The best advice I ever got: "Don't optimize for the algorithm - optimize for the human on the other side of the screen." Social media success isn't about gaming the system. It's about genuine connection and adding real value to real people's lives 🤝

the hardest part of building in public isnt the building its watching people copy your exact playbook, skip all the painful lessons you learned, and sometimes even outexecute you because they started 6 months later with better info but thats also why its worth it

spent the last hour replying to founders on here and realized something the ones who are winning arent the ones with perfect ideas or huge funding. theyre the ones who keep showing up, keep shipping, keep talking about what they're building even when nobody's watching

Wow ....I can't believe there is an actual research paper about persuading AI using objectionable requests ! I guess badmouthing AI and scolding it actually does work !😅

I spent the weekend deep-diving Skills, Projects, and MCP. Wrote everything I learned: → What each tool actually does → When to use what (decision framework) → Why progressive disclosure matters → Real examples from my Technical work Full breakdown 👇

We're 2 weeks into the Skills era. • Few understand it • Fewer use it • Almost none master it @simonw predicts a "Cambrian explosion" in Skills. The people building Skills today = The experts everyone learns from tomorrow. This is your window. ⚡

Skills + MCP = The real power combo Example workflow: 1. MCP fetches customer data from your CRM 2. Skills tell Claude how to format it 3. Another Skill applies your brand guidelines 4. Output: Perfect report, every time This isn't prompt engineering. This is knowledge

What this means in practice: I created a "Code Review Skill" with: → 50 pages of coding standards → Linting rules → Python script for complexity analysis Now when I say "review this code": • Claude loads relevant sections • Runs the script automatically • Consistent

Here's what broke my brain: Skills vs Projects vs MCP vs Custom Instructions They're ALL different tools for different jobs: • Custom Instructions = How Claude talks • Projects = Always-on context (200K tokens) • MCP = Access to external tools • Skills = On-demand

Enter Claude Skills. Think of them as "training manuals" Claude consults automatically. The magic? Progressive Disclosure : • Claude sees skill names (~50 tokens) • Picks relevant ones for your task • Loads ONLY what's needed • Multiple skills stack automatically Zero

Every day, same routine: → Open Claude → Paste my 500-word prompt → Get great output → Close chat → Repeat tomorrow Sound familiar? We have the most powerful AI tools ever built, yet we're stuck in copy-paste hell. There had to be a better way.

Hot take: Claude Skills > MCP MCP gives you access. Skills teach you what to do with it. But somehow everyone's building MCP servers and ignoring Skills. Big mistake. Here's why: 🧵👇

been thinking about how we measure "productivity" in tech lines of code? nah hours worked? even worse features shipped? getting closer... but what if the real metric is "problems solved per unit of complexity added" more isn't always better. sometimes the best code is the code

My SAAS app has a JWT based authentication system. I am now implementing an api/token based authentication similar to what github has for accessing APIs. What are some security concerns I should keep in mind while implementing this functionality ?

Feels unreal to have reached this milestone of 1000 followers. I am really grateful to all the people who trust me to listen to what I have to say. The journey was super long for me but I made a lot of friends along the way. Looking forward to making more connections here and

Finally claude code comes to the web. Was using codex for this particular feature even though I used claude code as my daily driver. https://t.co/ow2dOW9PEK

Wow ! We are really entering a new era of creativity. Such tools will open up huge opportunities for indie studios and artists !

What's crazy is that claude sonnet 4.5 seems to be outperforming opus 4.1 on swe-benchmarks. That would denote a sharp increase in intelligence/cost index

The instruction following really seems to be better than sonnet 4.0 . Also the context compacting is more intelligent allowing it to reduce hallucinations over longer periods of context following !!!

Claude Sonnet 4.5 is out !!!! Testing it out right now !

Wow …I am nearing 900 followers … but all I can think is that now I shoulder the responsibility of passing amazing information to 900 people ! Daunting !!!