How I AIThe AI content machine that turns ideas into posts that don't sound like slop | Alex Lieberman
At a glance
WHAT IT’S REALLY ABOUT
Alex Lieberman’s AI content machine turns work signals into posts
- Lieberman built the “Content Machine” to scale high-quality content despite limited personal time and to make employee advocacy easy for busy full-time operators.
- The workflow starts with an “Oracle” that scans internal systems (Slack/Notion/meetings/Git/email, etc.) plus curated internet sources to propose and score daily “content spikes.”
- Instead of letting AI invent copy, the system uses an AI “Interview Panel” to interrogate the human for concrete stories, examples, and opinions, then drafts primarily from the user’s own transcript.
- A personal voice/style guide and a continuously updated “content lessons” file create a reinforcement loop so the system increasingly matches the creator’s tone and avoids repeated mistakes.
- Tenex complements the tooling with behavioral incentives via the “Tenex Creator Cup,” a points-and-prizes challenge designed to normalize employee posting and amplify recruiting and distribution.
IDEAS WORTH REMEMBERING
5 ideas“AI slop” is usually an input problem, not a model problem.
Lieberman argues the system outputs slop mainly when the human doesn’t provide strong ideas, real anecdotes, or specific context during the interview; the AI should shape good raw material, not fabricate it.
Use AI to kill blank-page friction before using it to draft.
The Oracle’s biggest win is ranking actionable ideas from your actual week (plus your reading list), which helps creators build a consistent habit even if they never let AI write the final copy.
Interviewing beats prompting for authentic voice and specificity.
The Interview Panel (Ferriss/Rogan/Barbara Walters, etc.) is designed to force examples, edge cases, and concrete “what happened” details—then the draft is assembled largely from the transcript.
Codify voice like a product spec, not a vibe.
Tenex builds per-person folders with role/context, target assets, preferred formats, exemplar posts, hook formulas, and “latent patterns,” so the model has stable constraints for how to sound.
Quality control can be automated with multi-critic scoring loops.
A Writer’s Council of writer personas scores drafts; if the aggregate is below a threshold (e.g., 9/10), the system runs revision loops until it meets the bar, creating consistent editorial standards.
WORDS WORTH SAVING
5 quotesCreating content has afforded me so many opportunities, but I am capped on the amount of time that I can spend creating content every day. And so I've been basically thinking about how can I re-engineer my content process to be AI native or AI assisted?
— Alex Lieberman
My take is that AI slop is hilariously people just pointing the finger at themselves and saying, "I'm not intelligent enough."
— Alex Lieberman
And so the only time, in my view, that the content machine actually produces slop is more of an indictment of the person not sharing good enough ideas during the interview step than the AI writing bad stuff.
— Alex Lieberman
The alternative before this content machine was people just not creating content.
— Alex Lieberman
Someone who has, like, the ability to understand systems at a really deep level... but who have the malleability to fully lean into making their workflows agentic.
— Alex Lieberman
High quality AI-generated summary created from speaker-labeled transcript.
