What’s the Use In Physics?

[epistemic sta­tus: this is mostly vague im­pres­sions I’ve ac­cu­mu­lated]

I’m an un­der­grad­u­ate physi­cist in­ter­ested in effec­tive al­tru­ism, and have no­ticed in sev­eral places (here, here) that physics is offhand­edly men­tioned as a promis­ing re­search area and/​or skill to pos­sess. How­ever, so far, I haven’t seen any speci­fics; no in­dex­ing of re­search top­ics in physics with DALY num­bers and in­struc­tions, or or­gani­sa­tions that hire physi­cists, or any­thing like that.

I can’t write that ar­ti­cle ei­ther, though to get some­thing in this space (hey, it’s highly ne­glected!) I’m go­ing to de­scribe some uses of physics and EA ar­eas where I’ve seen physics be done.

Not Do­ing Physics

The EA recom­men­da­tion I tend to see for physi­cists is to not do physics. I mean, no­body’s said that ex­plic­itly, but physics has been passed over in recom­men­da­tions for other fields for po­ten­tially good rea­sons.

Physics is, in­stead, use­ful be­cause it gives you good quan­ti­ta­tive skills, and a physics qual­ifi­ca­tion is use­ful be­cause it gives you a re­li­able sig­nal of quan­ti­ta­tive skills. In­stead, con­sider a more use­ful field to em­ploy the power of math­e­mat­ics, like AI safety (via com­puter sci­ence) or global pri­ori­ties re­search (via eco­nomics).

This is some­what un­in­spiring to those ex­cited by physics, though not as much as at first glance: AI safety is still about dis­cov­er­ing uni­ver­sal truths, and global pri­ori­ties re­search is still about us­ing math­e­mat­ics to pre­dict ex­per­i­men­tal out­comes.

Still, there are some things that can be done even within the purview of the physics de­part­ment, which is what the rest of this post is about.

Study­ing the Large-Scale Struc­ture of the Universe

Most of the value that a civil­i­sa­tion will pro­duce lies in the far fu­ture, and that means that the shape of the far fu­ture dic­tates how civil­i­sa­tion will be­have and how much value it will cap­ture. The long-term evolu­tion of the uni­verse is dic­tated by cos­mol­ogy. What­ever con­clu­sions are drawn have big im­pli­ca­tions in what be­havi­our we should ex­pect to see from other civil­i­sa­tions, if any, and po­ten­tially in what our civil­i­sa­tion should do.

Some work I’ve seen in this area is Bostrom’s Astro­nom­i­cal waste pa­per which fol­lows from knowl­edge about the size of the uni­verse and the rate of loss of us­able mass-en­ergy. Know­ing more about this would give bet­ter bounds on the trade­off be­tween ex­is­ten­tial risk and has­ten­ing tech­nolog­i­cal de­vel­op­ment (also rele­vant here: the scale of s-risks).

The be­havi­our of other civil­i­sa­tions helps put bounds on the Fermi para­dox, which might give us some in­sight into ex­is­ten­tial risk for our civil­i­sa­tion via de­ter­min­ing the ex­is­tence, size and lo­ca­tion of the Great Filter(s). In this area we have Sand­berg’s Spam­ming the Uni­verse pa­per (sum­mary: it is rel­a­tively easy to colon­ise the en­tire uni­verse), and the Aes­ti­va­tion hy­poth­e­sis pa­per (sum­mary: the best use of mass-en­ergy might be to wait un­til the uni­verse is very cold be­fore do­ing any­thing).

Also in this field is the idea of Boltz­mann brains. If these will ex­ist, they would make up most ex­pe­rience in the cos­mos, which means that re­search to find out what the deal with them is might be valuable.

Philos­o­phy and Foun­da­tional Research

Physics, as the sci­ence about how the world works at its most fun­da­men­tal level, has a thing or two to say about how the world works at its most fun­da­men­tal level. Coun­ter­in­tu­itive dis­cov­er­ies from physics might lead to eth­i­cal in­sights that we wouldn’t have de­vel­oped oth­er­wise.

Us­ing spe­cial rel­a­tivity, Beck­stead ar­gues in On the Over­whelming Im­por­tance of Shap­ing the Long Term Fu­ture that there should not be an eth­i­cal asym­me­try be­tween space and time: if two civil­i­sa­tions ex­ist si­mul­ta­neously and out­side each other’s light cones, then there is a refer­ence frame where one civil­i­sa­tion lives en­tirely be­fore the other. There’s no clas­si­cal analogue to this.

