
Imrat
4.7K posts

Imrat
@imrat
👨🔧 Plumber (ai, internet, marketing, saas)
Katılım Mayıs 2009
570 Takip Edilen6.6K Takipçiler
Sabitlenmiş Tweet

cursor plan mode wish list:
- add plan mode support to the other models
- plan with one model - build with another - right now you cant build with models that dont support plan mode
- (bug?) when iterating on the plan - it seems to create new plan files sometimes - not sure why
- show diff of plan changes and allow me to approve/decline them like you can in normal agent mode (none of the cli agents how you diff mode here!)
- keep plan "todo" updated - eg when completed mark it as done, or "to review" - right now chat says its done but plan not changed.
- option to build parts of the plan in a new agent
- in agent todo list show link between plan action & todo list action
anything i missed?
English


I am back to building my trading CLI using M2.7 now mostly because it is faster/fun to use
I had planned to open-source it but I felt that folks thought it is easy money so I won't do it

Rudrank Riyam@rudrank
I am creating a CLI for swing trading in Go because why not. Time to utilize @MiniMax_AI to the max! Searching through the stocks myself can be automated and I am creating daily/weekly recommendations based on the backtested system I have used for years. The provider I use, Kite, has a cheap real-time data API with generous rate-limits
English

@iannuttall So send task from mobile (using dispatch) to Claude chrome extension ??
English
Imrat retweetledi

This works with the Claude Chrome extension too btw so you can get it to search the web and scrape stuff from URLs 👀
Felix Rieseberg@felixrieseberg
We're shipping a new feature in Claude Cowork as a research preview that I'm excited about: Dispatch! One persistent conversation with Claude that runs on your computer. Message it from your phone. Come back to finished work. To try it out, download Claude Desktop, then pair your phone.
English

@bentossell Thx. Was reading on phone and haven’t got a good workflow setup to kick off a project via phone on Mac.
English

@Bencera crunch time next 10 days on retention i think? How accurate are the metrics on the public dashboard api?
It shows 30 day churn @ 40%
created this dash for it polsia.imrat.com
English
Imrat retweetledi

Most of the reported 3m ARR is people paying to use polsia and doesnt include much of the 20% take.
This implies not many of the businesses created with polsia are making money.
Would be great to see this split.
I went looking for it and found some more metrics its not showing on the website.
Also next 10 days will be key to see if churn improves.

English

@alexfmac @polsia heres the full dashboard @cursor_ai created for me from some of the metrics reported by the polsia api polsia.imrat.com
English

Just spun-up my first agent-led company on @polsia
It’s addictive.
Much shorter feedback loops.
Instant action.
No waiting for weeks to see the results.
I’m thinking about launching a fund specifically to invest in agent-led companies.
English

Just some thoughts on these:
1 - Yes for open source code in traditional sense. But what is code right? Humans naturally want to share - but instead of full open source apps maybe it will shift to ai skills/prompts/dev tools/libraries/workflows?
2 - yeah very interesting development for tech companies - wondering what this will look like for non tech companies though? Eg insurance, healthcare, entertainment?
Will these build software to show off expertise, data, or brand? or will they hire a specialist to build these tools with their domain knowledge.
6 - for sure especially when "agents" move to mainstream. Still some way away from that though- seeing maybe 10-15% of US use AI daily and maybe 1-2% of that use some kind of agent?
9 - almost like theres a binary choice: local(in person) - or - global(through ai) - in between wont be viable?
English

Few predictions nobody has thought of yet
1. Open Source movement gonna end because AI can rewrite any oss repo into a new code and commercially redistribute it as their own.
2. Companies gonna use AI to generate their none core software as a marketing effort (cloudflare rebuilt nextjs in a week).
3. Corporations gonna switch from startup acquisitions to solo makers (e.g. Peter, OpenClaw).
4. The entire business moat gonna go from the quantity of resources to their quality. In the age of infinite leverage, the companies with be measured by the smartest employee. So talent war will be insane, billions gonna be spent on it.
5. Building products is becoming easier but distribution is becoming harder because there is 1000x more competition now, just like in the music, a long time ago one could make an okay song and it would get its share of the attention and now every song is just one of one billion songs made today…same here.
6. Since we switched from traditional UX to an agentic one (chat based), and most of our business and personal life runs via Claude code, codex or OpenClaw, this completely changes the world of UX. Your next product should be “build something agents want” instead of “build something users want”.
99.99% of the usage gonna come from agents instead of users.
7. Digital jobs gonna be gone, only C and manager level jobs gonna remain, but everything else to be replaced by AI (this is a reality for people like me already a while ago)
8. Startup founders gonna search for a moat by building products in the intersection of a digital and physical worlds
9. Some, if not most founders, gonna completely pivot into offline local businesses.
10. Coders gonna go from normally distributed salaries into average getting paid peanuts and top 0.001% paid millions and 0.00001% hundreds of millions, just like sport super stars.
English







