Nic Touron

I built a tool that outperforms ChatGPT 5 Pro, Claude Opus 4.1, and Claude Code on knowledge work tasks. It's not another wrapper. It's not another chat interface. It transforms a simple prompt into publishable work in one shot. Here is how it works 🧵

@martinamps @quionie This is amazing - love it. But I don’t know how I feel about making bladder ultrasound public… 😂 Also the clinic address? Stay safe out there

@quionie sleep is a lil down lately tbh #sleep" target="_blank" rel="nofollow noopener">health.martinamps.com/#sleep

Dear Anthropic team, - whats the avg sleep schedule look like - how many agents are you running per person - do you have time to work out - do you eat 2-3 meals a day - what’s it feel like to lowkey dominate the market rn

@yescynfria @github That would be sick! You guys should get or source it and get some cred!!

@NicTouron @github Thank you!! I actually don’t know if I can share it since it’s a work thing? I’ll find out!

for hack week, i made a @github visualizer desktop app! · see your github repo in a visual format · get alerts on branch health · see diffs written in plain english instead of code · view an actual visual diff between branch/main

@dotta Cheers!

@NicTouron Thanks!! Today we're finishing up making it a lot easier to use OpenClaw. Hope to cut a release this afternoon

Less than 24 hours into Paperclip and we've got 👀 500,000+ views on X 🌟 2,500+ github stars 👨‍💻 10+ pull requests 💬 a buzzing discord We also shipped OpenCode integration, Cursor, bugfixes. The humans want zero-human companies

@klarnaseb Use @cursor_ai for front end bits - select components to modify in the built in browser + add screenshots.. game changer!

Watching Claude Code mass-murder backend tasks in seconds then spend 45 minutes misaligning a button... ... has given me more insight into the frontend vs backend engineer wars than a decade in tech. I used to think frontend engineers were being dramatic... 😂😂😂

@hunterhammonds Down for it! Built 20+ products with AI in the last 2 years!

I’m starting a community for cracked AI builders. Claude Code. Codex. Cursor. Conductor. It doesn’t matter. All I care about is that you’re building or you’re hungry to learn. We’re sharing workflows, skills, repos, plugins, etc. Want to join? Comment below and I’ll reach out.

@danshipper Yep same

claude code has been freezing for me in warp the last few days anyone else?

@hunterhammonds Yes doing both, with multiple tabs of CC using warp

Who is using Obsidian + Claude Code? I want to talk to you.

@linator @boredGenius @claudeai That's sick @linator, bringing back babyAGI! I am just worried adding one more plugin will just pollute the context window and degrade performance. Did you notice anything with and without it?

@boredGenius @claudeai I came up with my own approach after some vibe coding, and it’s starting to work really well: github.com/msaelices/clau… It uses local tools for TTS (or ElevenLabs) and STT (Whisper) instead of phone calls, e.g. using bluetooth earphones.

Introducing CallMe, a minimal plugin that lets Claude Code call you on the phone. Start a task, walk away. Your phone/watch rings when Claude is done, stuck, or needs a decision. Free & open source (MIT). Underlying API costs are cents per minute of call.

@HipCityReg ChatGPT Pulse is doing a decent job at daily brief based on interest! It offered some cool sites to monitor for WW3 risk level assessments! chatgpt.com/share/695fbba5…

currently building a "situation monitor" dashboard for the boys

@mntruell @milichab @mckaywrigley I love Cursor, I think it’s the best hands-on ai coding experience. But I think you still need to do some work for the async experience (make a plan, validate it - then go to bed. Wake up tomorrow with everything tested and working exactly as expected)

@mckaywrigley We think of Cursor as a desktop app, not an IDE. Agree the ideal Cursor becomes a dashboard for working with agents.

Cursor seeks to be the best and most powerful way to code with AI. What are the ways in which we could be better?

@mert If you are doing any sort of front end work Cursor is better. IMO models > IDE Opus 4.5 in most places will do wonders

alright all prior vibecodes were from using opus 4.5 in cursor now going to try giving claude code a shot lmk any tips/learnings

Six months. Six products shipped - some of them pretty big. I've had maxed-out subscriptions to every major AI platform (Codex, Claude, Bolt, Replit, and more), but I always return to @Cursor_ai. The IDE, the interface, the different modes, the model updates as they drop - it's the best of all worlds. I'm genuinely impressed by the little things the team includes daily. But as an Ultra user pushing Plan Mode hard, it's still behind the classic "write me .md specs" workflow. Here's how I would make it a true collaborator you can trust: 1. Questionnaire Constraints ⛓️ Problem: Questionnaires are productive but too rigid. I feel boxed in - I've started asking Plan Mode to skip them and just ask questions in chat instead. Fix: Add "I don't know" options + free-text notes. Let me guide with context, not just A/B choices. 2. Plan Collaboration 📑 Problem: Refining a plan feels disconnected from the conversation. Fix: Let me highlight sections, leave comments Google Doc-style, trigger targeted updates - and select text to add to chat like we can in the terminal. 3. Deep Research 🧠 Problem: Sometimes Cursor should search the web before making product decisions for me. Fix: Connect to Perplexity/OpenAI Deep Research during complex planning. 4. View & Tracking 👓 Problem: Tables render as `|—|—|`, line returns get swallowed, and I can't track what changed between iterations. Fix: Clean up markdown rendering + add diff view for plan versions (like Agent mode has for files). You're building the best tool in the game - I've built six products to prove it. Nail these collaboration layers, and Plan Mode becomes untouchable. Happy to chat through anything @n2parko @amanrsanger, @shaoruu & team 🫶

@mal_shaik Using headframe.ai daily - beating Chat GPT-5.1 Pro

what founders on this app actually use their own product? be honest. lmk

@aryanlabde Cursor!

Which one do you use for coding? cursor claude gpt Any other? I need a coding partner.

@danshipper @every Keep dropping the goods @danshipper 🙏

ive been working on @every for 6 years as of today we’ve grown more in the last year than we did in the previous 5 combined. and so have i im incredibly grateful for this team and for all of you. excited for the next 6 :)

@danshipper Huge indeed @danshipper I am very happy with it as well - I mostly vibe coded with Opus 4.1 (and a little of sonnet) - so far so good!

Opus 4.5 represents a huge milestone in agentic coding i was vibe coding this morning and ran into a bug. i accidentally pasted an old screenshot from a previous version of the app into Claude Code and asked it to fix the bug. rather than immediately believing the screenshot and running off on a wild goose chase, Opus 4.5 said, "this looks like a screenshot from an old run" and asked for more details that is HUGE

@thisiskp_ I would say they are not all in the same race! Cursor / CC -> dev The others -> vibe builders?

Cursor v0 Claude Code Bolt Lovable Base44 Who will win this race in 2035?

@blader The anthropic documentation is pretty clear - you have an agent when you have autonomy + getting feedback from your environment This is just a workflow builder 🤷‍♂️

when did we all decide that "agents" are just zapier workflows that occasionally make an LLM call? seems shortsighted.