Build Before You're Ready: Andrey Marey on High-Agency, Discipline, and Refusing to Use AI With Friends

Andrey Marey has been building since before most people his age finished a college essay. By twenty he had built fraud detection systems in London, won an NVIDIA hackathon, lived in a hacker hotel in Finland, and shipped three different startups. He's a junior at Santa Clara studying math and computer science, but he calls himself something more specific. High-agency.

In this Still Human conversation, Andrey breaks down what that actually means. It's not motivation. Motivation is a dopamine spark that fires and dies. Discipline is the thirty-second pause before the scroll, where you catch yourself and choose something else. He talks about training that pause for months. He talks about setting impossible goals on purpose because your brain operates differently when the bar is real. And he talks about the line he refuses to cross: he will never use AI to text his friends, his family, or a partner. The moment you outsource emotional intelligence to a machine, he says, you've granted someone else access to your life.

What High-Agency Actually Means

Most people use "high-agency" as a vibe — Twitter shorthand for ambitious. Andrey treats it as the operating system itself. The cleaner definition: motivation is a dopamine spark that fires and dies, discipline is the work of building habits, high-agency is the structural choice to act before you feel ready. It's distinct from confidence. You can be terrified and still high-agency. The act is what defines it.

The 30-Second Pause

The smallest unit of discipline isn't a routine — it's the gap between an impulse and the action that follows it. Thirty seconds between wanting to open Instagram and tapping the icon. Andrey trained that pause for months. The skill isn't never wanting to scroll. The skill is reliably noticing the want and choosing again, every time, until the chain of small choices compounds into a different life.

Setting Impossible Goals on Purpose

Most goal-setting advice tells you to be realistic. Andrey argues the opposite for one specific reason: the brain literally operates differently when the bar is real. Aiming at something genuinely beyond your current capacity changes which actions feel relevant in the meantime. The realistic goal lets you optimize around what you already do. The impossible one forces you to invent the actions that make it possible.

Ideas → Execution → Distribution

Most founder advice ranks these wrong. Ideas matter. Execution matters more. Distribution matters most. The trap is assuming a good idea, well-built, will find its audience. It almost never does. Andrey's hierarchy isn't a slogan — it's a directive about where to spend your unscheduled time. Most of it should go to distribution, not the thing you're already comfortable with.

The Line He Won't Cross

Andrey uses AI heavily for technical work. He refuses to use it to text his friends, his family, or a partner. The moment you outsource emotional intelligence to a machine, you've granted it access to the most personal layer of your life. The cost is invisible at first. By the time you notice the warmth has gone out of the conversations, you've already trained yourself out of being someone who can produce it.

Pain as Operating System

Two knee surgeries before he was 18. Forced stillness, weeks of it, before he had any framework for what to do with it. He doesn't pretend the pain was good. What he does say is that what he built during and after it became the operating system he still runs. Most operating systems for ambition come from someone else's book. Andrey's came from a hospital bed.

Show Notes

Andrey Marey is a 20-year-old junior at Santa Clara University studying math and computer science, and a serial founder with a track record that runs through fraud detection work in London, an NVIDIA hackathon win, a stint at the FR8 hacker hotel in Finland, and three shipped startups including Vivora (a NotebookLM competitor) and a Bionic Reader project. He left Vivora before it raised funding and left FR8 early — both decisions he describes honestly rather than dressed up. His framework for choosing what to build is short: ideas matter, execution matters more, distribution matters most. His framework for how to build is shorter: discipline over motivation, and never automate the parts of your life that require you to be a person. For the Still Human audience, Andrey is the guest who turns "high-agency" from a Twitter slogan into a working operating system.

Articles & Research

No external research was cited in this episode.

