People are always measuring moving targets. They say 'AI helps with close reading' or something, but 'AI' is changing (and is so close reading). Instead of replicating results like that, I wonder if we should be trying to replicate measurement methods, which are often the actual contribution (e.g., 'we can measure close reading via its outputs in a conversation and how reactive').
If you are inspired by this idea, you can reach out to the authors for collaboration or cite it:
@misc{holtzman-measurement-replication-2026,
author = {Holtzman, Ari},
title = {Measurement Replication},
year = {2026},
url = {https://hypogenic.ai/ideahub/idea/CRGKvfn4LANe5NdaL58v}
}Please sign in to comment on this idea.
No comments yet. Be the first to share your thoughts!