[Discussion] Are academic papers a terrible discussion forum for effective altruists?

A few years ago, Alyssa Vance posted a bunch of rea­sons that aca­demic pa­pers are an im­perfect fo­rum for dis­cus­sion of top­ics such as those of in­ter­est to LessWrong. Lots of the rea­sons might ap­ply to effec­tive al­tru­ism too:

  1. The time lag is huge; it’s mea­sured in months, or even years.

  2. Most aca­demic pub­li­ca­tions are in­ac­cessible out­side uni­ver­si­ties.

  3. Vir­tu­ally no one reads most aca­demic pub­li­ca­tions.

  4. It’s very un­usual to make suc­cess­ful philo­soph­i­cal ar­gu­ments in pa­per form.

  5. Papers don’t have pres­tige out­side a nar­row sub­set of so­ciety.

  6. Get­ting peo­ple to read pa­pers is difficult.

  7. Academia se­lects for con­for­mity.

  8. The cur­rent com­mu­nity isn’t aca­demic in ori­gin.

  9. Our ideas aren’t aca­demic in ori­gin.

  10. Papers have a tra­di­tion of vi­o­lat­ing the bot­tom line rule.

  11. Aca­demic mod­er­a­tion is both very strict and badly run.

Of these, four is de­bate­able and eight is less true for effec­tive al­tru­ism than for LessWrong. But the rest seem as true and per­ti­nent for to­day’s effec­tive al­tru­ists as they were for the au­thor.

Diego Caleiro re­cently sug­gested that books might be su­pe­rior to pa­pers for pro­mot­ing his ideas.

[Books] are long enough to con­vey in­ter­est­ing ideas (of a more philo­soph­i­cal type) sec­ond, though they also fol­low ex­po­nen­tial dis­tri­bu­tions, there are sev­eral mar­ket­ing strate­gies that can aid pub­li­ca­tion and in­crease num­ber of read­ers. They can be, like pa­pers and un­like blogs, cited as de­cent aca­demic ev­i­dence in good stand­ing. They can be mon­e­tized as well, whereas one must pay to pub­lish pa­pers.

He con­tinues to put his point in the strongest pos­si­ble terms:

I am glad to hear coun­ter­ar­gu­ments, be­cause as it stands, this seems like the ul­ti­mate no-brainer, there is not a sin­gle thing pa­pers are bet­ter than books for.

  • Num­ber of readers

  • Con­vey­ing com­plex ideas

  • Get­ting feed­back on ideas

  • Money

  • Ca­reer prospects

  • Author’s name be­ing re­mem­bered and sought after

  • Re­silience over long stretches of time

  • Com­pat­i­bil­ity with Tech­nolog­i­cal ad­vances (cur­rent and ex­pected)

  • Odds of find­ing col­lab­o­ra­tors in virtue of hav­ing writ­ten them.

… The few prop­er­ties like “peer feed­back” and “short and eas­ier to write” that pa­pers beat books at have are com­pletely dom­i­nated by blog­ging, ’spe­cially on pub­lic sci­ence or philos­o­phy blogs.

So are books sim­ply su­pe­rior to aca­demic pa­pers for con­vey­ing world-chang­ing ideas? Now that peo­ple have been try­ing to ad­vance the case for the im­por­tance of safe de­vel­op­ment of ar­tifi­cial in­tel­li­gence (ar­guably one of the most im­por­tant ideas, and one that has been quite con­tro­ver­sial), we can ask which texts have most helped to ad­vance that idea so far.

In­stinc­tu­ally, Bostrom’s book Su­per­in­tel­li­gence seems to have been most helpe­ful for this. Although it has only been mod­er­ately cited, it’s in­fluence was felt in the dis­cus­sions that it pro­voked from Musk, Hawk­ing, Woz­niak and an ar­ray of other sci­en­tists and tech­nol­o­gists. It was also a best­sel­ler, and seems to have been a ma­jor con­trib­u­tor to the change in dis­course on this topic in the last cou­ple of years. The book stood out as a land­mark event. The Puerto Rico was similarly ground­break­ing—both helped grow the reach of Bostrom’s ideas in a way that dozens of pa­pers would not (and had not).

