Stacked

some numbers from this week that should be on your radar: - claude is the #1 app in the US and 15 other countries - 1 million+ new signups per day - daily active users tripled since january - 70% of companies choosing an AI tool for the first time now pick anthropic over openai - chatgpt's US mobile market share dropped from 69% to 45% in one year the shift is happening faster than anyone expected. and it started because one company told the government "no, you can't use our AI for mass surveillance." the lesson: trust is a growth strategy. I break down how to actually use these tools every tuesday. free @ readstacked.com

@CodevolutionWeb great read. started utilizing plan mode and it’s changed the game

@Suryanshti777 if you wanted a visual of a 1 person company with claude, this is it.

Claude Code just made the traditional startup team obsolete. I don't say that lightly. Look at this .claude/agents/ folder structure: 30+ specialized agents — each a single markdown file with one focused role. Engineering. Product. Marketing. Design. Legal. Finance. Testing. All of it. One folder. One person. No hiring. No managing. No overhead. Just: "Hey rapid-prototyper, build this." "Hey growth-hacker, find me users." "Hey legal-compliance-checker, is this okay?" This is the unfair advantage most founders don't know exists yet. Bookmark this before they do. 🔖

@heynavtoor absolutely love this

🚨 BREAKING: Claude can now write legal contracts like NDAs, freelance agreements, and LLC paperwork better than $800/hour corporate lawyers. Here are 12 prompts that replace $15,000 in legal bills: (Save this before it disappears)

anthropic just published the largest qualitative AI study ever conducted. 81,000 people. 159 countries. 70 languages. one week. the #1 thing people want from AI? it's not productivity. it's more time for life outside of work. 19% said "professional excellence" was their top answer. but when researchers pushed on WHY, the real answer came out: they want to finish work faster so they can be present for the stuff that actually matters. we've been thinking about AI wrong. it's not a productivity tool. it's a time tool.

@RoundtableSpace claude cowork is dank

If you can’t figure out Claude Cowork you need to read this article You will be a pro by the time you finish Thank me later x.com/jspujji/status…

@krishdotdev legal compliance checker is slept on

.claude/ └── agents/ ├── engineering/ │ ├── frontend-developer.md │ ├── backend-architect.md │ ├── mobile-app-builder.md │ ├── ai-engineer.md │ ├── devops-automator.md │ └── rapid-prototyper.md │ ├── product/ │ ├── trend-researcher.md │ ├──feedback-synthesizer.md │ └──sprint-prioritizer.md │ ├── marketing/ │ ├── tiktok-strategist.md │ ├── instagram-curator.md │ ├── twitter-engager.md │ ├── reddit-community-builder.md │ ├── app-store-optimizer.md │ ├── content-creator.md │ └── growth-hacker.md │ ├── design/ │ ├── ui-designer.md │ ├── ux-researcher.md │ ├── brand-guardian.md │ ├── visual-storyteller.md │ └── whimsy-injector.md │ ├── project-management/ │ ├── experiment-tracker.md │ ├── project-shipper.md │ └── studio-producer.md │ ├── studio-operations/ │ ├── support-responder.md │ ├── analytics-reporter.md │ ├── infrastructure-maintainer.md │ ├── legal-compliance-checker.md │ └── finance-tracker.md │ └── testing/ ├── tool-evaluator.md ├── api-tester.md ├── workflow-optimizer.md ├── performance-benchmarker.md └── test-results-analyzer.md

@trq212 fully agree.

Using Skills well is a skill issue. I didn't quite realize how much until I wrote this, the best can completely transform how your team works.

the biggest misconception about ai workflows: “you need to know how to code.” nah. the best workflow builders in 2026 are visual. drag and drop. connect this to that. done. n8n looks like a flowchart. zapier is basically filling out forms. make is somewhere in between. if you can build a playlist on spotify you can build a workflow. i’m not exaggerating. my grandma could set some of these up.

something i learned the hard way about ai outputs: if the response sounds like a linkedin influencer wrote it, your prompt was too vague. the fix is dead simple. tell it who you are, who you’re writing for, and give it an example of your actual voice. i keep a 3-sentence voice profile saved in my notes. paste it into every prompt. now everything sounds like me instead of a motivational poster. takes 5 minutes to set up once. saves you from editing every single output forever.

