Attention Deficit · weekly note · serial AD-W30-9C1 Roger Wong · July 16, 2026
The productivity squeeze

Faster became the new minimum.

Artificial intelligence (AI) saved an hour. The company added another task. A survey of 5,920 tech workers shows who gets the saved time and who gets a higher quota.

Based on Roger Wong's synthesis of Noam Segal and Lenny Rachitsky's 2026 tech-worker survey. The interaction at right is a thought experiment. The reported percentages appear below.
capacity converter · v2026 AI is the baseline
AI as baselinethe quota absorbs the saved timeAI as assistance

Start at the baseline. Slide right to return the saved time to the worker. This visual demonstrates Wong's argument, not a measured productivity ratio.

Same tool. Two completely different jobs.

Workers with agency keep the time AI saves. Everyone else inherits a higher baseline.

when the worker keeps the hour

The worker keeps the time

Experience turns automation into room for better judgment, deeper work, or fewer hours.

+ The team redesigns the role around judgment + The worker can spend more time on quality + Time saved stays visible as a benefit
when the company keeps the hour

The quota keeps the time

The old responsibilities remain while management raises the output baseline.

Same pay, higher expected throughput Human judgment carries more simultaneous work Every saved hour becomes capacity to fill
AI creates spare capacity. Job design decides who gets it.

Designers and researchers report the sharpest strain

Designers and researchers work through judgment, ambiguity, and influence. They reported some of the survey's highest anxiety.

designers
63%
overwhelmed by the pace of change
share of respondents in role
designers
61%
expected to do more for the same compensation
share of respondents in role
researchers
51%
anxious about their job security
share of respondents in role
founders
15%
anxious about their job security
share of respondents in role

Percentages as presented in Wong's synthesis of Segal and Rachitsky's survey. Role populations differ. The bars describe responses within each role, not equal-sized groups.

AI can remove the starter tasks where juniors learn judgment

Senior workers can use judgment built over years. Juniors built that judgment through first drafts, small fixes, and guided research.

the apprenticeship pipelineall stages connected
small fixeslearn the system
first draftslearn the bar
guided researchlearn the questions
independent workapply judgment
strategyshape the system
Pipeline breakIf the first three stages disappear, how does anyone reach judgment?

When AI saves an hour, who gets it?

Who gets the hour determines the job design. The model doesn't.

Source: Roger Wong, “How tech workers are feeling in 2026: a workforce splitting in two,” synthesizing Noam Segal and Lenny Rachitsky’s survey of 5,920 tech workers.

Read the source ↗