The many-wor­lds in­ter­pre­ta­tion of quan­tum me­chan­ics sug­gests that there ex­ists a mul­ti­verse, which might have some im­pli­ca­tions for what one should do: there are de­ci­sion-the­o­retic ar­gu­ments for why cer­tain ac­tions we take might af­fect other uni­verses, for ex­am­ple. Other po­ten­tial sources of mul­ti­verses in physics in­clude the pos­si­bil­ity of eter­nal in­fla­tion, or the math­e­mat­i­cal mul­ti­verse hy­poth­e­sis. More physics re­search could give in­sight into the dis­tri­bu­tion of wor­lds, and avoid con­fu­sion among non-physi­cist foun­da­tional re­searchers about what a mul­ti­verse ac­tu­ally en­tails.

There’s also Brian To­masik’s post on suffer­ing in fun­da­men­tal physics that you’ve read. Know­ing what fun­da­men­tal physics says about the world seems like it would be es­sen­tial for this ques­tion; even with a com­plete un­der­stand­ing of con­scious­ness it wold be im­por­tant to know what pro­cesses are im­ple­mented by physics to say any­thing at all about any ex­pe­riences, if any, that might be had.

Cli­mate Science

Tail-end risks from cli­mate change do have a chance of re­duc­ing the value of the long-term fu­ture (through, say, wiping out hu­man civil­i­sa­tion), and like ev­ery­thing that af­fects the long-term fu­ture that makes them very im­por­tant. Re­search into mod­el­ling the Earth’s cli­mate would give us a bet­ter idea of where these risks lie and how best to miti­gate them. This could be both di­rectly (via geo­eng­ineer­ing) and to bet­ter in­form policy de­ci­sions (such as by pre­dict­ing which pop­u­lated ar­eas will face the most heat­ing to fore­cast geopoli­ti­cal prob­lems).

This is also, as of the time of writ­ing, the topic of the only physics-re­lated the­sis on Effec­tive Th­e­sis.

Molec­u­lar Nanotechnology

Molec­u­lar nan­otech­nol­ogy seems in prin­ci­ple pos­si­ble, and also seems like it would lead to mas­sive eco­nomic growth if done prop­erly. Many of the prob­lems in molec­u­lar nan­otech­nol­ogy are physics prob­lems, since it in­volves han­dling ob­jects like in­di­vi­d­ual atoms on small scales.

A cru­cial con­sid­er­a­tion in nan­otech­nol­ogy is its effect on x-risk: in the worst case, there’s a grey goo sce­nario (do some physics to eval­u­ate the risk of this hap­pen­ing!), but even in or­di­nary cases nan­otech­nol­ogy might lead to weapons be­com­ing much cheaper and com­put­ers be­com­ing much cheaper (mak­ing the hard­ware over­hang much worse if AI safety isn’t solved yet). Nan­otech­nol­ogy x-risks seem to be very ne­glected; bet­ter knowl­edge of the fea­si­bil­ity of molec­u­lar nan­otech­nol­ogy might help with this.

Gen­eral Eco­nomic Growth Ac­cel­er­a­tion via Science

The stan­dard use of physics is to make dis­cov­er­ies, which are even­tu­ally con­verted into ac­tual ap­pli­ca­tions that do some­thing to im­prove some­one’s life later down the line. This doesn’t seem like it would be very ne­glected, be­cause physi­cists have a di­rect in­cen­tive to do this (an eas­ier time ap­ply­ing for fund­ing), though his­tor­i­cally it does seem to be the source of most of the good from physics. Since ev­ery­thing is a power-law, ex­cep­tional physi­cists might still be best placed here; I think Ein­stein do­ing physics was one of the best things that he could have done, sim­ply for be­ing so damn good at physics.

Other Things

And now I turn to you, Most Es­teemed Reader. There are, un­doubt­edly, things that I’ve missed, and prob­a­bly other ma­te­rial on this topic that’s been pub­lished el­se­where. What else could po­ten­tially be done with physics in or­der to do the most good? Is there any­thing in physics that seems much more promis­ing than other ar­eas?