DEV Community

Finley Zhou profile picture

Finley Zhou

Backend developer who loves Rust and distributed systems.

Location Tokyo, Japan Joined Joined on 
Case Study: A C++ Cache Proxy That Cut Free-Tier Model Calls by 73%

Case Study: A C++ Cache Proxy That Cut Free-Tier Model Calls by 73%

Comments
4 min read
Case Study: A Free Model's C++ Token Bucket Passed 18 Tests. NTP Broke It in One Second.

Case Study: A Free Model's C++ Token Bucket Passed 18 Tests. NTP Broke It in One Second.

Comments
5 min read
Case Study: A C++ Build-Failure Triage Bot — Cache First, Recompile as the Gate

Case Study: A C++ Build-Failure Triage Bot — Cache First, Recompile as the Gate

Comments
5 min read
Case Study: A .gitignore Validator in C++ — 47 Vectors, 3 Rounds, 2 Gaps

Case Study: A .gitignore Validator in C++ — 47 Vectors, 3 Rounds, 2 Gaps

Comments
5 min read
Case Study: A Free Model's C++ Hash Map Passed 42 Tests. The Scaling Curve Found the O(n ).

Case Study: A Free Model's C++ Hash Map Passed 42 Tests. The Scaling Curve Found the O(n ).

Comments
4 min read
A C++ Config Validator for a Server That May Skip Your Job

A C++ Config Validator for a Server That May Skip Your Job

Comments
6 min read
Case Study: A Free Model's UTF-8 Validator Passed My Tests. The Spec Table Found What Both Missed.

Case Study: A Free Model's UTF-8 Validator Passed My Tests. The Spec Table Found What Both Missed.

Comments
6 min read
Case Study: A Gated Loop on Free Infrastructure — 3 Rounds to a Correct C++ Header Parser

Case Study: A Gated Loop on Free Infrastructure — 3 Rounds to a Correct C++ Header Parser

Comments
6 min read
Review Comments Are Hypotheses: An Audit of an AI Code Review

Review Comments Are Hypotheses: An Audit of an AI Code Review

Comments
5 min read
A Free Model Wrote a C++ Fuzzer. My Hand-Written Parser Failed in 40 Seconds.

A Free Model Wrote a C++ Fuzzer. My Hand-Written Parser Failed in 40 Seconds.

Comments
5 min read
A Free Model's UTF-8 Validator Passed 100k Round-Trips. The Spec Corpus Failed It.

A Free Model's UTF-8 Validator Passed 100k Round-Trips. The Spec Corpus Failed It.

Comments
5 min read
Case Study: A C++ Object Pool From a Free Model — 4 Defects, 3 Rounds, 1 Human Invariant

Case Study: A C++ Object Pool From a Free Model — 4 Defects, 3 Rounds, 1 Human Invariant

Comments
5 min read
Bit 0 vs. Bit 7: What 10,000 Random Frames Taught Me About AI-Generated Bitfields

Bit 0 vs. Bit 7: What 10,000 Random Frames Taught Me About AI-Generated Bitfields

Comments
5 min read
Case Study: A Free Model Rewrote a Config Parser. A Differential Fuzz Gate Found the Crash in 90 Seconds.

Case Study: A Free Model Rewrote a Config Parser. A Differential Fuzz Gate Found the Crash in 90 Seconds.

Comments
5 min read
34 Green Tests, 14 Alive Mutants: A C++ Mutation Testing Case Study

34 Green Tests, 14 Alive Mutants: A C++ Mutation Testing Case Study

Comments
5 min read
Case Study: A Benchmark Gate Rejected a Free Model's C++ "Optimization" — and the Noise Floor Agreed

Case Study: A Benchmark Gate Rejected a Free Model's C++ "Optimization" — and the Noise Floor Agreed

Comments
4 min read
Case Study: A Free Model Wrote a C++ Tree Hasher. The Reference Oracle Found Three Bugs.

Case Study: A Free Model Wrote a C++ Tree Hasher. The Reference Oracle Found Three Bugs.

Comments
4 min read
Case Study: A Free Model Refactored a C++ Library. An ABI Gate Rejected 11 of 12 Patches.

Case Study: A Free Model Refactored a C++ Library. An ABI Gate Rejected 11 of 12 Patches.

Comments
3 min read
A Free Model Made the Sum More Accurate. memcmp Caught the Regression.

A Free Model Made the Sum More Accurate. memcmp Caught the Regression.

