e-man

Building @usecrafter | rooms (private beta)

[e-man @the_evangelizer]

Ask not what AI can do for you—ask what you can do for AI.

[e-man @the_evangelizer]

@sflorimm We're normalizing it, when we shouldn't. Its why I built Rooms.

[e-man @the_evangelizer]

@kobyjconrad The attention should be on interaction and systems building through conversation, not the code. The argument for some is that Cursor has it backwards. IDEs are just code viewers.

[e-man @the_evangelizer]

@escliu @byjmitch The application process was refreshingly direct. More focused on clarity and conviction than overproduced startup storytelling. Excited.

[e-man @the_evangelizer]

👋

[e-man @the_evangelizer]

@rauchg Building @usecrafter https://t.co/YfoB1NgkSY

Codex, Fast Mode, 5.5

[e-man @the_evangelizer]

@sama @usecrafter is solving the biggest pain across team execution -- transparency. And OAI is already helping us!

[e-man @the_evangelizer]

@xadh @the_engi_nerd Sadly many of these companies are not AI-native and don't hire smartly.

Oh, and @usecrafter isn't slow :)

[e-man @the_evangelizer]

@iamjeganb @cursor_ai Watch out for @usecrafter, we're building capability to also assign tickets to your Claude and Codex agents. More soon. #crafter #jira #AI

[e-man @the_evangelizer]

Most teams don’t miss deadlines because of bad planning. They miss because reactive work slowly takes over execution. Interruptions. Scope changes. Investigations. Stakeholder asks. Hidden overload.

This week reinforced something for us at @usecrafter

#Engineering #AI

[e-man @the_evangelizer]

Most teams don’t actually know why projects slip.

Jira, for example, tells you what’s assigned.

Crafter shows you what’s disrupting execution.

#crafter #AI #building

[e-man @the_evangelizer]

@gdb Wait. I thought it was already

[e-man @the_evangelizer]

@sama Sweet. Goblins unite 🧌

[e-man @the_evangelizer]

@sama Goblin party. Sign me up.

[e-man @the_evangelizer]

@bruvimtired Apology accepted. Welcome

[e-man @the_evangelizer]

@RhysSullivan Decision making becomes crucial

I use my several LLM threads as a team. I've been experimenting with rooms (where threads are personas that can check each other) - to make better decisions

[e-man @the_evangelizer]

@thsottiaux I use codex as an AI team. Threads act as dedicated personas (engineer, PM, marketing) with deep DM context.

Lately, I've expanded into the concept of rooms, where I invite these chat thread personas. Helps a lot with decisioning on my end before execution.

[e-man @the_evangelizer]

@Mrwhosetheboss Now read it in reverse

[e-man @the_evangelizer]

@garrytan This is right

But faster building just exposes the next problem: decision quality

A single AI thread collapses everything into one voice

I’ve been turning AI into a team via “Rooms”
specialized chat threads working together in one place

Way better than one thread, one answer

[e-man @the_evangelizer]

@petergyang That's the way to go.

[e-man @the_evangelizer]

@weswinder yup