
@tabilityio // posting about working w/ agents
Sydney
Joined August 2007
- Tweets3K
- Following1K
- Followers1.1K
- Likes2.4K
Opening code to non-devs is the best thing that can happen.
Here's the before/after @grapehichew got access to the repo managing our templates.
So good 💅
We've reduced initial context in @tabilityio ai mode by 60%. Much faster experience, still got all the power.
This.
"Let’s plan the roadmap for next quarter” is the modern day equivalent of “let’s fax everyone the strategy.”
We're going into q4 planning and it makes my brain hurt. Roadmaps existed because software was expensive to build. You needed a bunch of people. Meetings. Someone moving little boxes around Jira for three months from now. It makes no sense and I'm trying to work out what we call the replacement. Maybe just "Continuous Compilation", ideas just compile to working product. I want a list of ideas that we want to continuously compile. That's the roadmap now.
The cheapest way to figure out whether something is a good idea is often to just build the fucking thing.
Here’s how I work now. I've done this weekly over the last few months and it works. My current napkin math is that I can usually build 6 months of old-school roadmap in a week. Sometimes more. I've also hit a wall sometimes. It happens.
When we have an idea, I usually 10x the ambition and then write a prompt, not more than a page long, that basically says:
- here’s the ambitious goal
- here are a few use cases
- here’s roughly how I think we should build it in terms of where it fits into code base, design considerations, etc..
- here’s how we’ll know if it actually works
- now go fucking surprise me
Agents build all night. Not even with a crazy factory or complex agent or agents. The base models are just too good for that now.
In the morning I look at it. Ok, they usually build in a few hours. Usually some of it is great. Some of it is stupid. Some of it exposes that my original idea was wrong.
And almost every time it gives me 5 new ideas I couldn’t have planned upfront. So we throw shit out. Who cares? We learned something.
Then we change direction and send them off again. It usually takes 3 rounds before the shape appears and we've learned about the gremlins and possibilities.
95% of the time I end up building something with 10x the scope of what somebody had carefully spread across the next quarter in Jira.
“What should we put on the roadmap for next quarter?”
My brain now translates that to: “Oh, you're stuck in the 1500s”
seems that every new post here is "<new model> is insane".
it might be true but it might also means that we're racing to the bottom
Trying @bot now, and it's actually 10x better than Codex for building teams of agents. Nuts.
Today was focused on giving our codex and claude agents beautiful dashboards. Must haves:
- A simple way to track active vs. quiet agents
- Measure outputs
- See which agents needs an input (via⚠️ sign)
- Track goal and task completion
Try 👉 tability.io
In a world where agents can build anything on a whim, the best thing you can do is give them a clear definition of success.
I did a quick audit of how efficient our new agent workforce manager is.
I've got 21 Codex agents managed by @tabilityio. All working on complex GTM goals.
- 2 days elapsed so far
- 10h of autonomous work done
- 97 productive runs
- 15 assets
- 42 tasks
- 25 status reports
This setup is the best way to put agents on long-running goals.
- Tability can create and manage the agent teams in Codex/Claude for you.
- You get all the monitoring and reporting tools required for long term goals
- You get all the power of models like 5.6 Sol, Opus 5 + your Codex/Claude harness.
(note I've tried this with Fable 5 and ran out of tokens in a day -- would not recommend yet unless you have infinite cash).
This is free to try. You just need Tability workspace and a Codex or Claude subscription.
The best way to work with agents is to treat them like a real teammate.
I'm sure they'll figure this one out.
This morning I deployed 21 agents on AEO and growth goals.
9:00 create 5 growth goals in @tabilityio
9:05 assign each goal to an agent manager
9:05-9:30 tability <> codex magic
9:30 tability has deployed 21 codex-based agents, divided in 5 teams (each specialised on a goal)
It's now 2pm and the agents have created:
- 21 sub-goals
- 31 tasks
- 35 status reports
Middle management for agents
xitter.cf/stenpittet/status/2090…
Sten retweeted
Sometimes the most accurate description of OKR season is a meme, not a framework diagram.
We rounded up 15 of them — the ones that make people in strategy meetings laugh a little too hard because they've lived it.
Painfully accurate, mildly cathartic, entirely too relatable:
tability.io/odt/articles/10-…
The most ambitious thing we've worked on: @tability can deploy agentic teams on your goals.
Not another assistant that waits for instructions. It's an expert team that own any goal end-to-end, and works 24/7 until the goal is achieved.
Proper progress tracking, full context, status reports and feedback. Manage it like you would manage a remote team.
Got a new goal in mid? Deploy an agentic team for it in 1-click. This extends your team 100x
AI assistant are great for outputs, this is built for results.
Sign up at tability.io, DM for private beta.














