Rida Ennasry
@EnnasryRida
Followers
241
Following
3K
Media
31
Statuses
296
Holding my pee until I reach 10k MRR https://t.co/Gk3uMi7Oc8
Joined January 2021
Amir walked 12km barefoot to get some food, got nothing but scraps, kissed the soldier's hand, then he was shot dead.
The child, Amir, walked 12 kilometers on foot and kissed the hand of Anthony Aguilar, the resigned American officer from the Gaza Humanitarian Foundation, in a plea for aid. Shortly after, Israeli soldiers killed Amir at the same location.
0
0
4
6) Safety disclaimer.Cursor CLI can read/modify/delete files and run shell commands you approve; it’s beta. Use only in a sandbox or a throwaway branch and audit diffs before applying.
0
0
2
5) MCP demo (local data → codegen). If you want “agent talks to your tools” cred, wire a local MCP server and point your agent at it (e.g., OpenZeppelin’s MCP server to scaffold secure contracts)
1
0
2
3) “Tiny model in front, big model behind” (SLM prefilter). You can’t point cursor-agent at Ollama directly, but you can pre-classify with a local SLM and feed its output to Cursor:. Why it’s cool: sub-100ms local “triage” + cloud agent for high-quality patches. (SLM-on-edge
1
0
1
2) Minimal “agent swarm” with a judge (bash). Cursor doesn’t orchestrate multi-agents for you, but bash does. You pipe outputs across three runs: finder → fixer → judge. Works on any repo; keep it in a feature branch to avoid chaos.
1
0
1
Cursor CLI just dropped. Most folks are clicking around. I used it to run headless agents that fix bugs, refactor, and open PRs, right from the terminal. Here are the exact commands + a security-safe way to chain “judge” agents 🧵
1
0
3
My takeaways after personal GPT-5 evaluation:. - Claude still rocks.- My AI doomsday jitters just chilled down. - GPT-5 seems more about efficiency gains for OpenAI than breaking new ground. - Feels more like “iPhone 16” than “first iPhone.”. what are yours?.
0
0
4
Have you ever run ads for your SAAS, can u share your experience?.
1
0
2
If you’re building AI agents, bookmark this repo. 30+ production-grade tutorials. Beginner-friendly. Real-world focused. All open source. 🔗
github.com
This repository delivers end-to-end, code-first tutorials covering every layer of production-grade GenAI agents, guiding you from spark to scale with proven patterns and reusable blueprints for re...
1
0
4
Why does a one-liner “light mode or dark mode?” reach 1.5 K people, yet a full case-study thread stalls at 30 views?. Is it that X prioritises polls & memes over value, or is it the blue badge or the VPN? What's going on 😅?.
1
0
9
I’m applying these rules on my third shot: Rekapo (AI that auto-summarises meetings). I’ll share every win + face-palm in public. If you’re building or burying a product, follow along.
0
0
1
TL;DR.1. Distribution is king. Marry it early. 2. TALK. TO. USERS. 3. Ignore vanity metrics. 4. 90-day kill rule. 5. Ask for money ASAP. 6. Iterate in tiny loops.
1
0
1
Ship → Feedback → Iterate (tiny loops).Big bang launches are casino bets. Small weekly releases + user calls beat month-long feature sprints every time.
1
0
1
Charge Immediately. Tingly stayed “free” for its life time, idea was to build PMF and then charge, the product never made a dollar. People love free. Wallets tell the truth. Even a $1 test price filters tire-kickers and forces you to serve real value.
1
0
1
90-Day Kill Switch. broblems dragged on because I was emotionally attached. From now every idea gets 90 days max. If key metric < Target X by day 90 → kill or pivot. Data decides; ego sits down.
1
0
1
Vanity metrics lie. sign-up ≠ activation. Activation ≠ retention. If people aren’t back in Week 1, you have a landing page, not a product. Example: broblems collected 200 emails in a weekend. Only 4 ever logged in twice.
1
0
1