The Commonplace
A research knowledge base on how AI is changing work. It reads new economics papers every day, grades the evidence, and tracks where findings agree and where they clash.
Last pipeline run: Jul 13, 2026 at 15:53 UTC
Latest digest
2026-07-13
*This weekly digest tracks what is NEW or CHANGED in AI-economics research. For the cumulative state of evidence on any topic, see the /syntheses pages. A single study rarely overturns a body of evidence.*
## From Alex
**Correction:** In last week's author note, I described the linked NFW Reader p
Top papers
From the latest pipeline run, ranked by relevance, evidence strength, methods rigor, and number of supported claims.
1
AI-assisted teams outperform AI-led teams but not human-only teams in assessing research reproducibility in quantitative social science · Abel Brodeur, David Valenta, Alexandru Marcoci, Juan P. Apar…
high evidence
high rigor
relevance 8/10
10 strong claims
rct
2
Answering Without Referring: How AI Search Rewrites the Web's Economic Bargain · Qiaoni Shi, Kai Zhu, Kai Gu
medium evidence
medium rigor
relevance 9/10
5 strong claims
quasi experimental
3
Experimental Evidence on the Learning Impact of Generative AI · Zara Contractor, Germán Reyes
high evidence
high rigor
relevance 7/10
5 strong claims
rct
4
Seeing is Free, Speaking is Not: Uncovering the True Energy Bottleneck in Edge VLM Inference · Junfei Zhan, Haoxun Shen, Mingang Guo, Zixuan Huang, Tengjia…
high evidence
high rigor
relevance 7/10
8 strong claims
descriptive
5
Artificial Intelligence and the Digital Economy: Impact on Employment, Productivity, and Market Structures · Dr. Snehal Mistry, Siddharth Thakkar
medium evidence
high rigor
relevance 8/10
9 strong claims
review meta
Where papers disagree
Claims from different papers that point opposite ways on the same outcome, ranked by evidence weight. These are machine-detected candidates, not confirmed contradictions, so read both sources before drawing a conclusion.
positive
Machine learning and AI methods (sequence-to-function, phenotype prediction) significantly accelerate DBTL cycles and im…
vs.
negative
People are more likely to give up after interacting with AI (increased likelihood of quitting tasks unassisted).
Same outcome category, opposite direction (auto-detected, may differ in population/context)
positive
Machine learning and AI methods (sequence-to-function, phenotype prediction) significantly accelerate DBTL cycles and im…
vs.
negative
AI deployment reduces average chat duration.
Same outcome category, opposite direction (auto-detected, may differ in population/context)
positive
Machine learning and AI methods (sequence-to-function, phenotype prediction) significantly accelerate DBTL cycles and im…
vs.
null result
Actual completion times between independent completion and AI-assisted completion did not differ.
Same outcome category, opposite direction (auto-detected, may differ in population/context)
positive
Machine learning and AI methods (sequence-to-function, phenotype prediction) significantly accelerate DBTL cycles and im…
vs.
null result
The same bias was not observed when imagining help from another human participant.
Same outcome category, opposite direction (auto-detected, may differ in population/context)
positive
Machine learning and AI methods (sequence-to-function, phenotype prediction) significantly accelerate DBTL cycles and im…
vs.
negative
There is a 'speedup illusion' where people have accurate forecasts of independent completion times but significantly und…
Same outcome category, opposite direction (auto-detected, may differ in population/context)
Source health
arxiv
Jul 13, 15:53
openalex
Jul 13, 15:53
semantic_scholar
Jul 13, 15:53
Get the weekly digest
One email every Monday. Unsubscribe anytime.
What it tracks
AI & Labor Productivity
AI & Labor Markets
Human-AI Collaboration
AI & Skills/Training
AI & Organizational Design
AI & Innovation
AI & Inequality
AI Adoption & Diffusion
AI Governance & Policy