ทวีตที่ปักหมุด
Rohan
33 posts

Rohan
@agenticrohan
Founder @ Klaudius: An AI website agency that runs 24/7, finding, building, and pitching bespoke websites to businesses that don’t have one, at scale.
เข้าร่วม Mart 2026
7 กำลังติดตาม4 ผู้ติดตาม

built the entire nor.ma website with claude code.
layout, animations, the 5% off email flow. didn't touch a single other tool.
kinda insane how far this thing has come.
Drice@driceroland
rebuilt the norma site. clearer, more on-brand, should convert a lot better. nor.ma
English

@komal_uk01 Use klaudius.dev and start letting it sell websites for you while you sleep!
English

@MitchellKeller_ Nice one, you could easily productise this and sell it to PE firms!
English

I built a system that finds every Series A company funded in the last 24 hours, enriches the data, and pushes it straight to Clay — for under $2/month. In one Claude Code session.
Stop paying hundreds a month for stale funding signals.
Here's exactly how it works:
→ SerperDev for Google search at pennies per query — no ZoomInfo, no expensive APIs
→ The self-annealing loop (Karpathy's concept): Opus for strategy, Mini for execution, smarter every run
→ Spider to test your signal against fresh data before you ship it
→ Trigger. dev for scheduled deployment every morning — no n8n, no manual runs
→ 61 raw results → 37 confirmed Series A leads, pushed live to Clay + Supabase
The part people forget: building the ground truth first. That's what makes the signal self-correct over time instead of drifting.
This pattern works for any signal type - not just funding.
Full walkthrough is live. 10 minutes.
[link in comments]
English

@primemans The truth is, if you’re not from a technical background, this video will be an absolute gem.
English

Ali Abdaal just shared his Claude Code workflow.
And this might be the most beginner-friendly breakdown out there.
Most Claude Code tutorials feel like they’re built for developers:Terminal commands. API keys. Technical jargon that loses most people in the first minute.
Ali takes a different approach.
He starts from zero — and actually makes it under.
English

@heynavtoor When people ask how to use Claude to make money, stuff like this is indirectly doing just that
English

A scientist in Denmark figured out how to make Claude prepare his job applications. He open-sourced the whole thing.
His name is Mads Lorentzen. He is a PhD geophysicist. He built it on top of Claude Code and released it under MIT license.
Here is what it does. You fork the repo, fill in your background once, and it runs a five-step pipeline for every job you want to apply to.
Step 1. It reads the job posting and scores how well you fit.
Step 2. It drafts a tailored CV in LaTeX, picking only the experience that matches.
Step 3. It writes a cover letter framed around what you would bring to the role.
Step 4. A second AI agent reviews the first agent's work, points out weaknesses, and the first agent revises.
Step 5. It compiles both into clean PDFs you can send.
The whole thing is a folder of markdown files. The candidate profile, the writing style rules, the CV templates, the interview prep notes. Every step is plain text you can read and change.
The job portal search is built for Danish boards. The application workflow itself works for any country.
489 stars. 270 forks. A fork-to-star ratio that high means people are using it, not only bookmarking.
Mads is not a startup founder. He built this because he needed it for himself, then shared it.
This is the future of job hunting. Not a service you pay for. A workflow you own.
(Link in the comments)

English

@bridgemindai Just as Codex was getting ahead in this fascinating battle, a culmination of events has led to Claude being top dog again
English

Now that the Codex 2X promo has ended the usage limits on Codex are terrible.
Last night I hit my 5 hour limit in 1 hour.
These limits are just as bad as Claude Code limits in April.
I am now officially getting more usage out of Claude Code with Opus 4.8 then Codex with GPT 5.5.
So much for OpenAI being "generous".
BridgeMind@bridgemindai
Codex limits are NERFED. With my $200 ChatGPT Pro plan I used to never worry about my limits using Codex. I've been using GPT 5.5 the past 2 days and am already 45% through my weekly limit. This is a completely different Codex then what we had last week. Who else is noticing this?
English

@karthikponna19 Either - our AI web agency tool supports both, and that allows you to use your subscription to generate revenue!
English

@OnlyTerp @NousResearch Awesome, yep when you get the orchestration right, parallel agents are game changing
English

I built Caduceus for Hermes Desktop @NousResearch
Two snakes, one staff
Caduceus runs parallel AI agents in perfect sync, on any model you already have
It sizes up every task and picks the path: • Easy → answered instantly • Hard → a live to-do plan + agents fanned out in parallel, verified before it calls it done
Caduceus has a UI right inside Hermes so you watch every agent build your project in real time
Full demo 👇
English

Workflows are the biggest upgrade to Claude Code’s capabilities since skills and subagents.
I dove deep into it with @sidbid to figure out best practices, examples and more. I’m particularly excited about the non-technical tasks it enables for Claude Code.
Thariq@trq212
English

@ClaudeDevs Ask Claude to review its own work, and also ask it to launch a subagent with fresh context to review its own work in parallel, then fix the combined findings. That way, you combine the pros of fresh context + the pros of context awareness.
English

@svpino Yep, context fatigue is real and you need to be the one proactive about it, Claude won’t do it for you. Something we spent a lot of time figuring out for klaudius.dev was how to keep each pipeline run on a fresh session whilst still keeping it completely autonomous.
English

The 60% rule to compact your context:
Don't let your sessions go above 60% in context utilization. After that, your model will start feeling drunk and you'll risk Claude Code to trigger auto-compact for you.
Auto-compact is not good if you want control on what context stays and what goes.
Instead:
• Run /statusline to permanently display context %
• When > 60%, run /compact with instructions on what to keep
• For even more control, dump session context to a file and edit it
Many people don't know this:
You can run /compact with a prompt explaining what you want to keep. For example:
"/compact Keep the project description and change instructions"
This will anchor Claude Code's summary around those topics and will compress everything else more aggressively.
English

@RhysSullivan It does a really good job at overall context management as well by its subagent heavy architecture
English

Best part? Not one of these needed a phone call. Every deal closed over text (i.e. closed by AI).
The full write-up, the real quotes from the inbox, and the exact reply scripts that close:
klaudius.dev/blog/what-to-d…
English













