Is It Better to Blog or Formally Publish?

Summary

This piece reviews some benefits and costs of formal academic publishing as well as publishing in a popular online newspaper or magazine. If what you want to write is not original, then consider adding the content to Wikipedia rather than reinventing wheels. Of course, there will be pieces (like this one) for which you judge that it's not worth the effort to publish formally, and in that case, publishing on your own website or blog is still an excellent choice.

Update, 2015: Currently I incline against publishing in academia most of the time, since I find that it takes a lot of effort to write papers in the style that a journal demands, while the payoff from having a journal publication isn't necessarily that big unless you're trying to get tenure. However, if you can get funding by being a grad student, the cost-benefit calculation changes and may make academic publishing a good idea.

Further reading: In Wiblin (2018), Will MacAskill offers a number of insights regarding the strengths and weaknesses of academic publishing. Click "Show full transcript" and then Ctrl+f for the text "so many biases in what gets published" to start reading that discussion. Here's one helpful passage:

My impression with intellectual research outside of academia is that you can make a ton more progress more quickly if you’re focused on something that academics aren’t focused on. So, like the question of, “What charity does the most good with your money?” Academics aren’t going to work on that question. If you want the state of the art, just don’t do that within academia. Then there’s questions that academics do work on, and often you get very high quality research on like meta-ethics, for example. When there’s an area that is suitable for academics to work on, then the highest quality thought, it seems to me, is actually going to be found within there.

Arguments for publishing in journals/conferences

I look favorably on the option of formally publishing research in an academic setting if you have the time to be this ambitious. Some reasons:

Signaling quality

Lots of people have blogs; many fewer publish academic articles. Formal publication is a standard way of distinguishing yourself from everyone else. Like formal education, one of its main functions is signaling. In order to get a paper published, it has to pass thresholds of quality and peer review.

Of course, different journals have different quality thresholds, so journal publishing provides a scale of signal strengths. That said, I think publishing at all is already a pretty strong signal, so I wouldn't let the improbability of getting published in a top journal scare you away from publishing somewhere. Neil McKinnon recommends starting with a middle-tier journal and then working your way down until you stop getting rejected.

Appearing more normal

Many of my views are controversial and may turn people away on first glance due to inferential distance. As David Pearce has pointed out, when your message is more unusual, it becomes more important to portray it in a conservative way. Formal academic publication can help separate merely unusual ideas from crackpot ideas.

Unfortunately, the Machine Intelligence Research Institute (MIRI) sometimes has a quasi-crackpot image not because its ideas are genuinely crazy but more because it publishes them amongst its followers rather than in formal channels; in contrast, Nick Bostrom—whose ideas are not really any more mainstream—is seen as more respectable by the outside world. MIRI's image has improved in recent years, in part due to some formal publications, both in conferences and popular online magazines.

Building on previous literature

I think this point is one of the most important. There is a lot of good scholarship already published in a lot of fields. There are almost no topics where I've found zero prior discussion in the academic literature. But the fraction of this extant literature that you would happen to have already stumbled upon in your lifetime is minuscule. Often you need to explicitly search for it to find it. Academic publishing maintains a principle of reviewing prior literature and then focusing your discussion on what's novel. This is a healthy discipline to have, and one that is typically absent from more informal blog-type posts.

When I reviewed the existing literature on reinforcement learning and robot ethics for an academic paper I wrote, I found dozens of fascinating articles that I would not have discovered had I not been summarizing prior work. In addition, I was forced to read articles by authors with viewpoints quite different from my own, and I appreciate the way this broadened my horizons.

Reducing insularity

Related to the previous point, writing in mainstream channels can open up the flow of ideas. The academic principle of discussing prior work forces you to examine many viewpoints, which is an epistemic virtue. And then by publishing in mainstream channels, you share your work with a wide audience, rather than just a more narrow group of people who already tend to think alike. In general, mainstream media have this benefit of exposing viewers to a broader array of ideas than they might encounter by staying within a familiar circle of friends or authors.

