Changing our
perspectives on AI
Most of software and product engineering has been forced into a rigid structure and relegated to basic CRUD tasks.
Meanwhile, most of the providers have been threatening that software development has already been replaced by AI.
But that just means most companies are not properly leveraging their most valuable assets.
Despite all the hype, it's the people who hold the keys. If we hand off the tedious work to AI agents, we unlock capacity for humans to solve our most interesting business problems.
Mastering the context window. Every tool has the same guardrails.
Whether through prompts, agents, MCP services, or desktop apps, they all have one job — tokens in, tokens out.
Tokens in
Context window
The model reads input tokens one by one and produces output tokens in response.
Tokens out
The augmentation skills matrix.
Whether through skepticism, perceived lack of capacity, or concerns about job security, most developers are operating at Level 0 — and many are reluctant to collaborate on their AI strategy, because the industry has been threatening that they are no longer needed.
Prompt Engineering
Get really good at typing things into a browser window.
Spend most of your day swearing at ChatGPT for looping the same suggestions over and over.
Context Engineering
Tells an agent what to know and what to build.
Spend most of your day swearing at Cursor for asking you to edit every file and looping the same two bugs over and over.
Intent Engineering
Tells agents what to want and why it matters.
Codifying purpose: define goals, tradeoffs, decision trees into an approach that agents can act on.
From developer to product engineer.
Operating at Level 0 isn't a character flaw — it's the natural reaction to an industry that has spent two years telling developers they're about to be replaced.
But the role isn't disappearing. It's expanding. The new space is called Product Engineering, and it is arguably more important than what came before.
Product Engineers don't compete with agents. They direct them — shaping intent, owning outcomes, and turning judgment into the most valuable thing on the team.
The shift
- Writing every line by handDesigning the system that writes the lines
- Owning the keyboardOwning the intent, the guardrails, and the outcome
- Reviewing your own pull requestsReviewing the work of an agent team you orchestrate
Specification Mastery.
Building documents, specs, guardrails, guidance, and the right code so autonomous agents can iterate over longer time horizons without supervision or intervention.
- Agent teams
- Orchestration
- MCP Integrations
- Custom Tooling
- Defined Outcomes
Outcomes
- Strict team focus on solving the most interesting problems
- Near zero time spent on non-value added work or routine tasks
- Get the actual results you need out of AI teams and very few hallucinations
- Changes take hours, not weeks
The product engineering operating system.
- /agents
- /skills
- /commands
- /mcp
- /hooks
- Smart Index
- Second Brain
- Product Canvas
- PRD Tooling
Context mastery
Budget: 1M tokens
More than 95% available for work.
- Business Logic
- Source Code
- IaC
- DevSecOps
- Design System
- Telemetry
- User Experience
- Test Suite
The magic of a well-tuned product engineering ecosystem is that we can define substrates, jobs, conditions, and indexes that significantly reduce overhead on the context window — leaving much more headroom for processing and problem solving.
CTOforge preaches AI augmentation.
- A seasoned product and technology executive here to guide your team on how to build a sustainable Product Engineering Operating System.
- A developer experience and ergonomics coach to collaborate directly with teams on how to develop these new skills and adapt to this new environment.
- A visionary business leader supporting integration of this tooling without handing most of your gross margin to LLM providers.
- An agents-as-a-service firm that “handles everything” by installing themselves between you and your technology.
- Incentivised by burning the most tokens possible in the shortest amount of time.
- Prescriptive in our approach. Every team needs tailored, customized tooling that they can iterate and improve upon.
Your technology is at your core, but it's not your identity.
Let's talk about the company you're actually trying to build.