Jeremy Theocharis

Boring is Awesome | Co-Founder & CTO at UMH

Things I Believe

Strong opinions, loosely held.

Think in structures

  • MECE
  • Break problems down to fundamental truths. Don’t just copy “best practices” or “industry standards”. Think from the ground up

Speed compounds

  • Think carefully, then move decisively
  • Speed compounds. The reason to go fast is to learn faster, not to get more done
  • Going fast without feedback is stupid
  • Speed is not an excuse to skip the plan
  • Shipped means the target audience is actively using it, not that you released it

You own the outcome, not the task

  • If it doesn’t move the needle, don’t do it
  • Choosing what not to do is as important as choosing what to do

Your best idea might be wrong

  • Hard work beats talent. But hard work requires curiosity, and curiosity can’t be taught
  • Have strong opinions, loosely held. Change your mind with evidence
  • Mistakes aren’t failures, hiding them is
  • More interested in being right than being seen as right
  • Best work comes from following curiosity

First agree on what matters

  • Don’t argue details when the real disagreement is about the framework underneath
  • Opinions and values are choices, not facts. But once you have them, everything can be evaluated against them

Writing is thinking

Observe, don’t ask

  • If you want to know something, don’t ask for feedback. Observe behavior instead
  • If people use it wrong, it is wrong, not the people
  • Watch what people do, not what they say
  • There are no stupid questions. If someone asks one, something failed to make it clear
  • What someone says matters less than why they see it that way. Guess their intent

Products should have opinions

  • Don’t aim for “this is good.” Aim for “I want it now, when do I get it?”
  • There should be only one way to do it
  • Easy to use, hard to abuse
  • Progressive disclosure: power users use the same UI, they just go deeper. New users should never be confused by depth they don’t need yet
  • Use good products yourself. Product taste comes from what you use, not what you study

Hire for growth, not credentials

  • Two hiring answers: hell yes or no
  • Growth potential > current skill
  • Curiosity > CV
  • The best credential is something you built that others actually use
  • Most important hiring question: What have you built that wasn’t assigned to you?

The job of leadership

  • The person closest to the problem should make the decision
  • Protect employees from bullshit. Bullshit is everything that pulls people away from the problem they need to solve
  • Never build tracks you haven’t run on yourself. Otherwise you don’t know if they help solve the problem, or just create more of them
  • If you can’t evaluate the output, you can’t delegate it
  • Build the framework together, then hand it off
  • Once someone proves they understand the philosophy, and has experienced the problems themselves, they own it and can evolve it

Be authentic

  • I absolutely despise non-authentic people
  • Have an opinion, say it, defend it, accept when you’re wrong
  • Non-authentic people try to please everyone. It wastes time and produces bad decisions
  • The truth can be painful. Say it anyway
  • If your gut says something is wrong, stop. You either don’t understand it, or you shouldn’t be doing it. It’s a self-fulfilling prophecy

LLMs are amplifiers, but don’t have agency

  • LLMs amplify what you already have. Opinions, structure, frameworks make you more powerful with them
  • Without those, you probably will not get much value from LLMs
  • If you have clear opinions and don’t use LLMs, you’re leaving power on the table
  • Code generation got easier. Good engineers always made the planning, testing, and thinking explicit. That is becoming the only part that matters
  • The closer you interact with AI like you would with humans and human work processes, the better the results
  • If you give a human “write me an article about coaches,” you get crap. Same with an LLM
  • Generate, freeze, test

Boring is awesome