Most people collect prompt templates like trading cards. They chase every model launch the second it drops. Then they wonder why nothing sticks.
A recent video from Jeff Su flips that habit on its head. The creator argues you should skip 80% of AI advice and lock onto the 20% that still matters a decade from now. That reframing hit me harder than I expected.
*Ad
Have you been noticing how Claude is the most talked about tool on the internet right now? CLaude has been launching new models & features like claude co-work, claude design, skills, connectors literally every week.
The World’s first Claude-a-thon, A 2 day deep dive into Claude, its user-cases and 10+ other AI tools happening this weekend from 10 am-7pm EST – and they have just 1000 free seats for a limited time.
In the workshop, you'll do deep research on Claude. Build your own artifacts and dashboards. Create full presentations using Claude. Set up Claude Connectors like Indeed to automate your job search.
Register NOW! (free for the next 48 hrs)
🧠Live sessions- Saturday and Sunday
🕜10 AM EST to 7PM EST
Pick One Model And Go Deep
The old playbook: bounce between ChatGPT, Claude, and Gemini hunting for the "best" one. The author calls that a waste of energy now. The top models have clustered so close that the gap is tiny for everyday work, and skills you build on one carry straight to the rest.
His three rules for choosing land clean. Prioritize paid tiers, because free versus paid is night and day. Match the model to your work, then go with the vibes: the AI whose personality you enjoy is the one you'll actually use.
One more move he flags: change your defaults. Companies route you to the weakest model because it's cheaper to run. Pick the most capable one you have, every time.
Context Beats The Perfect Prompt
Here's where I got excited. The expert makes a bold claim: your prompt is no longer the biggest factor in output quality. Models got powerful enough to infer the role, format, and tone on their own.
His workout example sold me. Instead of writing a giant prompt about your equipment, schedule, and goals, paste an article on the push-pull-legs routine as context, then ask for a 4-day muscle plan based on it. The second version wins every time.
Say user_id. Get user_id.
Wispr Flow recognizes variable names, file references, and framework syntax mid-dictation. Speak your prompt, get developer-ready text for GitHub, Jira, or your editor. No mangled syntax. Ever.
*Ad
Connect Projects Into A System
The only framework worth keeping, the contributor says, is OC: Outcome plus Context. Name a real framework, show two or three examples of good work, and connect your tools. Your best context already lives in Gmail, Drive, and Notion.
Projects in ChatGPT and Claude (Gems in Gemini) store your rules, files, and memory. But each one is a silo. Your workout coach can't see the health report sitting next door.
This is where the system comes in. It pulls context across projects to catch what one alone would miss: flagging no cardio despite borderline cholesterol. Then it updates itself from your edits, so learnings compound. If you want the live demos and the storage tips, check out the original article.
The honest takeaway from the original breakdown: most people aren't at the system level yet, and there's no rush. Start at Level 1. Go deep on one model. Climb when you're ready. This is a smart approach for anyone tired of collecting tips that expire in a month, so read the full roadmap here.
Credits to Jeff Su.
Why 400,000+ professionals switched to Lindy.
Lindy reads your email as it arrives, drafts replies in your voice, and texts you a meeting brief before every call. Works over iMessage. One minute setup. Try it free.
*Ad
Quick poll
Where are you on the AI roadmap right now?