Rather, Bostrom’s pa­pers have tended to be non­con­formist, and their de­gree of read­er­ship and rep­u­ta­tion has suffered from this to a greater de­gree than his new book. In two other points of con­cern to Alyssa Vance, his book was also more widely available, and was ac­cessed in much wider parts of so­ciety. It was also ad­van­taged on some counts raised by Diego: it al­lowed Bostrom to con­vey un­usu­ally com­plex ideas, and made his name more well-known.

The ex­am­ple of AI safety seems to speak to the value of books and in-per­son con­ver­sa­tion.

If we look at the found­ing of CEA, we see a similar story, of ini­tial con­nec­tions be­ing made be­tween Toby Ord and William MacAskill, as well as with some in-per­son stu­dent groups at Oxford Univer­sity. Even though the con­text was aca­demic, jour­nals were not the main mode of pass­ing in­for­ma­tion or in­spira­tion. How­ever, it jour­nal­is­tic pub­li­ca­tion does seem to have helped peo­ple like Ord and MacAskill to earn a liv­ing and main­tain their aca­demic ca­reer prospects. More­over, in the cases of Bostrom and CEA, it may have been valuable to gain a sig­nifi­cant aca­demic rep­u­ta­tion be­fore perform­ing out­reach ac­tivi­ties. Such would cer­tainly be the case for Peter Singer, who also ad­mit­tedly gained most of his even­tual rep­u­ta­tion through books.

So aca­demic pa­pers aren’t gen­er­at­ing big break­throughs in in­ter­est, but a lot of the lead­ing thinkers tend to be aca­demics.

So how helpful does be­ing an aca­demic seem to be? Well, it’s not strictly nec­es­sary. If we look at the in­cep­tion of MIRI (when it was the Sin­gu­lar­ity In­sti­tute), we see Eliezer, who was writ­ing on on­line mes­sage­boards, and was aller­gic to academia. De­spite of­ten writ­ing in a way that was tech­ni­cal and in­formed by sci­ence, he was not in any aca­demic role, and had no aca­demic pres­tige. Eliezer is as always a bit of an anomaly, though, as his writ­ing tal­ents have re­peat­edly proven to be ex­tremely pop­u­lar, and this may have al­lowed his writ­ing to cut through in the ab­sence of aca­demic sta­tus. If you look some­where like Edge, TED or Big Think a lot lead­ing pub­lic in­tel­lec­tu­als and writ­ers are aca­demics. How­ever, it’s hard to dis­cern the cau­sa­tion. It could be that academia is equip­ping them to bet­ter pro­mote their use­ful thoughts, or it could just be that academia sucks in more than its share of the best thinkers in the first place.

So what rea­sons do we see for writ­ing aca­demic pa­pers? For a start, they can help one to iden­tify an aca­demic ex­pert, which is pos­si­bly use­ful. MIRI, CSER, FHI and oth­ers de­rive a lot of their rep­u­ta­tion (which in turn gives them some poli­ti­cal and per­sua­sive strength) in this way, as have Ord and Bostrom, and this might be good for any­one who wants to be a pub­lic in­tel­lec­tual. Another rea­son to write aca­demic pa­pers would be to be able to point to the ex­is­tence of an aca­demic field. This seems to be what MIRI wants to be able to do—to point tech­ni­cal and con­scien­tious AI re­searchers to such a body of work. Th­ese kinds of in­vest­ments will not pay off for some time, and may prove hard to eval­u­ate. For the rest of us, why write pa­pers? Well, it could be strate­gi­cally use­ful for those who wish to se­cure an oth­er­wise un­de­mand­ing pro­fes­so­rial po­si­tion. They might help us se­cure a re­search-based job. But oth­er­wise, and es­pe­cially from an al­tru­is­tic point of view, they don’t seem to do much. So Alyssa and Diego’s reser­va­tions about the value of pa­pers seem pretty com­pel­ling to me.