🎮 The Next Input — Issue #162

Why Normal People Think AI is a Cult

Aaron Bost
April 14, 2026

In partnership with

⚡ The Briefing — 60 sec

Stanford report highlights growing disconnect between AI insiders and everyone else They think we’re loonies. We think they are loonies. Maybe we’re all just a bit nuts. The gap between people deep in this stuff and everyone else is getting harder to ignore.
Goldman Sachs chief ‘hyper aware’ of risks from Anthropic’s Mythos AI 👀 When people at that level start talking like this publicly, you know the conversation has shifted from “interesting model” to “serious institutional risk.”
Claude Mythos and Project Glasswing: why AI ‘superhacker’ has tech world on alert Mythos is not landing as just another model release. It is landing like a line in the sand.

🛠️ The Playbook — The AI Translation Engine

Mission
Bridge the gap between AI insiders and everyone else by turning frontier-model noise, fear, and hype into workflows people can actually understand and act on.

Difficulty
Intermediate

Build time
3–5 hours

ROI
Less confusion, better internal adoption, and a much stronger chance your team uses AI with clarity instead of suspicion, panic, or blind faith.

0) Why This Matters

There is a widening language gap in AI.

The people closest to the models talk in terms of agents, evals, capability jumps, safety thresholds, inference, and model behaviour. Everyone else hears some version of:

this will take your job
this thing is dangerous
nobody can explain what it actually does
the people building it sound insane

That gap matters because misunderstanding creates bad outcomes on both sides.

You get:

teams resisting tools they could benefit from
executives forcing adoption they do not understand
genuine risks buried under hype
serious breakthroughs dismissed as just more AI chatter

The play is not just to “educate the market.”

It is to build a translation layer that makes AI legible to normal operators.

1) Architecture

Component	Tool	Purpose	Owner	Failure mode
Signal layer	News / research / model releases	Collect relevant AI developments	Strategy / Ops	Teams drown in noise
Translation layer	GPT / Claude / internal prompts	Convert technical developments into plain-language implications	Operator	Oversimplification or fear-mongering
Risk layer	Review checklist / internal policy	Classify what is interesting, urgent, or dangerous	Leadership / Security	Everything feels equally important
Workflow layer	Docs / Notion / briefings / dashboards	Deliver AI understanding into real work channels	Operations	Insights never reach users
Feedback layer	Team questions / adoption feedback	See where confusion still exists	Team lead	Same misunderstandings repeat
Action layer	Training / workflow pilots	Turn understanding into practical next steps	Leadership / Ops	Awareness with no follow-through

2) Workflow

Capture major AI developments that could affect your team, customers, or workflows.
Translate each one into plain English with a simple framing: what happened, why it matters, what changes now.
Separate signal into three buckets: curiosity, workflow impact, and real risk.
Package the output into short internal briefings, decision notes, or pilot ideas.
Gather feedback from the team on what still feels unclear or overhyped.
Use that feedback to refine how you explain AI and where you introduce it in practice.

3) Example Prompts

Translation Prompt

You are translating an AI development for a non-technical business audience.

Explain:
1. what happened
2. why people in AI care
3. why normal operators should care
4. what is hype versus what is real

Keep it concise, clear, and grounded.

Risk Framing Prompt

You are reviewing an AI development for organisational relevance.

Classify it as:
- curiosity
- workflow impact
- serious risk

Then explain:
- why it belongs there
- what the team should do next
- whether leadership needs to care yet

Team Brief Prompt

Turn the development below into a short internal briefing.

Include:
- headline
- plain English summary
- likely impact on our work
- one recommended action

Keep it sharp and readable.

Mythos Reality Check Prompt

You are reviewing a frontier AI release.

Assess:
- what appears materially different
- what risks are being signalled
- what people are likely to misunderstand
- what a business should actually do in response

Return 4 bullet points only.

4) Guardrails

Do not confuse technical novelty with immediate business impact.
Do not water down real risks just to make people comfortable.
Avoid explaining AI in insider jargon.
Separate “interesting” from “actionable.”
Do not force adoption before the use case is clear.
Treat fear, hype, and confusion as translation problems first.

5) Pilot Rollout — 3 hours

Pick three recent AI developments your team is likely to misunderstand or overreact to.
Translate each one into plain English with a short implications brief.
Share the brief with a small team and gather questions.
Identify which parts created clarity and which parts still felt vague.
Turn one of the developments into a practical workflow pilot or policy note.
Repeat weekly until AI discussion inside the team gets sharper and less theatrical.

6) Metrics

Number of AI developments translated into internal briefs
Team understanding score after briefings
Reduction in repeated misconceptions
Number of AI pilots launched from translated insights
Leadership decision speed on AI topics
Ratio of hype stories to actionable stories
Adoption quality after briefing cycles

Pro Tip: A lot of AI resistance is not actually resistance. It is people being asked to care about something nobody has explained properly.

🎯 The Arsenal — Tools & Platforms

ChatGPT / Claude · useful for turning dense model news into plain-English operator briefings · ChatGPT · Anthropic
Notion · simple hub for internal AI briefings, definitions, and rolling notes · Notion
Google Docs · fast way to circulate short AI translation memos across a team · Google Docs
Internal briefing templates · the difference between “everyone is confused” and “everyone knows what this means”
Project Glasswing / Mythos coverage · a strong example of why frontier model releases now need translation, not just headlines · UQ News

Copy-paste prompt block:

You are helping me build an AI Translation Engine.

For the development below:
1. explain what happened in plain English
2. explain why AI insiders care
3. explain why everyone else may think it sounds insane
4. separate hype from real impact
5. classify it as curiosity, workflow impact, or serious risk
6. suggest one practical next step
7. turn it into a short internal briefing

Development:
[insert development here]

Return the answer in markdown with sections for:
- Plain English summary
- Why insiders care
- Why others may resist it
- Hype vs reality
- Classification
- Recommended action
- Internal briefing

💡 Free Office Hours

If your team feels like AI people are speaking a completely different language, I run free office hours to help translate the signal, cut through the noise, and turn it into something operational.

Book here: https://calendly.com