companies, one thing is clear: there is already an AI capability overhang in the industry. The technology is ready, but adoption is not. Most of the roadblocks are operational. Imagination: teams do not yet see

A quick thought for anyone looking for a job… We’re in a strange moment where 90% of non-tech candidates either: • Think AI will replace them • Think it’s going to destroy humanity • Or they’re curious but “too busy” to learn it This is your edge. It’s a massive arbitrage

Day idk im sick