Medium

Ali Süleyman TOPUZ

Pinned

In

Towards AI

by

Ali Süleyman TOPUZ

·

Oct 23

Skills That Maintain Themselves: SkillOpt, AutoHarness, and Why I Do Not Trust Either Yet

A skill that rewrites itself is a deployment. I want it to go through the same door as every other deployment.

Pinned

In

Towards AI

by

Ali Süleyman TOPUZ

·

Oct 26

Cutting Agent Token Costs in .NET: Tool Search, Output Gating and What the Numbers Really Say

Every few days another post tells me I can cut my agent’s token bill by 60%, 90%, 98%. I believe most of these numbers are real in the…

Pinned

In

Towards AI

by

Ali Süleyman TOPUZ

·

Oct 13

The Hybrid Beats Both: Routing Only the Uncertain Decisions to a Bigger Model, in .NET

A small C# router with a confidence band, a calibration step, and an escalation path. Local models only. Some of it worked, some of it…

Pinned

In

Towards AI

by

Ali Süleyman TOPUZ

·

Sep 19

Your AI Agent’s Reasoning Isn’t as Private as You Think

Do you actually know whether your encrypted reasoning trace is private? Not “probably,” not “the docs say it’s encrypted so it must be…

Pinned

Ali Süleyman TOPUZ

·

Jun 8

Agentic Architectures — Article 6: Multi-Agent Orchestration Patterns

I spent about three weeks convinced that a single agent with a really well-engineered harness could handle anything I threw at it. Article…

In

Towards AI

by

Ali Süleyman TOPUZ

·

5h ago

Your Harness Didn’t Get Better, It Got Lucky: Noise Bands for Agent Evals

A 2.5-point drop, a 20.4-point drop, and the one question that tells them apart

In

Towards AI

by

Ali Süleyman TOPUZ

·

6h ago

Measuring Review in Human Minutes: A Risk-Tier Gate for Agent Pull Requests in .NET

I stopped trying to read every line an agent writes. Here is the gate I built instead, what it cost in minutes, and the one bug that walked…

In

Towards AI

by

Ali Süleyman TOPUZ

·

2d ago

“Verified Publisher” Is Not a Trust Boundary: What the Trojanized Claude Code Extension Teaches…

For about eleven hours on 29 October, the extension marketplace inside Cursor handed out a tampered build of Anthropic.claude-code. It…

In

Towards AI

by

Ali Süleyman TOPUZ

·

2d ago

Your Sandbox Wraps a Process, but Your Agent Now Lives Behind a Service

I have been building a sandbox around my own coding agent for a few months, and I had quietly assumed the hard part was the kernel stuff…

In

Towards AI

by

Ali Süleyman TOPUZ

·

4d ago

A Decision Model as a Gate: Fast, Well Calibrated, and Still Wrong 16% of the Time

Can a small decision model like Jev sit between an agent and its tools in .NET? Here’s what other people measured, what I built to test it…

Ali Süleyman TOPUZ

Ali Süleyman TOPUZ

Staff Architect & Player-Coach. Building & writing on AI Agents (A2A), Polyglot Backends (.NET/Py/Node), Cloud, and the Psychology of Leadership.📍🇹🇷

Following

Help

Status

About

Careers

Press

Blog

Store

Privacy

Rules

Terms

Text to speech