Historians have an increasingly strong incentive to tell dramatic stories

This is a system that's problematic for science, but crazy in humanities. One reason in science is it discourages replication. But there is a parallel here in something like history, reinvestigation. Something like the Roman Empire is well studied and there's lower novelty to be gained. But do we want to discourage such research? Discourage modern people for maintaining such records?

Under the current system, we can encourage these people to continue studying popular areas but do we encourage maintaining historical truth and integrity? I conjecture that the answer is no. It encourages making things up, especially to fit a modern narrative. It's a system that makes people not know they're lying because they are true believers and no credible experts push back.

This isn't a problem unique to the study of history, but you can find issues like this is all research in academia. It is very hard to measure the quality of research. Citations, h-index, and so on do not mean a lot. We've seen in the past few months very popular researchers found to be frauds, with high positions in prestigious institution like Stanford. Even in practical areas that should be treatable: psychology, medicine, machine learning, and more. We want to believe, but do we have strong evidence and strong understanding? We make easy metrics at first to be "good enough" or because we lack better ones. But we often forget the limitations of those metrics as time goes by because we get caught up in the rat race. You can't just ask yourself if the metric makes sense, but instead need to ask if there's an easier way to score on the metric that isn't the intended object of measurement. We can see here that citations can be hacked easily by fun stories, surprising results, and we'll never uncover them because the only people who attempt to replicate are those in junior positions like undergrads and grad students, who are more likely to question their results than the original work. Those failures never get written about.

Science is really about progressing knowledge. Publications are really about communication. Both these are very vague and can be accomplished in many different ways.

I say instead of making narrow easy to define metrics that will never be good enough we embrace the chaos and noise. Let the researchers be free and let ̶G̶o̶d̶ time sort it out. We don't know what is good research until 5-20 years down the line. We don't know who's a good teacher until the same. By removing the metrics you force nuance to not be ignored. Most of the time it's a "know it when you see it" definition and if that's true we have to be careful about levels of confidence. The thing that makes us human is the ability to handle nuance. It is what differentiates us from animals and LLMs. If the downside is that in a field where people are required to think hard and carefully have to... Think hard and carefully, then that's okay. ML is very successful with the arxiv model because despite much hype, noise, and even many bad actors, us researchers can sniff it out. We have to do it anyways, as that's our job.

I fear if we don't fix things we will continue to just degrade our foundation. We've seen this experiment fail over the last century. It's time to just get rid of the bureaucracy. Let academics be academics. Micromanagement isn't helping. Research can take decades, but people concerned with the next quarter will never allow that. Academics is about risks, challenging the status quo, and nuance. It is not to be run like a business. I am happy for this sector to not be profitable because I know the profits are indirect and benefit everyone. But returns on investments may take centuries, and that's okay.