Key Takeaways

  • Lenny Rachitsky points out that a growing share of modern tech work is simply cleaning up AI slop created by coworkers.
  • Molly Graham frames AI models not as omniscient partners, but as lazy summer interns who require onboarding, tight context, and constant coaching.
  • Forwarding raw AI output to colleagues is the workplace equivalent of handing an intern's unreviewed first draft directly to your CEO.
  • Executives who send AI-generated strategy memos signal to their entire company that outsourcing core thinking and shedding personal accountability is acceptable.
  • True efficiency requires moving conversations forward with vetted conclusions, while raw token generation only drowns teams in text they must spend hours wading through.

The Real Cost of Token Generation

Silicon Valley spent the last two years celebrating raw generation speed. Companies boast about writing ten times more code, generating thousands of words of documentation, and drafting instant customer replies. Yet actual organizational speed has stalled because teams spend their afternoons filtering noise.

As Rachitsky puts it: “A growing part of everyone's job is cleaning up the AI slop from other people trying to do your job.” When everyone can generate five pages of prose with one prompt, reading queues explode.

Graham distinguishes between sheer volume and real leverage. “Productivity is can you generate, you know, more, which at this point I think we've proven we can all generate a bunch of [trash] that other people then have to wade through,” Graham notes. “But is that efficient? Does that actually move the conversation forward? And the answer lately has been why no.”

The Lazy Intern Rule

To stop the spread of internal slop, Graham suggests a mental reset: treat every AI model like an unvetted intern who lacks context and takes shortcuts whenever possible.

“AI needs the same coaching and training that a human does,” Graham says. “It needs context, it needs onboarding, it needs all of these things. It's kind of much more like an intern, like a bad intern, often I call it a lazy intern.”

Nobody with good judgment takes a slide deck from a summer intern and emails it straight to the executive team without reading it. “You would never take the presentation that an intern handed you and just like straightforward it to your boss. Like, that is a little bit insane,” Graham explains. “You would iterate on it, you would edit it, you would make sure that it had all the context.”

When you prompt a model, you own the result. The moment you hit send, your name is on the work. If you did not verify the facts, tighten the logic, and cut the fluff, you shipped broken work.

Never Outsource the Strategy Memo

The most dangerous breakdown occurs when leadership adopts lazy habits. A founder who uses AI to draft a company vision memo or quarterly priorities destroys organizational trust.

Graham warns against this practice directly: seeing executives “shipping a strategy memo that was clearly written by AI, that is you role modeling for your organization that it's fine to outsource your thinking and it's fine to outsource your strategy. It's you role modeling that you don't need accountability for the work that you send.”

Strategy is not the act of filling a template with corporate buzzwords. Strategy is making trade-offs, choosing what not to do, and owning the outcome. When a leader prompts an LLM to decide or articulate direction, they abandon the primary duty of their seat.

What to Do With This

Tomorrow morning, audit the last three long messages or documents you sent to your team. If you copied raw text from ChatGPT or Claude without editing down the word count by at least 40%, rewrite them. Establish a team rule: any AI-assisted draft must be reviewed, fact-checked, and condensed by its author before anyone else is asked to read it.