I get asked the same questions often. Instead of writing the same email repeatedly, I’m putting my standard advice here. Take what’s useful, ignore what isn’t.

Disclaimer: This is what worked for me. Your situation is different. Adjust accordingly.

💻 For Engineers

“How do I become a senior engineer?”

The short answer: Ship things, take ownership, and learn to communicate.

The longer answer:

  1. Own outcomes, not just tasks. Juniors complete tickets. Seniors own features. Staff own systems. The level is about scope of responsibility, not years of experience.

  2. Learn to write. Seriously. Technical writing, design docs, RFCs, postmortems. The ability to articulate complex ideas clearly is rare and valuable. Practice.

  3. Understand the business. Why does your company exist? How does it make money? What problems does your code actually solve? Engineers who understand business context make better decisions.

  4. Mentor someone. Teaching forces you to understand deeply. It also builds your reputation and influence.

  5. Pick a specialty. Generalists hit a ceiling. Having deep expertise in one area + broad knowledge elsewhere is the winning formula.

What doesn’t matter as much as you think:

  • Language purity (knowing 10 languages shallowly vs. 3 deeply)
  • Algorithm trivia (useful for interviews, not for most actual work)
  • Always being right (it’s more important to be effective in a team)

“Should I specialize or stay generalist?”

Specialize, but with breadth.

The best career shape is T-shaped: deep expertise in one area, broad familiarity across many. Being “okay at everything” becomes a liability after mid-level.

How to pick a specialty:

  • What do you enjoy enough to do for years?
  • What’s in demand and likely to stay that way?
  • Where can you develop genuine expertise faster than others?

The intersection of these is your specialty.

For now (2026): AI/ML, systems programming, and security are high-value specialties. Frontend frameworks are a race to the bottom.

“How do I get into AI/ML?”

Two paths:

Path A: The Research Path

  • Take Andrew Ng’s courses (they’re still good)
  • Read papers (start with “Attention Is All You Need”)
  • Implement papers from scratch
  • Contribute to open source projects
  • Target research labs or ML-heavy companies

Path B: The Applied Path (what I did)

  • Learn enough theory to understand what’s happening
  • Focus on deployment, scaling, and productionization
  • Learn the ecosystem: LangChain, vector DBs, fine-tuning pipelines
  • Build projects that use ML to solve real problems
  • Position yourself as “the person who makes ML actually work in production”

Path B is faster and has strong demand. Most companies need more ML engineers than ML researchers.

What to learn first:

  1. Python (well, not just “enough to copy examples”)
  2. PyTorch or TensorFlow (I’d pick PyTorch)
  3. Transformers and attention mechanisms
  4. RAG patterns and vector databases
  5. LLM orchestration (LangChain, LlamaIndex)

📝 For Writers

“How do I start writing?”

Just publish something. Today. It doesn’t have to be good.

The first 10 posts will be bad. That’s fine. The only way to develop voice is through practice. Waiting until you’re “ready” is procrastination.

Practical steps:

  1. Pick one topic you know well
  2. Write for 30 minutes
  3. Edit for 15 minutes
  4. Publish
  5. Repeat

On frequency: Don’t commit to daily or weekly if you’ll burn out. Better to publish monthly for years than weekly for months.

On voice: It takes 50+ posts to find your voice. Stop trying to sound like your favorite writer. You’re not them. Write like you talk.

“How do I get readers?”

The secret: There is no shortcut. It’s quality + consistency + time.

What actually works:

  • Write things people want to share (contrarian takes, surprising insights, useful frameworks)
  • Cross-post to Twitter/X, LinkedIn, relevant communities
  • Engage with other writers in your niche
  • Guest post if invited
  • Be patient — it takes years

What doesn’t work:

  • SEO gaming (works short-term, dies long-term)
  • Clickbait titles (destroys trust)
  • Growth hacks (exhausting and temporary)

My honest take: don’t optimize for readers at first. Write to think, write to clarify, write for yourself. If the writing is good, readers come eventually.

🏋️ For Fitness

“How do I start working out?”

Start ridiculously small.

Don’t commit to 6 days a week. Commit to going to the gym twice this week. Just show up. Do something. Leave.

The beginner program I recommend:

3 days per week, full body:

  • Squat (or leg press)
  • Bench press (or dumbbell press)
  • Row (or lat pulldown)
  • Overhead press (or machine)
  • Deadlift or RDL

3 sets of 8-12 reps each. Add weight when you can complete all reps with good form.

