Patrick Robinson retweeted
"The first thing you need to do to make effective use of AI and autonomous AI is to improve your own process"
-Gartner
Totally agree. I keep seeing teams over-trusting AI. They lack the automated quality controls to safely move faster.
-Gartner
Totally agree. I keep seeing teams over-trusting AI. They lack the automated quality controls to safely move faster.
11
5
34
2,993
Patrick Robinson retweeted
Subagent overuse and abuse has been a problem since mid 2025.
There is a time and a place and they can be very high leverage, but the vibeslop dopamine casino + twitter hype machine has let the average user down by encouraging more and more parallelism instead of thoughtful context management
There is a time and a place and they can be very high leverage, but the vibeslop dopamine casino + twitter hype machine has let the average user down by encouraging more and more parallelism instead of thoughtful context management
Anthropic is now recommending that companies REMOVE unnecessary subagents from their AI systems.
This is a pretty significant shift in agent architecture, with their team finding that a single agent equipped with skills can frequently outperform complex multi-agent systems, and one customer seeing roughly 2x faster execution.
The emerging rules of thumb:
→ Skills: Teach the agent HOW to do something, including procedures, expertise, and instructions.
→ Tools / MCP: Give the agent the ability to DO things through APIs, actions, and live data.
→ Subagents: Use for parallel work, isolated context, or tasks a single agent struggles with.
The interesting part is progressive disclosure. Instead of spawning specialized agents or stuffing everything into a giant prompt, one agent discovers and loads skills and tools as needed.
Their advice is to fold existing subagents into the main loop, turn their instructions into skills, expose their actions as tools, and run evals to see if performance holds up.
As models have gotten smarter, improving agent architecture is becoming less about what you can add and more about what you can delete.
claude.com/resources/arti…
52
26
492
32,593
Patrick Robinson retweeted
Mistake: Creating a plan via AI that answers many ambiguities in the ticket.
Instead: Update the ticket with all decisions made in the plan.
This way, the ticket is a *permanent*, centralized, comprehensive record of *all* decisions that's easy to reference later.
Instead: Update the ticket with all decisions made in the plan.
This way, the ticket is a *permanent*, centralized, comprehensive record of *all* decisions that's easy to reference later.
22
5
59
6,941
Can't say how many times in the last year or so I've said, "Have you asked AI?"
More recently, "Have Claude do it."
More recently, "Have Claude do it."
0
0
0
8
Patrick Robinson retweeted
"Smaller things get through the pipe faster." - Plumber
Good advice for devs too.
Good advice for devs too.
10
7
66
6,026
Patrick Robinson retweeted
If you track software factory metrics, don't celebrate the wrong signals.
Bad: LOC written, tokens used
Good: Cost-per-PR, cycle time from conversation start -> merge
A factory leaderboard should take cost into account and champion shipped work, not token spend.
Bad: LOC written, tokens used
Good: Cost-per-PR, cycle time from conversation start -> merge
A factory leaderboard should take cost into account and champion shipped work, not token spend.
21
10
185
8,964
Patrick Robinson retweeted
Wonder if your tests are helping? Ask Claude.
Prompt: “Look through my past sessions. How many times did a test catch a bug?”
I just asked Claude and it found 5 recent instances where a test caught a bug.
Prompt: “Look through my past sessions. How many times did a test catch a bug?”
I just asked Claude and it found 5 recent instances where a test caught a bug.
21
3
39
5,850
Patrick Robinson retweeted
I’ve seen many teams struggle because their tickets are too big.
I’ve never seen a team struggle because their tickets are too small.
I’ve never seen a team struggle because their tickets are too small.
27
1
48
6,965
You can see when I started using my personal GitHub to do things... and when I created my factory
i wanted to visualize what my github history looked like before and after the AI inflection point.
see yours here - mo.rgan.co/gh/
i've looked at a few friends, eng and design, and it's really interesting to see the patterns.
0
0
0
20
I've never been on a team that struggled with tickets being too small.
What's your team's right-sized ticket?
If you respond in story points, tell me why you use points.
What's your team's right-sized ticket?
If you respond in story points, tell me why you use points.
0
0
0
7
Hit weekly limit on Grok @bot
Also noticed additional usage option was gone 👀
What's everyone up to these days?
Also noticed additional usage option was gone 👀
What's everyone up to these days?
1
0
0
37
Did the mega.dev assessment today.
Had it evaluate me and my factory sessions.
My factory won.
Not sure that's good or bad?
@kentcdodds what do you think?
Had it evaluate me and my factory sessions.
My factory won.
Not sure that's good or bad?
@kentcdodds what do you think?
0
1
1
16
Patrick Robinson retweeted
This post outlines an interesting new approach that AI unlocks: Non-code automated tests.
For example, QA doesn’t typically have the tech chops to write and maintain code-based tests.
But they can certainly ask Claude to click around for them!
For example, QA doesn’t typically have the tech chops to write and maintain code-based tests.
But they can certainly ask Claude to click around for them!
Before you launch anything, have Opus 5.5 send 5 subagents in as your first users.
Paste this into Claude Code:
"Act as my launch QA lead. Spin up 5 subagents, each a different first-time user:
1. The impatient one: skims, clicks the first button, quits at the first error
2. The skeptic: reads pricing, terms and privacy before signing up
3. The mobile user: does everything on a 375px screen on slow 3G
4. The non-native English speaker: only follows what the UI literally says
5. The breaker: pastes junk into forms, hits back mid-flow, double-clicks everything
Each one opens the app in a real browser, goes from landing page to first success, and screenshots every moment they get confused, stuck or annoyed.
Merge everything into one list ranked by how many users hit it. Fix it all in one pass, then rerun the 5 users to prove it's fixed."
Your first 100 real users won't tell you why they left. these 5 will.
(Works best with a browser MCP connected, Playwright or Chrome DevTools)
22
4
80
12,051
Passed Anthropic's Claude Certified Developer exam. So it's official: I'm a certifiable Claude user. credly.com/badges/b7eb2d8…
0
0
0
32
Right?!
@bot can you help us get off of whatever list we're on, or find any legal recourse for receiving unsolicited texts?!
@bot can you help us get off of whatever list we're on, or find any legal recourse for receiving unsolicited texts?!
Dear Texas politicians,
I’ll vote for whoever stops texting me 🤬
2
0
5
2,869