Skip to content

Just Trying to Figure Out My Job

2026-08-07 · letter vlog · Hank · 5:29 · watch on YouTube

Topic: Internet & YouTube (also: Projects, Personal life) · mode: Personal essay · tags: timely hook

After a week of criticism of his public post about being enticed by LLMs, Hank warns against audience capture while valuing honest listening, thanks John, apologizes to Complexly staff whose work was questioned, and announces a personal AI policy: his job is to be himself and make fewer, better things.

What is said

The parts of the video and what is said in each (★ — how much it matters to the video).

00:00:00 — Audience capture, Hank's search for certainty, and how community members move each other

  • ★★ Hank explains audience capture: a creator takes signals about what the audience likes, dislikes or is angered by and slowly becomes that perspective, until creator and audience reinforce each other's certainty. — 00:00:00
  • ★★ Hank says he has been searching a lot for certainty, not just this last week (week of 27 July 2026) but over the last couple of years, which have felt destabilized. — 00:00:00
  • ★★ Hank worries about dynamics that bring people to a certainty he thinks reality does not deserve, since our record at predicting the future hasn't been great. — 00:00:29
  • ★★★ Hank says creators do not have zero obligation to their audience. — 00:00:29
  • ★★★ Hank says community works by listening to each other's deeper views, not to move everyone to one person's position but to be moved by understanding; that is how this community and the brothers' brotherhood have always worked. — 00:00:58 · Nerdfighteria, John Green

00:01:24 — Hank thanks John for his kind frustration and for not mentioning the novel's release

  • ★★ Hank tells John he is a good brother and thanks him for being frustrated with and worried about him kindly, with compassion and love the biggest part of his feelings. — 00:01:24 · John Green
  • ★ Hank thanks John for not complaining that this distracts from the imminent release of the novel he worked on for a decade, though he is sure John feels it. (playful) — 00:01:24 · Hollywood, Ending

00:01:54 — Hank's post about LLMs, bad-faith criticism, Complexly's AI policy and his apology to Complexly staff

  • ★★★ Hank says that last week (week of 27 July 2026) he posted publicly about being enticed by LLMs, using them to start projects, prototype a game and research many topics at once, until every problem looked LLM-shaped, which he calls dumb. — 00:01:54
  • ★★ Hank says that, because of how online discourse works, some of the response has been silly or bad-faith criticism. — 00:01:54
  • ★★★ Hank says no one is using AI to write Crash Course videos, whatever an AI-generated headline said. — 00:01:54 · Crash Course
  • ★★ Hank says Complexly's AI policy is years old, old enough to mention Bard, and extremely clear. — 00:01:54 · Complexly
  • ★★★ What Hank feels worst about: the reputational hit is his to take and earn back, but Complexly staff who write SciShow, Endlings, Eons and Crash Course and work with experts are now questioned because of him. — 00:02:23 · Complexly, SciShow, Eons, Crash Course
  • ★★★ Hank apologizes to the people who work at Complexly: they are so good, he is lucky to work with them, and whole teams there have each other's backs trying to help people understand the world. — 00:02:52 · Complexly
  • ★★ Hank says that whatever he does, in any context, reflects on Complexly, so he should be more careful. — 00:02:52 · Complexly

00:03:21 — Hank's personal AI policy, making fewer and better things, and his job being to be himself

  • ★ Hank says he should not be on Twitter, and everybody agrees. (playful) — 00:03:21 · Twitter
  • ★★★ Following Complexly's lead, Hank is creating an AI policy for himself, having become too reliant on LLMs for advice, editing and sourcing. — 00:03:21 · Complexly
  • ★★ Hank encourages creators to have an AI policy, even a private one, and wishes he had had one. — 00:03:21
  • ★★ Answering viewers who think he rushes to check boxes, Hank says he has never felt that way: he only makes videos he thinks are valuable and keeps no schedule. — 00:03:56
  • ★★ Hank stopped doing brand deals on Hank's channel at the beginning of the year because he did not like the lock-in. — 00:03:56 · hankschannel
  • ★★ Hank says his problem is the opposite: he wants to make too much stuff and is convinced it is possible. — 00:03:56
  • ★★ Hank likes making videos, but wants to be sure he likes making them, not just having them done. — 00:04:23
  • ★★ Hank says that people feel he might be checking boxes is a fault in itself, and being more choosy would bring only good outcomes for him, the content and the audience. — 00:04:23
  • ★ Hank says, honestly, there are not too few YouTube videos. (playful) — 00:04:52 · YouTube
  • ★★ Hank says he still lacks the certainty he would love to have about where we are now or where we are heading. — 00:04:52

  • ★★★ After a week of listening to the conversation, Hank feels more aware of his job right now: a huge part is to be himself, and making more things does not make him make better things. — 00:04:52

    “Making more things does not make me make better things.”

  • ★★★ Hank says people need to be able to absolutely trust that his takes are his own, which they are. — 00:04:52

00:05:26 — Sign-off

In the description and comments

  • ★★ The description says Hank has been hearing mostly honest critique and thoughtful perspectives rather than bad-faith nonsense or unconditional support: he should create honestly and make clear viewers hear only from him. — video description
  • ★★★ Hank's AI policy, rule 1: no portion of any script will be written, edited or outlined by an LLM. — video description
  • ★★ Hank's AI policy, rule 2: the thesis of a video will always originate with a human. — video description
  • ★★ Hank's AI policy, rule 3: no image or music in a video will be generated by AI, and anything included by accident will be removed as far as possible. — video description
  • ★★ Hank's AI policy, rule 4, added on August 7 after reading the comments: LLM outputs will never be trusted as a source. — video description
  • ★★ As a more general policy, Hank's informational and educational videos will link to primary sources in the description. — video description

Threads in this video

Where the stories that moved here went next

Mentioned

Central: Audience capture · Complexly · large language model

Also: Projects: Crash Course · Endlings · Eons · hankschannel · SciShow

Works: Hollywood, Ending

Organizations: Twitter · YouTube

Concepts: apology · artificial intelligence · sponsorship

Products: Gemini

Only in the description or credits: generative artificial intelligence

Also centred on large language model: How The Internet Is Rigged

← Previous: Esther Day 2026: Some Thoughts on Loss · Next → Hollywood, Ending