Common Signals

← Back to comments

Codebook: TikTok comments on a two-minute Diary of a CEO clip of Nate Soares

Central claim

Advanced AI could become impossible for humans to control and poses a serious risk, up to human extinction, that justifies slowing or stopping development.

Stance note

agree = accepts the risk case, sides with Soares or Yampolskiy, or attacks McAfee; disagree = rejects it (AI is a tool, hype or fearmongering), sides with McAfee, or attacks Soares or Yampolskiy; mixed = explicitly holds both positions.

Stance Shared across the series

Agree
accepts the risk case, or sides with Soares or Yampolskiy, or attacks McAfee
Disagree
rejects the risk case, or sides with McAfee, or attacks Soares or Yampolskiy
Mixed
explicitly holds both positions
Unclear
about AI or the debate, but the position cannot be read
Na
does not engage with AI or AI risk (format, host, adverts, tags, emoji, off topic)

Frame Specific to this study

General agreement
Agrees with the risk case without giving a further reasone.g. "This really is the whole argument in one sentence."
Debater verdict
Judges the panellists (who won, who was rude, who talked too much) more than the argumente.g. "Pretty sure AI is already smarter than the bald guy"
Fatalism
Doom, too late, resignation, grief for the futuree.g. "We're all doomed...there's no other way of putting it."
Scifi
Reads the risk through a film, TV or book storye.g. "Strong 'Don't Look Up' vibes with this interview."
Blame people
The danger is the humans who control or misuse AI (greed, power, bad actors), not AI itselfe.g. "the main risk is people with such enormous power going rogue"
Human mirror
AI is dangerous because it is trained on and copies human naturee.g. "you design an artificial mind based on humans and it acts like humans"
Capability scepticism
AI is just a tool or LLM, it is hype, the fear is scaremongeringe.g. "AI is so dumb. What are you talking about"
Unplug
We can switch it off, cut power or destroy data centres, or rebuttals that we cannote.g. "Can't we just unplug it ?"
Takeover scenarios
Speculation about how AI would harm or outwit us: hiding, replicating, sabotage, competing for resourcese.g. "We won't be seen as cattle more like weeds"
Policy leaders
Regulation, slowing down, prosecuting executives, political leaders, international racee.g. "Roman should speak to congress"
Present harms
Harms happening now: jobs, education, dependence, misuse todaye.g. "we stopped paying attention to what's happening now"
Benefits
AI's upside, or that the risk is worth takinge.g. "I would love to use AI to discover a cure"
Religion spiritual
Religious or spiritual readinge.g. "No one here wants to consider Revelation ch 13?"
Conspiracy other target
Conspiracy theories aimed at groups or institutionse.g. "Scare mongering, just like covid, fals flag incoming"
Format only
About the podcast itself: format, adverts, guests requestede.g. "Thumbs down for all these advertisement breaks."
Other
Fits none of the above

Format reaction Shared across the series

Praise
praises the debate or clip itself
Mock
mocks the debate or clip
Condescended
feels talked down to by the debate or clip
Imitates
imitates or parodies the delivery
None
no reaction to the delivery

Emotion Shared across the series

Fear
No definition recorded.
Resignation
No definition recorded.
Anger
No definition recorded.
Humour
No definition recorded.
Hope
No definition recorded.
Neutral
No definition recorded.

Mentions a speaker's expertise or track record Shared across the series

Descriptive field: not a coded value with its own definitions.

Mentions governments, politicians, regulation or law Shared across the series

Descriptive field: not a coded value with its own definitions.

Mentions AI companies, executives or billionaires Shared across the series

Descriptive field: not a coded value with its own definitions.

Mentions a film, TV show, book or meme Shared across the series

Descriptive field: not a coded value with its own definitions.

Asks what people or society can do Shared across the series

Descriptive field: not a coded value with its own definitions.

Names or refers to Soares or Yampolskiy Specific to this study

Descriptive field: not a coded value with its own definitions.

Names or refers to McAfee Specific to this study

Descriptive field: not a coded value with its own definitions.

Mentions China or competition between countries Specific to this study

Descriptive field: not a coded value with its own definitions.

Refers to the OpenAI agent-swarm or Hugging Face incident Specific to this study

Descriptive field: not a coded value with its own definitions.

Written mainly in a language other than English Shared across the series

Descriptive field: not a coded value with its own definitions.

Download this codebook as markdown