🎮 The Next Input — Issue #169

In partnership with

⚡ The Briefing — 60 sec

• Meta to cut one in 10 jobs after spending billions on AI ... Another week, another “AI investment means people are leaving the building” story. Meta is reportedly cutting about 8,000 jobs and closing 6,000 open roles while spending heavily on AI.

• Bosses don’t like the sound of a ‘four-day workweek’. Maybe it’s time to rebrand it Bosses hate it! Employees love it! The funny part is the four-day week is already sneaking in through flex time, half-day Fridays, compressed schedules, and generous leave. Just don’t call it that in front of management.

• AI Intelligence Ceiling Collapses: GPT-5.5 Pro’s Visual IQ Hits 145, Breaks Mensa Club Threshold GPT-5.5 is amazing. Throw your hardest tasks at it and get ready to be impressed. The benchmark chatter may be loud, but the feeling is simple: this thing is starting to make “hard” feel very negotiable.

🛠️ The Playbook — The Productivity Dividend Engine

Mission
Turn AI productivity gains into better operating models instead of defaulting to layoffs, burnout, or fake flexibility branding.

Difficulty
Intermediate

Build time
3–5 hours

ROI
Higher output, lower fatigue, and a cleaner way to prove AI is improving work rather than just shrinking teams.

0) Why This Matters

AI productivity is here. The fight now is over who gets the dividend.

One version looks like layoffs. Big companies spend billions on AI, then workers are told the machine has entered the chat and their badge has left the building.

Another version looks like better work design. The four-day workweek keeps getting framed as some radical employee fantasy, but plenty of workplaces are already using versions of it through flex time, compressed schedules, half-day Fridays, and generous leave.

And then there is the model capability side. GPT-5.5 seems like one of those releases where the obvious move is to throw genuinely hard work at it and see what breaks.

So the play is not:

• automate people out by default

• make everyone work harder with better tools

• pretend a four-day week is impossible while quietly doing half of it already

The play is:

• measure what AI actually saves

• convert saved time into better throughput or better work design

• decide deliberately where the productivity dividend goes

1) Architecture

2) Workflow

1. Pick one team with repeatable knowledge work and visible workload pressure.

2. Map the top 10 recurring tasks and estimate time spent on each.

3. Use AI on the hardest three tasks first, not the easiest vanity tasks.

4. Track time saved, quality retained, and human correction required.

5. Decide where the saved capacity goes: faster delivery, deeper work, fewer hours, or reduced backlog.

6. Review the outcome with workers before leadership declares victory.

3) Example Prompts

Hard Task Prompt

You are helping me test a frontier model on a genuinely difficult work task.

Task:
[insert task]

Requirements:
- reason carefully
- identify assumptions
- show where the answer may be uncertain
- produce a usable output, not a generic summary
- suggest one way to verify the result

Productivity Dividend Prompt

You are reviewing an AI-assisted workflow.

Calculate:
- time spent before AI
- time spent after AI
- quality difference
- human correction required
- where the saved capacity should go

Options:
- faster delivery
- fewer hours
- deeper work
- reduced backlog

Four-Day Rebrand Prompt

You are helping reframe a flexible work proposal so managers do not instantly reject it.

Avoid the phrase "four-day workweek."

Frame the proposal around:
- performance
- retention
- output quality
- reduced burnout
- measurable delivery

Return a concise internal memo.

Layoff Risk Prompt

You are reviewing whether an AI rollout is creating leverage or just becoming a layoff narrative.

Check:
- what work is actually being removed
- whether output quality improves
- whether remaining workers get more pressure
- whether the team has a fair productivity dividend

Return:
1. biggest risk
2. best use of saved time
3. one recommendation

4) Guardrails

• Do not count AI speed as a win if human correction wipes out the gain.

• Do not let productivity gains silently become more workload.

• Track worker pressure, not just output volume.

• Avoid using AI adoption as a lazy layoff story.

• Test frontier models on hard tasks, not party tricks.

• Rebrand flexibility around performance if the phrase itself triggers management panic.

5) Pilot Rollout — 3 hours

1. Choose one overloaded team and list its recurring tasks.

2. Select three hard tasks where AI might materially reduce time or effort.

3. Run those tasks through GPT-5.5 or your strongest available model.

4. Measure time saved, quality, and correction load against the old process.

5. Decide how the saved time will be used before expanding the rollout.

6. Share results with the team and adjust based on what actually improved their work.

6) Metrics

• Time saved per task

• Human correction rate

• Output quality score

• Worker workload pressure score

• Backlog reduction

• Percentage of saved time reinvested into deeper work or flexibility

• Number of hard tasks successfully improved by AI

Pro Tip: The real AI productivity question is not “How much faster can we go?” It is “Who gets the time back?”

🎯 The Arsenal — Tools & Platforms

• GPT-5.5 · throw the hard stuff at it and see where the capability shift actually shows up · 36Kr coverage

• Airtable · map recurring tasks, owners, time sinks, and workflow redesign decisions · Airtable

• Google Sheets · simple tracker for time saved, correction load, and productivity dividend allocation · Google Sheets

• Team retrospectives · essential for finding out whether AI made work better or just faster for management

• Flexible work policy docs · because sometimes the work model is already changing; it just needs a better label

Copy-paste prompt block:

You are helping me build a Productivity Dividend Engine.

For the workflow below:
1. map the current task steps
2. estimate time spent before AI
3. identify where AI can handle hard or repetitive work
4. estimate time saved after AI
5. identify human correction or validation required
6. recommend where the productivity dividend should go
7. define success metrics

Workflow:
[insert workflow here]

Return the answer in markdown with sections for:
- Current workflow
- AI-assisted workflow
- Time saved
- Quality risks
- Human validation
- Productivity dividend
- Metrics

💡 Free Office Hours

If your team is trying to use AI for real productivity gains without turning it into layoffs, overload, or corporate theatre, I run free office hours to help map the workflow and decide where the time should actually go.

Book here: https://calendly.com

The ops hire that onboards in 30 seconds.

Viktor is an AI coworker that lives in Slack, right where your team already works.

Message Viktor like a teammate: "pull last quarter's revenue by channel," or "build a dashboard for our board meeting."

Viktor connects to your tools, does the work, and delivers the actual report, spreadsheet, or dashboard. Not a summary. The real thing.

There’s no new software to adopt and no one to train.

Most teams start with one task. Within a week, Viktor is handling half of their ops.

Add Viktor to Slack for free.

🕹️ Game Over

Bosses want output. Employees want breathing room. AI might finally force the real conversation.

— Aaron Automating the boring. Amplifying the brilliant.

Subscribe: link