Reposted by brendan o’connor
Looking forward to the Text as Data conference on Monday at UC Berkeley! I will be presenting our new paper "Multi-Perspective LLM Annotations for Valid Analyses in Subjective Tasks" which is a joint effort with Navya Mehrotra and @gligoric.bsky.social. 🧵
![The proposed rule change is a terrible idea - it would impose biased, arbitrary, political control over scientific funding, and greatly harm U.S. research, technology development, and higher education. The extremely biased language in the document indicates the worst of intentions to implement the policy. Also note:
- Congress already conducts robust oversight of federal grant-making agencies, including regular reporting and testimony from agency directors. For example, the National Science Foundation director is legally required to submit biennial reports to Congress (42 U.S. Code § 1885d), and in the past NSF directors and assistant directors have regularly given testimony to Congress. Congressional oversight is completely sufficient to investigate all of the issues alleged here, and remedy them if they turn out to be actual issues. Members of Congress have indeed issues critical reports in the past about particular grant-funded projects they wished scientific grant-making agencies to change, as is their proper right and authority.
[...]](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:rff2etl7xl6ffchnxgsg66bp/bafkreih2t6ehfpmlc3o3bwzumkxbdrywug7fcwduo2sf7uwjxf63hokg7q)
