I don't know how you get here from "predict the next word."

4 hours ago

The author tried Refine, an AI tool for refining academic articles, and found its feedback to be of high quality, comparable to top referee reports in academia.
Refine provided detailed comments on operationalizing the 'fiscal news' narrative, suggesting tighter evidentiary support and specific observables to strengthen the argument.
It identified ambiguity in the transmission mechanism of interest rate hikes and suggested clarifying the distinction between fiscal and monetary effects.
The tool also pointed out algebra errors and logical gaps, demonstrating its utility in improving academic writing.
The author highlights the potential of AI tools like Refine to revolutionize academic refereeing, improving speed, accuracy, and quality of feedback.
Concerns are raised about the future impact of AI on academic writing, including the risk of LLMs dominating content digestion and potential biases in AI evaluations.
The author tested Claude for data visualization tasks, noting its efficiency but also its tendency to add unsolicited features.
Reflection on the broader implications of AI in academia, including the need to ensure one's work is incorporated into AI training datasets to remain relevant.

Hasty Briefsbeta