Formal and informal feedback

The peer-review process varies in its utility, but sometimes it actually helps you improve your ideas or catch errors. In addition, soliciting informal feedback on a draft of your paper may improve your thinking. Of course, the same kinds of feedback may accumulate in the comments of a blog post, but when the feedback gets incorporated into the paper itself, this is more helpful for readers, because they glean more insight per word they have to read, compared against following a string of debates in a comment thread. Of course, reading comments definitely has its place as well, but in general it's more valuable to read, or at least skim, a fully polished paper.

Citability

Academics are typically expected to cite other academic publications, or at least formal publications of some sort. Having at least the core ideas you want to express in the academic domain makes this possible. In addition, professors composing syllabi about an altruism-relevant topic may require that the assigned readings come from academic publications.

Preservation

Many journal-published articles are likely to be preserved for posterity, both in official databases like JSTOR and in unsanctioned ones like Sci-Hub. In contrast, blog posts tend not to be preserved by other people except Internet Archive, and preserved versions of websites on Internet Archive can't be found via Google.

Personally, I hope to keep my websites alive until the day I die (and ideally beyond), but if you don't have the time to devote to this task, then the risk of your blog eventually disappearing is an important consideration.

Arguments against publishing in journals/conferences

Time investment

The main barrier to publishing is often just that the process takes a lot of time and effort, and many people don't have the bandwidth for this, especially if they have another job. If you write a full paper, the topic needs to be important enough for you to spend possibly a few months conducting research on it.

For example, I would probably not write a journal paper on the topic of this essay. I think the topic is important to think about and important enough to share some ideas about, but I don't think it's necessary to read the literature on this issue extensively. Many people have probably already written very similar articles as this one, and I don't see a formal publication on this question as being highly counterfactually non-replaceable.

In contrast, I do think papers on suffering-focused ethical views, suffering from space colonization, insect suffering, and so on may be worth publishing formally, because these are issues on which I should become an expert. I ought to be reading the literature on these topics anyway, and in that case, writing a formal paper is a way to put that reading to good use rather than letting it sit in my head.

Still, writing a formal paper requires a higher degree of thoroughness than may be useful in practice. An 80/20 rule suggests that a lot of the important contours of a situation can be garnered from less-than-complete research on a topic. Publishing a formal article, however, tends to require more and narrower investment.

Grunt work

Formal publishing has more grunt work than informal writing, particularly as concerns formatting and citations, the conventions for which vary by journal. Probably the most useless part of formal publishing is compiling bibliographic information and checking its accuracy (since citation data from services like Google Scholar isn't always correct or consistently formatted).

This is a cost I don't incur with online publishing because it's typically unnecessary. For stable links, I often just hyperlink to an article and rely on Internet Archive to be able to retrieve the page if it later disappears. For links more likely to vanish, I often list the title of the piece in the "title" field of the hyperlink so that readers can search and find its new location relatively easily. Usually nothing more than a title and maybe an author is necessary for finding an article, yet formal citation style still asks for the journal volume and issue, page numbers, and even city of publication for a book. This is a big waste of time, but the convention hasn't yet gone away.

Slow turnaround

The turnover time between writing down your thoughts and having them available publicly is likely to be many months, or maybe even a year, when you publish formally. Of course, you can circulate the draft in private, but this may not always be sufficient.

Between submitting your paper and getting comments, several months may pass. During that time, you're likely to forget some of the details of what you wrote, which makes it harder to tweak or rewrite the article. You may have moved on to other projects, so editing an article submitted months ago may require a significant mental context switch.

Caprice of peer reviewers

In my experience, some reviewers may give unhelpful feedback or may reject your article because of their own personal views and stylistic preferences. Approval of your paper may then partly come down to luck of the draw. Michael Huemer agrees, at least with respect to philosophy journals: "After many years in the profession, it seems to me that referee decisions are mostly random. One referee’s report is no predictor of another referee’s report; one journal’s decision is no predictor of another’s."

Articles can't be changed or removed

I update the essays on this website very frequently—whenever I have a new idea, learn a new relevant fact, or change my views on something I wrote previously. In that sense, these essays are more like a private wiki than a publication. Being able to edit essays has been valuable for me because my views have changed a lot over time. I'm not necessarily likely in ten years to agree with what I write today.