Tools & Resources

Relevant to this episode:

• Vivora (NotebookLM competitor) — One of Andrey's startups, referenced as the project he chose to leave before it raised funding
• Bionic Reader project — Earlier project where Andrey explored reading-acceleration technology
• FR8 (Finland) — The "hacker hotel" Andrey lived in; he describes leaving early as being "too early" rather than dressing it up
• NVIDIA hackathons / developer programs — The competitive AI environment where Andrey won; part of the early track record that shaped his execution-first instincts
• Engineering-led sales — Andrey's framing for why technical founders need to learn distribution, not just shipping

Related Still Human Episodes

You might also enjoy:

• AI Is the Ultimate People Pleaser — Bailley Georgieva on Hypersonics & Critical Thinking — oshenstudio.com/episode/ai-people-pleaser-bailley-georgieva-hypersonics
• Why AI Can't Reason Over Time — Michael Iwashima on BCIs, Biosensors & Building From Nothing — oshenstudio.com/episode/brain-computer-interfaces-michael-iwashima-ai-limits
• Execution Culture — Sean Wu on Raising $2M for Robotics & the Sim-to-Real Gap — oshenstudio.com/episode/execution-culture-sean-wu-synphony-robotics

People Mentioned

No additional people were cited by name in this episode beyond the host and guest.

Key Takeaways

• High-agency is the operating system, not the slogan. Andrey distinguishes it sharply from motivation. Motivation is a spark that dies; high-agency is the structural choice to act before you feel ready.
• Discipline is a 30-second pause. The window between the urge to scroll and the next click is where discipline actually lives. He trained that pause for months.
• Set impossible goals on purpose. The brain operates differently when the bar is real. Aiming for something genuinely hard changes the actions you'll take in the meantime.
• Ideas → execution → distribution. Ideas matter. Execution matters more. Distribution matters most. Most founder advice gets the order wrong.
• Pain becomes the operating system. Two knee surgeries before 18 forced a stillness that reshaped how Andrey thinks about time and goals. He runs on what came out of that.
• Don't automate being a person. AI for technical work, yes. AI for messages to friends, family, or a partner — no. The moment you outsource emotional intelligence, you've granted someone else access to your life.
• Creators get replaced for a beat, then come back stronger. Andrey's contrarian take on the AI-and-creators debate: short-term displacement, decade-long resurgence.

In This Episode

• What "high-agency" actually means — Andrey's distinction between motivation, discipline, and high-agency, and why the difference is operational rather than philosophical
• The 30-second pause — How Andrey trained the gap between impulse and action, and why that's the smallest unit of discipline
• Setting impossible goals on purpose — His case for aiming high enough that your brain genuinely operates differently
• The framework: ideas → execution → distribution — The hierarchy he uses to choose what to build, and why most founder advice inverts it
• Two knee surgeries before 18 — How a forced stillness shaped the way he runs his career
• NVIDIA, London, FR8, Vivora, Bionic Reader — The actual track record behind the philosophy
• Why he left Vivora before the raise — His honest answer on respect for other people's time
• Why he left FR8 early — "Too early" instead of dressing it up — and what that taught him about reading rooms
• The line he won't cross — AI is never allowed to text his friends, family, or a partner. The argument for why
• Creators in the AI era — Replaced for a beat, back stronger within the decade
• The signature answer — The most successful people of the next decade will be the ones who use technology to deepen human relationships, not replace them

About Andrey Marey

Andrey Marey is a 20-year-old junior at Santa Clara University studying math and computer science, and a serial founder with three shipped startups, a NVIDIA hackathon win, fraud detection work in London, and a stint at the FR8 hacker hotel in Finland behind him. He calls himself high-agency — and uses it operationally rather than as a vibe. His framework for what to build is ideas matter, execution matters more, distribution matters most. His framework for how to build is discipline as a 30-second pause and goals set deliberately out of reach. He left Vivora (a NotebookLM competitor) before it raised funding, and left FR8 early because he was, in his own words, "too early" — not because he failed. He carries an operating system shaped by two knee surgeries before 18, and a hard line: AI is allowed in his technical work but never in his messages to people he loves. For the Still Human audience, he's the guest who turns the high-agency conversation into something concrete enough to use this week.

Connect With Andrey Marey

• Watch the episode: youtube.com/watch?v=ZWhkF8q3-g0
• Website: andrey.marey.com
• Instagram: @andrey.marey

Follow Still Human Podcast

Still Human Podcast is a biweekly show by Oshen Studio, hosted by Perkin — exploring what it means to stay human in the age of AI. Real conversations with builders, creators, founders, and thinkers doing it in real life.

• YouTube: youtube.com/@Oshen.studio
• Instagram: instagram.com/perkin0909
• Listen everywhere: Search Still Human Podcast on Spotify, Apple Podcasts, or wherever you listen

New episodes drop every two weeks. Subscribe so you never miss a conversation.