That’s it. It’s not complicated. The program matters less than consistency.

The real challenge isn’t the program — it’s showing up when you don’t feel like it. That’s the only thing that matters.

“How do I build muscle naturally?”

The unsexy truth:

  1. Progressive overload. Add weight or reps over time. If you’re not progressing, you’re not growing.

  2. Protein. 0.8-1g per pound of bodyweight. Every day. Not negotiable.

  3. Sleep. 7-9 hours. This is when you actually grow. Cutting sleep is sabotaging your gains.

  4. Consistency. Years, not months. The people with great physiques have been training for 5-10+ years.

  5. Patience. Realistic natural gains: 15-25 lbs of muscle in your first year, then diminishing returns. Anyone promising faster results is lying or selling something.

What doesn’t matter as much:

  • Supplements (except creatine, which actually works)
  • Optimal program selection (any reasonable program works)
  • Meal timing (total daily intake matters more)

🚀 For Builders

“Should I quit my job to build a startup?”

Probably not yet.

Only quit when:

  • You have 12+ months of runway saved
  • You have validated demand (people paying or strong signals)
  • You can’t make meaningful progress on the side
  • You’ve accepted that it will probably fail

Until then:

  • Build on evenings/weekends
  • Validate with real users
  • Get to some revenue before quitting
  • Use your job’s stability to take risks with the side project

The romantic “quit and go all in” narrative makes for good stories but bad advice. Most successful founders I know built on the side first.

“How do I validate an idea?”

Talk to people. Not friends. Not family. Potential customers.

The Mom Test (from the book of the same name):

  • Don’t ask “Would you use this?” (they’ll say yes to be nice)
  • Ask “How are you solving this problem today?”
  • Ask “What have you tried? What didn’t work?”
  • Ask “When’s the last time you looked for a solution?”

Signs of real demand:

  • People have already tried to solve the problem
  • People have already paid for solutions
  • People ask when they can buy (not “let me know when it’s ready”)

Signs of fake demand:

  • “That’s a cool idea!” (means nothing)
  • “I’d definitely use that!” (also means nothing)
  • No one is currently paying for alternatives

The best validation is a sale. Can you get someone to pay before you build? Pre-sales, LOIs, or even paid pilots are the strongest signal.

🧠 For Life

“How do I figure out what to do with my life?”

You don’t figure it out — you try things.

There’s no amount of introspection that replaces actually doing. Run experiments. Try different work, different hobbies, different lifestyles. Notice what energizes you vs. what drains you.

The framework I use:

  • What do I enjoy enough to do even when it’s hard?
  • What would I do if I had enough money?
  • What do people come to me for?

The intersection of these is probably meaningful to you.

What helped me:

  • Reading widely (exposes you to possibilities you didn’t know existed)
  • Talking to people with different lives (same reason)
  • Actually trying things instead of just thinking about them
  • Accepting that “figuring out your life” is ongoing, not a one-time event

“How do I deal with burnout?”

Recognize it first. Burnout isn’t tiredness. It’s loss of meaning combined with exhaustion. You can sleep for a week and still feel burned out.

What actually helps:

  1. Time off. Real time off. Not “working from home.” Not “just checking Slack.” Actually off.

  2. Exercise. The research is clear. Physical activity helps. Even a walk is something.

  3. Address the cause. If you go back to the same conditions, you’ll burn out again. What needs to change?

  4. Lower the bar temporarily. In burnout, “good enough” is the goal. Excellence can wait.

  5. Get help. Therapy, coaching, talking to someone. Burnout is lonely. It doesn’t have to be.

What doesn’t help:

  • Pushing through (makes it worse)
  • Meditation alone (helps, but doesn’t fix root causes)
  • Vacation followed by returning to the same situation

See also: my note on Why Engineers Burn Out

📫 Still Have Questions?

If your question isn’t answered here, email me. I read everything, though I can’t promise to respond to everything.

Before emailing:

  • Search this site — I may have written about it
  • Be specific — “How do I succeed?” is unanswerable
  • Tell me what you’ve already tried

This page is a living document. I add to it when I notice I’m giving the same advice repeatedly.

See also: My Personal Rules | What I Believe | What I’m Doing Now