📊 Full opportunity report: The Model Is Only 10%: The Real Lesson of the New SDLC on ThorstenMeyerAI.com — validation score, market gap, and execution plan.
TL;DR
A recent whitepaper from Google emphasizes that in AI-assisted software development, the core value lies in how developers configure and control AI systems, not just the models themselves. The model is only 10% of the system, with harness and context engineering representing the remaining 90%.
A new Google whitepaper released in early 2026 states that the most significant factor in AI-driven software development is not the AI model, but the harness and context engineering. This insight challenges the common focus on models, emphasizing instead the importance of configuration, verification, and judgment in AI systems.
The whitepaper, titled ‘The New SDLC With Vibe Coding,’ by Addy Osmani, Shubham Saboo, and Sokratis Kartakis, reports that 85% of professional developers use AI coding agents regularly, with 51% using them daily. It states that roughly 41% of new code is generated by AI. The key takeaway is that the model constitutes only about 10% of the system’s behavior, while the harness—scripts, prompts, tools, and policies—accounts for 90%. Evidence from benchmark experiments shows that optimizing the harness can significantly improve AI performance, even with the same model.
The model is only 10%
A Google whitepaper argues software’s biggest shift is from writing code to expressing intent. Its sharpest claim: the model you obsess over is the smallest part of the system — the scaffolding around it does the real work.
The clearest map yet of how serious AI development works — and mostly tool-agnostic. But it’s a Google funnel: the concepts are neutral, the on-ramps point to Gemini, Jules & the ADK. If the harness is 90% and it’s yours, your moat and your costs both live there — so own your scaffolding, route across models, and remember: AI amplifies whatever engineering culture it lands in.
Why Configuration and Context Matter More Than the Model
This shift has major implications for organizations investing heavily in AI models. It suggests that cost, reliability, security, and performance depend more on how systems are configured and controlled than on the raw model quality. Leaders should prioritize building robust harnesses and effective context engineering to gain a competitive edge, rather than solely chasing the latest model advancements.
YAML Made Simple: A Beginner’s Guide to Configuration and Data Structuring
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Background on the Evolution of AI Development Practices
Until early 2026, the industry largely viewed the AI model as the core driver of system performance. Recent developments, including widespread adoption of AI coding agents—used by over 85% of developers—have shifted focus toward how these models are integrated and managed. The whitepaper builds on prior insights into ‘vibe coding’ versus ‘agentic engineering,’ emphasizing the importance of structured workflows, verification, and context management in AI systems.
“The model is only 10% of what determines behavior; the harness is 90%. Focus on configuration, verification, and judgment.”
— Addy Osmani
The AI Prompt Playbook: Master AI Prompt Engineering with 140 Ready-to-Use Templates for ChatGPT, Claude, Gemini & Copilot
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Remaining Questions About Implementation and Best Practices
It is not yet clear how organizations should best structure their harnesses or what standards will emerge for effective context engineering. The precise methods for scaling these practices across different types of AI applications are still under development. Additionally, the long-term security implications of complex harness configurations remain to be fully understood.
Observability in the AI-Native Era: Leveraging AIOps to build, observe, and operate resilient systems
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Next Steps for Organizations Adopting AI Development Strategies
Organizations are encouraged to evaluate their current AI workflows, invest in building robust harnesses, and develop standards for context management. Future research and industry collaboration will likely produce best practices and tools to streamline this process. Monitoring how these shifts impact cost, security, and performance over the coming months will be crucial.
AI Model Validation & Testing: Ensuring Reliable AI Systems — Bias Testing, Robustness Evaluation & Regulatory Compliance (AI Compliance Toolkit)
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
Why is the model only 10% of system behavior?
The whitepaper explains that most of an AI system’s behavior depends on how the model is configured, guided, and integrated with tools and policies—collectively called the harness—rather than the model itself.
How can organizations improve their AI systems based on this insight?
Organizations should focus on building better harnesses, including prompts, tools, guardrails, and context management, to control and verify AI outputs effectively, rather than solely upgrading models.
What are the risks of over-relying on the model?
Focusing only on models can lead to higher costs, security vulnerabilities, and unreliable outputs, as the core control lies in how the system is configured and managed.
Will this shift change AI development practices industry-wide?
Yes, the emphasis on harness and context engineering suggests a fundamental change in how AI systems are built, maintained, and optimized, prioritizing configuration over model selection.
What remains uncertain about this approach?
It is still unclear which specific methods and standards will become industry best practices for harness design and context management, and how security will evolve with more complex configurations.
Source: ThorstenMeyerAI.com
