Build logBuild Logs
I Gave AI Agents a Real SaaS Build. Here Is Where It Broke.
Agents wrote most of the code. They did not write the parts that mattered most, and the gap between those two facts is the whole story.
Hamza
Build Logs6 min read
The Practical AI Coding Publication
Hamzify
Real tools. Real projects. Real results.
Build log #02
Why this exists
Marketing does not explain how these tools behave on real work.
Hamzify tests AI coding tools, documents builds, and records both what worked and what failed. The point is a practical answer, not a claim that every product is worth using.
Stated method
Every test says what was built, which version was used and how long it ran. You can judge how far the result transfers.
Failures included
Build logs record the decisions that went badly. A write-up where nothing went wrong is a write-up that left something out.
Corrections in the open
When something here turns out to be wrong, it gets fixed on the page with a note saying what changed.
Start here
New to Hamzify? Pick the question you actually have.
01If you are evaluating AI coding tools
Reviews and comparisons02If you want to see what happens in real projects
Build logs03If you are experimenting with AI-assisted development
Vibe coding04If you want a process you can run again tomorrow
Workflows
Editor's picks
A few places to start reading
Recent
Recent experiments and AI build logs
What happened when real work was handed to an AI coding model. Experiments and agent build logs, written with the setup and the limitations attached.
Experiment #02
Can an AI agent build a complete Chrome extension from one prompt?
- Result
- Partially successful
· 5 min read
Build logBuild Logs
Rebuilding My Portfolio With AI: A Weekend Build Log
A build log of rebuilding a developer portfolio in a weekend with AI writing most of the code — the design decisions that had to stay human, and the performance work that got missed.
Build Logs5 min read
Prompt to Deployed Product in One Evening: What Held Up
Vibe Coding5 min read
AI coding tools
Tested, compared, and explained
Reviews with a stated method, comparisons decided per use case, and guides to how the categories fit together.
Comparison
CursorvsGitHub Copilot
Cursor vs GitHub Copilot: Same Four Tasks, Both Tools
There is no single winner here. There is a winner per use case, which is the only useful kind of answer.
· 5 min read
- Review
Claude Code Review: A Terminal Agent on Unfamiliar Code
Hands-on notes from using a terminal coding agent for a week on a codebase I did not write: where autonomous multi-file work paid off, and where supervision was cheaper than delegation.
5 min read - Resource
A Context Checklist for Briefing Coding Agents
A reference checklist for what an AI coding agent needs to know before it starts: the seven inputs that change output quality, and the ones that only add noise.
4 min read
AI agent build logs
The project archive
Each AI build log is a record of something actually built with coding agents: what it was for, the stack, and where it ended up.
Ship a usable invoicing tool for freelancers — auth, invoices, PDF export, Stripe checkout — with coding agents writing the bulk of the implementation.
- Stack
- Next.js · TypeScript · Postgres · Prisma · Stripe
- AI tools
- Cursor · Claude
- Status
- Shipped, still running
- Time invested
- 26 hours over 9 days
I Gave AI Agents a Real SaaS Build. Here Is Where It Broke. · 6 min read
Replace a four-year-old portfolio with a static, fast, accessible site — case studies, writing index, contact — using AI for implementation while keeping design and content decisions human.
- Stack
- Next.js · TypeScript · MDX · CSS
- AI tools
- Cursor · Claude
- Status
- Shipped
- Time invested
- 11 hours over 2 days
Rebuilding My Portfolio With AI: A Weekend Build Log · 5 min read
AI development workflow
The AI Pair Programming Loop I Actually Use
Most bad output from a coding model is a briefing problem, not a model problem. This is the loop that fixes the briefing.
Workflow5 min read
Read the workflow- 01Brief
- 02Constrain
- 03Generate
- 04Verify
- 05Integrate
Reference
Resources worth keeping open
Checklists and guides updated in place. Use them while you work, not after.
- GuideResources
The AI Coding Toolbox: What Each Category Is Actually For
A guide to the AI development tool landscape organised by the job each category does — editors, agents, CLI tools, review bots and model routers — and how to tell when you need one.
HamzaResources5 min read
The lab
Coming to the lab
Nothing here is built yet. These are the three utilities on the list, described honestly as plans rather than products.
- Planned
Stack picker
Answer six questions about a project and get a defensible starting stack, with the trade-offs named.
Built because 'which stack should I use' has a decent answer once you know the constraints — team size, hosting, whether you need a database, how much you care about cold starts.
- decisions
- architecture
- Planned
Agent brief builder
Turn a rough task description into a structured brief a coding agent can act on.
The context checklist from the resources section, as a form: goal, acceptance criteria, files to imitate, scope limits and constraints, copyable as markdown.
- agents
- prompting
- Planned
Diff review prompts
Generate review questions tuned to the specific failure modes of generated code.
Paste a diff summary, get the questions worth asking about it — duplication, boundary conditions, error paths — rather than a generic checklist.
- code-review
- quality
Newsletter
One useful AI development workflow at a time.
New experiments, tool notes and workflows when they are worth sending. There is no fixed schedule. If email is not your thing, the feed carries everything the newsletter does.