everyone’s chasing the perfect ai tool. meanwhile the actual unlock is connecting the tools you already pay for. your calendar. your email. your task manager. your notes app. most of these already talk to each other if you set it up. zapier connects them for free (up to 100 tasks a month). n8n does it with no limits if you want more control. you don’t need a new tool. you need a bridge between the ones collecting dust.

a prompt i use almost every sunday: “here’s everything i accomplished this week. [paste your done list] tell me: ∙what patterns do you see in how i spent my time? ∙what should i do more of next week? ∙what should i stop doing? ∙give me a brutally honest productivity score from 1-10” it’s like having a coach who doesn’t care about your feelings. 15 minutes. every sunday. changed how i plan my weeks.

hot take: the best ai workflow is the one you actually use. i’ve seen people build these wild 47-step automations that look incredible on a whiteboard and then never turn them on. start with one thing that annoys you every day. just one. automate that. get the win. then stack the next one. compounding beats complexity every time.

i automated my entire meeting follow-up last month. the old way: sit in meeting. take notes. forget half of them. spend 20 minutes writing a recap email nobody reads. the new way: record it. let ai pull out action items, deadlines, and who owns what. route it all to my task manager automatically. total setup time: about 2 hours. time saved every single week: about 3 hours. the math ain’t hard.

most people use ai like a search engine. type a question. get an answer. close the tab. the people saving real time are using it like a coworker. they give it context. their role. their goals. what they tried that didn’t work. one extra sentence of context in your prompt changes the output from generic to “wait how did it know that.” try it today. before your next prompt, add one line: “here’s what i’ve already tried and why it didn’t work.” watch what happens.

@itsolelehmann everyone. please 10x your claude skills. Ole is trying to warn you.

one skill. that's all you need to add. it takes any other skill you have and auto-improves it: 1: runs your skill and scores the output 2: finds what's failing 3: makes one small change to the skill prompt 4: runs it again to see if the score went up 5: keeps the change if it helped, reverts it if it didn't 6: repeats until the skill actually works full breakdown + the file to run it yourself:

if you missed today's issue here's what you missed: anthropic tracked 800 jobs and found that 94% of computer and math tasks could theoretically be done by AI. but only 33% actually are right now. that 61-point gap is the window. the people who learn the tools now are the ones who become irreplaceable when the gap closes. i also broke down a system i built called the plaud cascade. you record a phone call or meeting. you hang up. and before you sit back down: notion has your tasks with deadlines. google calendar has your follow-up meetings. gmail has draft emails written in your voice. slack has a summary for your team. and a knowledge base remembers every person you've ever talked to. one recording. five outputs. no typing. the full build guide is in the issue. step by step prompts you paste into claude and it walks you through every click. free every tuesday @ readstacked.com

the best automation advice i ever got: do it manually 5 times first. write down every step. find the bottleneck. fix the process. then automate. most people automate chaos and wonder why it doesn't work.

@JoshKale skills has to be the most underrated thing people are using with claude.

Anybody using Claude Code needs to read this. 'Skills' are the single biggest hack to get more out of Claude. Anthropic just cataloged every skill they use internally and found they all fit into 9 categories. After reviewing hundreds of skills, here's their top 9: 1. Library & API reference (edge cases Claude gets wrong) 2. Product verification (automated testing with video recording) 3. Data fetching & analysis (connecting to your monitoring stack) 4. Business process automation (standup posts, ticket creation) 5. Code scaffolding & templates 6. Code quality & review (spawns a "fresh eyes" subagent to critique code) 7. CI/CD & deployment (babysits PRs, retries flaky CI, auto-rollback) 8. Runbooks (takes a symptom, walks through investigation, writes the report) 9. Infrastructure ops (finds orphaned resources, gets confirmation, cleans up) The insight: most individuals or teams have skills in maybe 2-3 of these categories If your org is only using skills for code scaffolding and style enforcement, you're leaving 80% of the value on the table

chatgpt uninstalls surged 295% after the pentagon deal. claude became the #1 app in 16 countries. people vote with their installs.