How I think about building from zero
The useful version is not romance. It is choosing a narrow problem, making the first version work, finding the first users, and cutting everything that distracts from the loop.
Eugene Bochkov
I build software companies and write about the work.
Software companies, AI, financial systems, and the messy path from idea to something people actually use.
person.file
default mode
Eugene Bochkov
Builder with product taste and engineering patience. I am interested in software that changes how money, work, information, or markets move.
current lane
Software, markets, writing
This site collects writing, projects, and public links around the work I keep returning to: software companies, AI, financial systems, and markets.
latest.notes
Financial software is constraint design
Compliance, liquidity, risk, UX, and distribution are not separate tracks. Good financial software makes those constraints usable.
AI needs a job, not a demo
The difference between a gimmick and something useful is whether the model improves work someone already cares about.
proof.signals
$20M TVLfinancial systems built from early stage
10k MAUsoftware scaled from zero
AI before hype cycleconversation systems and workflow automation
10+ yearssoftware companies, markets, and product work
consulting.signal
book a consultation
Software company strategy, AI, financial systems, and market positioning.
Useful if you are working through a new software idea, an AI product, a financial system, a market entry question, or the shape of a company before it becomes obvious.
writing.index
Medium
Building a conversational AI startup from zero to one
Substack
Newsletter archive on software companies, AI, and markets
Substack article
Essay from Eugene Finch Substack
Substack article
Essay from Eugene Finch Substack
note / building
Building from zero is mostly subtraction.
The early version of a company is usually too loud. Too many features, too many explanations, too many imagined users. The work is to cut it down until the real loop is visible: problem, user, workflow, reason to return.
I keep coming back to this because most software does not fail from lack of ideas. It fails because nobody can tell what the product is actually asking the market to do differently.
note / ai
AI is only interesting when it changes the work.
A model demo is cheap. A workflow that becomes faster, clearer, or possible for the first time is not. That is the difference I care about.
The useful questions are boring on purpose: who uses it, what does it replace, what does it make measurable, what breaks when the model is wrong, and why would someone keep using it after the novelty disappears.
note / financial systems
Financial software has no clean room.
Money products sit inside constraints: trust, risk, liquidity, compliance, distribution, user fear, and bad incentives. The interface is only the visible part.
Good financial software does not pretend those constraints are separate. It makes them legible enough for someone to use the product without understanding the entire machine underneath.
topics.map
software companies
AI
financial systems
markets
capital
distribution
company building
writing