Once published, formal articles can't be edited. This means no updating with new information, and it also means you can't take down naive views that you used to espouse. If you run for president, the words you wrote in your 20s may come back to haunt you. That said, you can publish a new article clarifying your change of position, and in time, people tend to catch up to this. From the perspective of information as a public good, it's also plausibly helpful to have a paper trail of the progression of a writer's thinking. Perhaps the old articles are still convincing to someone else, even if you don't presently share their intuitions. We can see an analogy to the concept of the "crowd within."

In general, I'm not too worried about this drawback because I like to be open and don't mind publicly changing my positions.

Narrowing of focus

Some ideas are academically interesting and readily publishable but not the most important from an altruistic standpoint. Some topics, especially state-of-the-art discoveries in computer science and neuroscience, may increase risk in the sense of potentially accelerating technology before society has a chance to build safeguards. And some topics are highly relevant altruistically but are not "academic material." While academic interest can play a role in the decision to write a paper, it's important to make sure there's altruistic value as well.

Keep in mind that there are many types of journals. Outside academia there are trade journals for various fields. If it's something that humans do a lot of, almost certainly there's a journal or other publication for it.

Potentially small audience

Many academic papers are read by just a few hundred people, or even only a few dozen. This is more likely for highly technical papers; discussions of ethics and altruism have potentially wider readership. Also, even if just a few people read your papers, those readers may be much more influential than a typical person and may take your ideas more seriously than someone browsing Slate would.

Still, academics don't typically optimize real-world impact directly; they optimize getting tenure, which often means publishing technical material in highly respected journals, without consideration for whether it's useful. Li-Shih Huang observes:

Advice offered by fellow academics that we find percolating on Twitter, in blogs or in newspapers are reminders of the downside of writing for these media. We are told that we "get points for high-profile publications. Points get you tenure." Or we're advised to "re-budget [your] time towards traditional research publications to build the strongest possible case for tenure," or that "outreach work is a 'stupid idea' and a huge career mistake."

She goes on to report that she has "been bluntly asked" to focus on high-impact journals rather than doing more direct outreach to professional communities and trade publications.

Thus, it's plausible that in some cases, your message is better taken to other venues than an academic journal. I think the topics I write about are relatively more journal-friendly because they're unusual and often focus on to long-term issues, but I certainly think more action-relevant writings often belong better in popular publications.

Also note that some academic papers do become very popular. For instance, sometimes bloggers pick up on a study and share its findings with a wider audience. Perhaps your publication in an ethics journal will wend its way to Wikipedia. (But I don't recommend adding it yourself, lest your addition be seen as spam.) Some papers by people like Peter Singer or Nick Bostrom are fairly widely read even among non-academics.

Paywalls

Many journal articles are paywalled, preventing most members of the general public from knowing how to access your work. Sometimes you can pay to make your journal article open-access, but often this costs several thousand dollars.

Publishing in online magazines and news sites

If mass readership is more important than signaling credibility, there's still another option besides blogging: Publishing in a formal online magazine or news site. To start off, you might aim for a lower-tier publication, to build up a profile of writings and avoid discouraging yourself. In the long run, with enough publications in increasingly popular outlets, it might be possible to get an op-ed in the New York Times—especially if you co-author it with someone who's already famous.

There are some benefits to writing newspaper-like articles:

1. Mass audience.
2. You can cover topics that aren't amenable to or publishable in the academic literature.
3. If you conduct interviews with experts for the piece, you'll expand your network of connections and learn new things.
4. You might get paid, depending on the publication. Remember, some people earn a living as freelance writers.

The costs are similar as for academic publications, including the extra time required to submit a single paper to possibly many venues until you stop getting rejected, the delay in waiting for publication, the static nature of the published article, and the possibility of pressure to write something interesting rather than something important.

Here are some examples of existing publications in popular media by friends of mine:

1. "Want to Kill Fewer Animals? Give Up Eggs" by Julia Galef in Scientific American blog
2. "Can We Really Upload Johnny Depp's Brain? by Luke Muehlhauser and Stuart Armstrong on Slate
3. Various pieces by Will MacAskill on Quartz.

If your organization has something to announce, consider distributing a press release. This way you get other reporters to do the work of writing articles for you. Of course, if you don't have something to make a news story about but instead want to research a topic where others are the experts, then you'll have to do it yourself—or persuade a writer friend to do it on your behalf.

