Experts who know their field and find AI drafts bland
No metric reported. Ruben Hassid says the first draft from this setup gets about 60% of the way, steering with short follow-ups gets it to about 95%, and you write the final 5%. He says the file takes 2 to 3 hours to build once and is reused for months.
Hassid's steering lines: 'Paragraph 2 is too generic. Make it specific to B2B SaaS.' 'Wrong angle. They don't care about features. They care about not looking stupid in front of their board.'
These are examples of writing I think are excellent. I'm building a markdown context file I'll upload to future conversations so you can match my standards. Analyze what makes each example work. Extract specific patterns I can reuse. Format each one as a rule starting with 'Always' or 'Never'. Now I'll tell you my personal constraints. These are rules I learned the hard way. Ask me questions one by one to extract them: what I never do, what I always do, where I've been burned before, what mistakes I see others make. Now ask me about my specific audience. Who they are, what they've tried before, what they're afraid of, what they won't say out loud, what makes them trust or distrust. Compile everything into a single context file I can reuse. Use this format: STANDARDS: [what good looks like] CONSTRAINTS: [my rules] LANDMINES: [where things go wrong] AUDIENCE: [who I'm writing for] --- Every time you write a post --- I uploaded my context file. Read it fully before starting. Then, I want you to do this task: write a LinkedIn post about [topic]. Before you write anything, list the 3 rules from my file that matter most for this task. Then give me your execution plan.
Legacy Builder Lab is Clifton Sellers' community for founders who write their own content with AI. Members get his prompts and bot setups, five new post templates every Monday and a monthly live teardown.
Join the Lab waitlist