Human judgment. Intelligent systems. Ambitious software.
We ♥ demanding problems.
We use AI to understand software and build better ways to create it.
We assess codebases, examine how advancing AI changes their technical defensibility, and build tools for engineering and collaboration. Through the Collaboratory, we develop products with people who want to become AI builders. Through our sponsorship of Big Lever Institute, a nonprofit, we support free practical AI education to broaden access to what AI makes possible.
Our work is about finding problems worth solving and bringing solutions to market; we're equally interested in the problems our partners bring us.
01 / SOFTWARE ASSESSMENT
Software assessment
Understanding software requires examining both its present condition and what changing technology makes possible.
01
Codebase assessment
We examine source code, architecture, dependencies, configuration, and development history to establish how a system is built, where its weaknesses lie, and what addressing them would require.
The work covers security, maintainability, test quality, technical debt, and constraints on growth. Findings carry their supporting evidence, independent confirmation, and human review. The assessment also identifies what could not be examined and how those limits affect the conclusions.
02
Engineering assessment of AI obsolescence
Advancing AI changes which software capabilities are difficult to build and which can be reproduced more readily.
We assess where existing software is exposed, where technical advantages remain, and what adaptation or reengineering would require. That includes examining reliance on AI models and whether a product's capabilities amount to more than what those models can already provide.
This is an engineering assessment of software capabilities and technical defensibility. It is distinct from a market assessment or financial valuation.
Big Lever AI · Illustrative report specimen · Select to enlarge
Applied in Big Lever AI
Our flagship offering BigLever.AI applies our software-assessment capabilities to the decisions facing software acquirers and owners. It combines technical evidence with investor and operator judgment to address software condition, repair exposure, and the implications of advancing AI.
We build tools that support our own work and make them available for others to use.
Methodical · Illustrative workflow · Select to enlarge
Methodical Gated Engineering
An engineering workflow for Claude Code that combines deliberate planning, assumption testing, independent review, and human judgment at key decisions.
Methodical gives people and AI agents a shared structure for deciding what to build, testing assumptions, and reviewing the result.
Commons · Fellow Commons board · Select to enlarge
Commons
A platform for communities where people float ideas, find collaborators, and organize around projects they choose to pursue. It supports distributed decision-making, with ideas earning support through their merits and people's willingness to contribute.
Commons supports this way of working, and we make it available for others to use.
AI is part of our daily work: development, code review, research and operations. We adopt new capabilities as they become practical, and evaluate them through the work they can perform.
That includes choosing models for particular tasks, managing token use and cost, and designing secure workflows that preserve access to frontier capabilities. Human judgment remains central to setting direction, interpreting evidence, and approving consequential decisions.
The tools we share and the products we develop come out of that practice.
Build with us through the Collaboratory
The Collaboratory brings people together to turn ideas into working products. Choose a project, bring your experience, and develop new capabilities by building alongside others. You can contribute to an existing venture or find collaborators for an idea of your own.
People take on responsibilities and share in the performance of the products they help create.