What if your piece is rejected?

It seems unlikely that a good piece would be rejected from all places where you submit it. At a minimum, you could send it to a relatively obscure newsletter. But in the absolute worst case, if what you write is rejected by everyone, you can still publish it on your own website/blog. Among the costs of writing for formal publications, not being able to publish at all is not one of them.

You might have psychological hesitation due to fear of rejection, depending on your personality. I'm choosing to submit my first formal pieces to venues where I have a decent chance for them to be accepted, as a way to give myself positive reinforcement with my early attempts. Later on I may take longer shots.

There's a famous quote that "If you're not failing, you're not trying hard enough." Whether this is true I think depends a lot on what you're doing; in many domains, there isn't a clear definition of failure. But the idea has some merit in some cases, and this may be one of them. However, the advice may not work well for risk-averse personalities, so take it with sodium chloride.

Solo vs. social outreach

I personally love to read and write in private, without distraction. I learn best by absorbing the polished writings of others, so I find reading literature extremely valuable. And writing in response helps me generate, consolidate, clarify, and remember my own ideas. As a result, I find writing for an impersonal audience quite rewarding.

Other people may prefer more social engagement—conversations with other people they know, or one-on-one interaction. Such people may not enjoy solitary scholarship as much, but I think they can still play a very useful role in the process: namely, by popularizing the research of others and sparking discussions on important topics. Without channels for getting people excited, dry academic writings would accomplish much less in altruistic terms. Even Darwin needed his "bulldog."

I don't have a good estimate for the relative value of a marginal researcher vs. outreach person vs. organization builder, but it seems likely the values don't differ by more than a factor of 2 or something on average. Of course, for any given person, she may be much more suited to one path or another, and the difference in impact by that choice might be more than a factor of 2.

Where should I publish a given piece?

There's a good place to publish most altruistically relevant ideas that you might want to share, as is summarized in this table:

Optional: My history with this topic

Before late 2006, most of the authors I read were formally published, either in books/articles or on news/magazine websites. In fall 2006, I began reading the Felicifia and Overcoming Bias blogs, and in the following years, I started participating on forums. By 2011, I began to join Facebook discussions and met hundreds of new people that way.

Most members of the effective-altruist community currently publish informally, either on blogs, static webpages like this one, or just on Facebook or forums. Most of the organizations associated with the effective-altruism movement publish blogs and organization tech reports but less often formal venues.

Around 2008, I was discussing the work of MIRI (then named SIAI) with a friend at my college. He said MIRI would be taken more seriously if they published in academic channels rather than just writing informal papers and blog posts to a hobbyist crowd following. When I brought this up with some people who were more supportive of MIRI's approach, their reply was that academic publishing is slow and not optimal when your goal is to actually figure things out rather than just get tenure. I see merit in both sides of this debate.

Historically I've published by informal means, because I began writing these essays when I was just a student (around age 17), and I generally assumed my writings weren't of publication quality, although I did toy with the idea of submitting one or two pieces to a journal. As the years went on, I continued publishing online informally because I was too busy with school or my job to take on anything more ambitious. Now that I'm doing research on full-time basis (as of 2014), I have flexibility to consider more options for sharing my ideas.

In Mar.-Apr. 2014, I wrote an article on the ethics of artificial reinforcement learning for submission to a journal. I did this as an experiment to see what the whole publishing process would involve and whether I wanted to pursue it further. I had previously published a few papers in machine-learning conferences in 2009, but those were relatively more structured and required less background reading because they focused on empirical results.

In the future, I'm planning to try writing a magazine-style article as well. It would not be academic but would instead involve interviewing experts and synthesizing their ideas. I think it's valuable to try new things like this in order to broaden your skill set. I'm not so worried about what's optimally cost-effective now because I'm in the learning phase, and if I made estimates at this point, they could be unrealistic without first-hand experience.

Acknowledgments

Thanks to several friends, including Jonah Sinick, for inspiring some of these points.

Reader feedback

I shared this piece on the "Effective Altruists" Facebook group, and several people replied with additional insights. The comments include discussion of the authority of academic papers, circulating paper drafts ahead of time, and the value of book publishing.