docs: consolidate README aims from three bullets to two

Merge "hard signals make agents reliable" into "quality is a system
property" — they were saying the same thing from different angles.
Keep "do big things with small models" as a consequence, not a
separate thesis.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Author: FL4TLiN3

README.md

@@ -19,9 +19,8 @@ If you want to build practical agentic apps like Claude Code or OpenClaw, a harn
 
 Perstack is a harness for agentic apps. It aims to:
 
-- **Do big things with small models**: If a smaller model can do the job, there's no reason to use a bigger one.
-- **Quality is a system property, not a model property**: The harness provides the quality floor—hard signals that guarantee the output works. You provide the quality ceiling—domain knowledge that ensures it solves the right problem. Neither requires a frontier model or an AI science degree.
-- **Hard signals make agents reliable**: Output quality depends on verification, not model capability. When each micro-agent's output can be checked by a compiler, test suite, or screenshot diff—independent of LLM judgment—the system self-corrects and converges instead of oscillating.
+- **Quality is a system property, not a model property**: The harness provides hard signals — compiler output, test results, screenshot diffs — that let agents detect and fix their own mistakes. You provide the domain knowledge that defines what "correct" means. The combination sets both the floor and the ceiling.
+- **Do big things with small models**: When the system handles verification, a focused agent on an affordable model outperforms a generalist on a frontier one.
 
 ## Getting Started