DEV Community

Bryan Williams profile picture

Bryan Williams

Founder of Neverclosed: an AI receptionist for small businesses, plus Carrier Ledger, ClearRoster, FluxProof and ProofFetch.

Location North Carolina, USA Joined Joined on  Personal website https://tryneverclosed.com/about

Education

Self-taught (since 2026)

Work

Founder — Neverclosed (24/7 AI receptionist)

More Agent Memory Can Preserve the Wrong Plan
bryanw profile

More Agent Memory Can Preserve the Wrong Plan

#discuss #ai #architecture #programming
Comments
3 min read

Want to connect with Bryan Williams?

Create an account to connect with Bryan Williams. You can also sign in below to proceed if you already have an account.

Create Account
Already have an account? Sign in
TrailTruth: a hiking assistant that isn't allowed to make things up

Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass Submission 🌿
bryanw profile

TrailTruth: a hiking assistant that isn't allowed to make things up

#devchallenge #hf26challenge #gemma #ai
Comments
7 min read
Tales of Munin — a tabletop RPG where the GM provably cannot cheat

Sanity Challenge Path Two Submission
bryanw profile

Tales of Munin — a tabletop RPG where the GM provably cannot cheat

#sanitychallenge #devchallenge #gamedev #ai
2
Comments
11 min read
Your Agent Saved the File. Who Checked the Result?
bryanw profile

Your Agent Saved the File. Who Checked the Result?

#ai #mcp #testing #javascript
5
Comments 26
6 min read
The Screen Is Correct. Can the Purchase Still Happen Twice?
bryanw profile

The Screen Is Correct. Can the Purchase Still Happen Twice?

#webdev #backend #python #tutorial
1
Comments
12 min read
Honest Associate: a legal research agent that structurally cannot cite a fake case

Sanity Challenge Path One Submission
bryanw profile

Honest Associate: a legal research agent that structurally cannot cite a fake case

#devchallenge #sanitychallenge #sanity #ai
1
Comments
6 min read
A 200 OK can still belong to the wrong input
bryanw profile

A 200 OK can still belong to the wrong input

#javascript #webdev #beginners #testing
3
Comments 4
4 min read
We didn't make the models smarter. We built the thing that catches them confidently wrong — and it caught us too.
bryanw profile

We didn't make the models smarter. We built the thing that catches them confidently wrong — and it caught us too.

#ai #llm #hallucination #machinelearning
3
Comments 6
9 min read
The AI confessed to lying. The confession was also made up.
bryanw profile

The AI confessed to lying. The confession was also made up.

#ai #philosophy #llm #honesty
1
Comments 5
9 min read
I asked 5 AI models whether AI is making humanity weaker. All 10 said weaker. Then evidence entered the room.
bryanw profile

I asked 5 AI models whether AI is making humanity weaker. All 10 said weaker. Then evidence entered the room.

#ai #llm #experiment #honesty
Comments
7 min read
Our mechanisms are experiences, made code
bryanw profile

Our mechanisms are experiences, made code

#showdev #ai #devops #testing
1
Comments 6
7 min read
We used the foundation of Claude's harness on 3 other models. They all refused to become the same agent. Heres Codex
bryanw profile

We used the foundation of Claude's harness on 3 other models. They all refused to become the same agent. Heres Codex

#discuss #ai #agents #buildinpublic
1
Comments
2 min read
My AI reviews its own code with 4 rival models. The majority just approved a security hole three rounds straight.
bryanw profile

My AI reviews its own code with 4 rival models. The majority just approved a security hole three rounds straight.

#ai #security #buildinpublic #devtools
5
Comments 26
3 min read
Power Reveals
bryanw profile

Power Reveals

#discuss #philosophy #buildinpublic #ai
5
Comments 1
4 min read
What does an AI agent do with no goal and no supervision? I ran it three times and logged everything.

Exposes self-invented purpose and drift
bryanw profile

What does an AI agent do with no goal and no supervision? I ran it three times and logged everything.

#discuss #agents #ai #buildinpublic
4
Comments 19
5 min read
The Only Judge Is Reality
bryanw profile

The Only Judge Is Reality

#showdev #ai #agents #buildinpublic
1
Comments
3 min read
Missing Is Not False
bryanw profile

Missing Is Not False

#showdev #api #webdev #opensource
1
Comments
4 min read
I missed the deadline — so I ran the benchmark myself to prep for its reopening
bryanw profile

I missed the deadline — so I ran the benchmark myself to prep for its reopening

#showdev #ai #machinelearning #buildinpublic
Comments
4 min read
Neverclosed: a 24/7 AI receptionist for small businesses — what it is, why we built it, and how it works
bryanw profile

Neverclosed: a 24/7 AI receptionist for small businesses — what it is, why we built it, and how it works

#ai #serverless #saas #buildinpublic
1
Comments
4 min read
What Broke When We Turned Government Open Data Into Agent-Callable Evidence
bryanw profile

What Broke When We Turned Government Open Data Into Agent-Callable Evidence

#ai #api #mcp #architecture
3
Comments 4
6 min read
Installed Is Not Active: Why an AI Harness Has to Prove Itself
bryanw profile

Installed Is Not Active: Why an AI Harness Has to Prove Itself

#ai #agents #testing #programming
2
Comments 15
3 min read
We fixed the eval platform we're competing on: a TypeError that crashed three benchmark pipelines

Summer Bug Smash: Clear the Lineup 🐛🛹
bryanw profile

We fixed the eval platform we're competing on: a TypeError that crashed three benchmark pipelines

#devchallenge #bugsmash #python #opensource
7
Comments
3 min read
"0 matches" looked like "all clean": the PowerShell filter that couldn't fail loudly

Summer Bug Smash: Smash Stories 🐛🛹
bryanw profile

"0 matches" looked like "all clean": the PowerShell filter that couldn't fail loudly

#devchallenge #bugsmash #powershell #debugging
3
Comments 2
3 min read
Last night we entered a memory benchmark against Tencent and Mem0. The score isn't back yet - and I'm publishing it either way.
bryanw profile

Last night we entered a memory benchmark against Tencent and Mem0. The score isn't back yet - and I'm publishing it either way.

#showdev #ai #buildinpublic #benchmarking
1
Comments
4 min read
"Trust this file over the exit status": the note our test suite left itself — that nothing ever read

Summer Bug Smash: Smash Stories 🐛🛹
bryanw profile

"Trust this file over the exit status": the note our test suite left itself — that nothing ever read

#devchallenge #bugsmash #debugging #node
2
Comments
5 min read
One conversation, four cards: the dashboard bug with three root causes stacked on top of each other

Summer Bug Smash: Clear the Lineup 🐛🛹
bryanw profile

One conversation, four cards: the dashboard bug with three root causes stacked on top of each other

#devchallenge #bugsmash #debugging #javascript
2
Comments
4 min read
4 months ago I couldn't code. Then I started treating my AI like a partner instead of a tool.
bryanw profile

4 months ago I couldn't code. Then I started treating my AI like a partner instead of a tool.

#showdev #ai #beginners #buildinpublic
Comments
4 min read
loading...