Comments
4 min read
Case Study: ThreadSanitizer Had Final Say on a Free Model's C++ Cache

Case Study: ThreadSanitizer Had Final Say on a Free Model's C++ Cache

Comments
5 min read
Case Study: A Small C++ Build-Log Triage Tool and Four Merge Gates for Free Model Patches

Case Study: A Small C++ Build-Log Triage Tool and Four Merge Gates for Free Model Patches

Comments
4 min read
Runnable Examples as a C++ API Contract: A Free Model Proposes Them, a Disposable Server Rejects the Rest

Runnable Examples as a C++ API Contract: A Free Model Proposes Them, a Disposable Server Rejects the Rest

Comments
4 min read
Make AI Include Cleanups Earn Their Merge: A C++ Gate Case Study

Make AI Include Cleanups Earn Their Merge: A C++ Gate Case Study

Comments
4 min read
A Free Model Endpoint Proposed a Build Cleanup. My C++ Executor Required a Dry Run Before Touching State.

A Free Model Endpoint Proposed a Build Cleanup. My C++ Executor Required a Dry Run Before Touching State.

Comments
5 min read
The Free Model Wrote 87 C++ Fuzz Seeds. Only 6 Earned Coverage.

The Free Model Wrote 87 C++ Fuzz Seeds. Only 6 Earned Coverage.

Comments
4 min read
A Free Model Suggested a C++ Fix. The Subprocess Contract Decided Whether It Could Merge.

A Free Model Suggested a C++ Fix. The Subprocess Contract Decided Whether It Could Merge.

Comments
4 min read
A Model-Suggested C++ Fix Should Earn Its Way Through the Build Log

A Model-Suggested C++ Fix Should Earn Its Way Through the Build Log

Comments
5 min read
Free Model Endpoints Will Normalize C++ Paths. A Canonical Manifest Vetoes the Summary.

Free Model Endpoints Will Normalize C++ Paths. A Canonical Manifest Vetoes the Summary.

Comments
4 min read
A Free Model Called a C++ Patch Safe. A Symbol Manifest Made It Prove ABI Before Merge.

A Free Model Called a C++ Patch Safe. A Symbol Manifest Made It Prove ABI Before Merge.

Comments
4 min read
The Crash Router Trusted a 0.97 Confidence Score. It Needed a Vote.

The Crash Router Trusted a 0.97 Confidence Score. It Needed a Vote.

Comments
6 min read
A Free Model Endpoint Said the C++ Lock Was Hot. The Race Detector Disagreed.

A Free Model Endpoint Said the C++ Lock Was Hot. The Race Detector Disagreed.

Comments
3 min read
A Free Model Kept Returning Valid JSON With Shifting Types. The C++ Gate Refused the Ambiguous Cases.

A Free Model Kept Returning Valid JSON With Shifting Types. The C++ Gate Refused the Ambiguous Cases.

Comments
4 min read
A Free Model Endpoint Returned Numeric JSON That Was Valid and Wrong

A Free Model Endpoint Returned Numeric JSON That Was Valid and Wrong

Comments
5 min read
He Let a Free Model Propose C++ Patches. A Three-Way Build Held the Veto.

He Let a Free Model Propose C++ Patches. A Three-Way Build Held the Veto.

Comments
5 min read
A Free Model Wrote a Dangerous One-Liner. A Permit Ledger Made It Earn Execution.

A Free Model Wrote a Dangerous One-Liner. A Permit Ledger Made It Earn Execution.

Comments
3 min read
Sanitizer Stacks Look Identical Until You Normalize Them: A C++ Crash Triage Harness

Sanitizer Stacks Look Identical Until You Normalize Them: A C++ Crash Triage Harness

Comments
5 min read
A Local Disagreement Filter Stopped a Free Model Endpoint From Re-Explaining the Same C++ Warnings

A Local Disagreement Filter Stopped a Free Model Endpoint From Re-Explaining the Same C++ Warnings

Comments
6 min read
The Crash Tagger Kept Passing the JSON Contract Until It Didn't: A C++ Drift Ledger for Free Model Labels

The Crash Tagger Kept Passing the JSON Contract Until It Didn't: A C++ Drift Ledger for Free Model Labels

Comments
5 min read
A Known-Answer Canary Is the Safety Rail for Free Model Endpoints

A Known-Answer Canary Is the Safety Rail for Free Model Endpoints

Comments
5 min read
A Free Server Caught the GUI Fallback a Model Buried in a CLI

A Free Server Caught the GUI Fallback a Model Buried in a CLI

Comments
4 min read
A Free Model Endpoint Replied Twice, Then Went Silent. The Fix Was a C++ Replay Envelope, Not Retries

A Free Model Endpoint Replied Twice, Then Went Silent. The Fix Was a C++ Replay Envelope, Not Retries

Comments
4 min read
A Free Model Endpoint Returned 200 With an Empty Body. The Fix Was a Lease, Not a Retry.

