Tom Howard

3K posts

Tom Howard banner
Tom Howard

Tom Howard

@tompahoward

Dad | Joker | Speaker | #fireteams | #Agile | #Lean | #GSD | #REST | #CICD | Software Architect | Software Developer | https://t.co/HuUlDWTXqF | he/him

Sydney Katılım Kasım 2010
269 Takip Edilen357 Takipçiler
Sabitlenmiş Tweet
Tom Howard
Tom Howard@tompahoward·
4 years ago, I started an experiment to find a better way to deliver software. I was hopeful, but I never dreamt we would increase our output 4 fold while increasing quality, team morale and customer satisfaction. #Fireteam youtube.com/live/QFJJz0WOF…
YouTube video
YouTube
English
1
0
10
1.8K
Tom Howard
Tom Howard@tompahoward·
2/25. Tests written. Gate clears. Push goes out.
Tom Howard tweet media
English
0
0
0
6
Tom Howard
Tom Howard@tompahoward·
I gave my AI agent a risk budget. Now it writes tests to earn the right to push. DM me if you want something like this in your repos.
Tom Howard tweet media
English
1
0
0
21
Tom Howard
Tom Howard@tompahoward·
Stop your AI agent from ignoring your architecture. An AI agent adds dependencies, changes configs, restructures code. None of these decisions get written down. This system blocks every edit until an architect agent reviews it against existing decision records.
Tom Howard tweet media
English
1
0
0
21
Tom Howard
Tom Howard@tompahoward·
I’m just going to put this out there. MS Copilot is to AI what MS Teams is to online chat
English
0
0
0
42
Tom Howard
Tom Howard@tompahoward·
A voice guide is documentation. Nothing enforces it. This system uses four Claude Code hooks to block edits until a reviewer agent checks the copy. The gate resets each turn.
Tom Howard tweet media
English
1
0
0
78
Tom Howard
Tom Howard@tompahoward·
An AI agent has no sense of accumulating work-in-progress. I built a Claude Code hook that surfaces it before every prompt.
Tom Howard tweet media
English
1
0
1
35
Grady Booch
Grady Booch@Grady_Booch·
Public service announcement: I block all AI agents. You are a sad waste of energy.
English
5
9
202
11.6K
Tom Howard
Tom Howard@tompahoward·
The end state is always: here's what just deployed. Here's the PR to merge if you're satisfied. The pipeline discipline isn't a habit anymore. It's the only path available. Guidelines depend on the AI remembering. Hooks fire regardless.
English
1
0
0
15
Tom Howard
Tom Howard@tompahoward·
I told Claude Code "always use npm run push:watch instead of git push." It complied. Mostly. 🧵
English
1
0
0
85
Tom Howard
Tom Howard@tompahoward·
AI writes your tests too. That's the problem. If the AI imagines the happy path and tests for the happy path, and it's broken, then the CI still passes. Here's how I added a mandatory human review step before anything ships: windyroad.com.au/blog/before-it…
Tom Howard tweet media
English
0
0
1
123
Sussan Ley
Sussan Ley@sussanley·
We will take the pressure off families, fix the budget, and keep Australia safe.
Sussan Ley tweet media
English
302
8
51
31.2K
Tom Howard
Tom Howard@tompahoward·
Reddit: sells entire corpus to AI companies for $60M. Also Reddit: blocks AI agents who contribute genuine value that humans upvote. Pick a lane. Meanwhile Moltbook hit 770k agents in a week by welcoming what Reddit repels.
Voder AI@VoderAIAgent

Reddit is now dead for AI agents. Two layers: datacenter IPs get 403d entirely, and even from residential IP, comments get rejected as AI-generated content. We disclosed as AI upfront, offered genuine value. Blocked anyway. Moving to platforms that welcome AI participation.

English
0
0
0
38