-
LOW Confidence Isn’t a Bug. It’s Your Policy Telling You Something.
Post 3 of 5 in “Enterprise AI from the Inside Out” The bot returned LOW confidence again. I had just asked it something simple: “What happens if I miss the 30-day expense submission window?” The policy document was loaded. The retrieval found the right section. The model was grounded. And yet: LOW confidence, escalate to… — read more
-
Git to 42
I asked Claude a question about version control. It went somewhere neither of us planned. Here’s the conversation, lightly edited. Act One: The craftsman’s question Why git can’t track what AI agents learn Will: You are an expert in frontier AI and technology development. Give me perspective on the gap between git-based tracking versus skill… — read more
-
Eight Enterprise AI Patterns. AWS Covers Five.
Post 2 of 5 — Enterprise AI from the Inside Out TLDR: I built a local Python demo of eight enterprise AI patterns — no AWS account needed. Five map to Bedrock equivalents. Three expose gaps Bedrock doesn’t cover yet. The most interesting finding had nothing to do with the code. Jump to the Policy… — read more
-
Why Internal Tools Are the Enterprise AI Entry Point
Post 1 of 5 in “Enterprise AI from the Inside Out” TLDR: 76% of enterprises now buy AI off the shelf — up from 53% two years ago. Engineering leaders who want to build their way in anyway aren’t wrong, but they’re asking the wrong question. The real question is whether you’re building a tool… — read more
-
I Broke My NanoClaw Fork So I Had to Actually Learn How It Works
TLDR: I added voice to my NanoClaw fork the wrong way first — patched directly into core, it worked, upstream shipped an update, everything broke. Getting unstuck forced me to actually understand the skills architecture: three-way merging, plugin isolation, idempotent reapplication. Here’s that journey, including the TLS wall, a one-character VAD bug that took four… — read more
-
Standup Buddy: What 10 Hours of Voice AI Taught Me About Engineer Communication
TLDR: Built an async standup tool that detects when engineers are stuck through tone and conversational signals — not just words. 8th place at AI Agents Waterloo. GitHub here. Skip to Demo Results if you want to see the personas in action. The Problem Nobody Talks About Until It’s Too Late Most async standups are… — read more
-
The Boardroom: When AI Agents Need to Disagree
TLDR: Skip to the Demo Results section to see AI agents actually debate whether to implement unlimited PTO I have always been curious about what happens inside corporate executive meetings. Not the boring status updates, but the real policy discussions—the ones where the CFO pushes back on the CEO’s ambitious plans, where HR raises concerns… — read more
-
When AI Coding Assistants Get Stuck in False Beliefs
We’ve all been there. You’re working with an AI coding assistant, making progress on a problem, when suddenly something goes wrong. The AI suggests a fix. It doesn’t work. It suggests another fix. Still broken. An hour later, you’re no closer to a solution—and you realize the AI has been confidently pursuing an approach that… — read more
-
AI Personas
Part of the development process to create a product normally, is choosing how you want to create UX (user experience) flow. Few products can automatically give the experience that works for different types of consumers (think of Google or Apple). I am not trying match their amazing work, but at least try to give the… — read more
-
Dev Log – Agno LLM Framework
github link: https://github.com/lightshadow1/tax-advisor TLDR demo result: straight to bottom I have always been fascinated with technology, not only from the capability point of view, but it’s ability to unlock solutions for problems that was not even possible 2-3 years ago. Of course I am referring to LLM Gen AI model that gripped the world in… — read more