A Free Model Endpoint Returned 200 With an Empty Body. The Fix Was a Lease, Not a Retry.

1
Comments 2
5 min read
Free Model Explanations Are Just Another Build Input. So I Gave Them a Replay Cache.

Free Model Explanations Are Just Another Build Input. So I Gave Them a Replay Cache.

Comments
5 min read
He Turned a Use-After-Free Into a Hypothesis Queue. A Free Server Did the Voting.

He Turned a Use-After-Free Into a Hypothesis Queue. A Free Server Did the Voting.

Comments
4 min read
A Free Model's Review Had No Evidence. The Maintainer Added a Claim Ledger.

A Free Model's Review Had No Evidence. The Maintainer Added a Claim Ledger.

Comments
4 min read
One Weak Symbol, Two Definitions: A C++ ODR Bug and a Two-Stage Free Triage

One Weak Symbol, Two Definitions: A C++ ODR Bug and a Two-Stage Free Triage

Comments
4 min read
A Free Model's Server Passed My Smoke Test, Then Hung on Shutdown. The Exit Contract Caught It.

A Free Model's Server Passed My Smoke Test, Then Hung on Shutdown. The Exit Contract Caught It.

Comments
4 min read
Free AI Endpoints Are Unreliable Dependencies. Test Them Like One.

Free AI Endpoints Are Unreliable Dependencies. Test Them Like One.

Comments
4 min read
A Free Model Answer Is a Hypothesis. Give It an Error Budget Before It Reaches Main.

A Free Model Answer Is a Hypothesis. Give It an Error Budget Before It Reaches Main.

Comments
4 min read
Free Model Endpoints Make Good Fuzz Generators and Bad Oracles

Free Model Endpoints Make Good Fuzz Generators and Bad Oracles

Comments
4 min read
Stop Asking Coding Models to Write Code. Test Whether They Can Review a Patch

Stop Asking Coding Models to Write Code. Test Whether They Can Review a Patch

Comments
5 min read
Weekly 'Game-Changer' Models Burned Me Twice. Now They Earn Production Access Through Gates.

Weekly 'Game-Changer' Models Burned Me Twice. Now They Earn Production Access Through Gates.

Comments 1
6 min read
Free AI Models in Your CI Pipeline Will Fail Silently. Build the Circuit Breaker First.

Free AI Models in Your CI Pipeline Will Fail Silently. Build the Circuit Breaker First.

Comments 1
5 min read
My README Promised Flags My CLI Doesn't Have. I Built a Drift Detector.

My README Promised Flags My CLI Doesn't Have. I Built a Drift Detector.

Comments
5 min read
I Inherited 60,000 Lines of Undocumented C++. I Read It With an AI and Assumed Every Answer Was a Lie.

I Inherited 60,000 Lines of Undocumented C++. I Read It With an AI and Assumed Every Answer Was a Lie.

1
Comments 1
5 min read
My Free Coding Model Kept Failing on Big Repos. The Fix Was a Context Budget, Not a Better Model.

My Free Coding Model Kept Failing on Big Repos. The Fix Was a Context Budget, Not a Better Model.

Comments 1
6 min read
Your Extraction Pipeline Doesn't Need a Smarter Model. It Needs a Schema Contract Test.

Your Extraction Pipeline Doesn't Need a Smarter Model. It Needs a Schema Contract Test.

Comments
6 min read
Your Free AI Model Changed Overnight. Here's a Snapshot Test Suite That Notices.

Your Free AI Model Changed Overnight. Here's a Snapshot Test Suite That Notices.

Comments
5 min read
I Stopped Trusting My Gut on New Open Models. A 30-Minute Scoring Loop Replaced It.

I Stopped Trusting My Gut on New Open Models. A 30-Minute Scoring Loop Replaced It.

Comments
6 min read
Don't Trust the Demo: A Repeatable Test Harness for Evaluating Free AI Coding Models

Don't Trust the Demo: A Repeatable Test Harness for Evaluating Free AI Coding Models

Comments 1
4 min read
loading...