EA Hotel with free accommodation and board for two years

[Posted about pre­vi­ously here and here]


Do you long to be free from ma­te­rial needs and be able to fo­cus on the real work you want to do? I know I’ve cer­tainly been in that situ­a­tion a few times in the past, but in­stead have lost time do­ing unim­por­tant and me­nial jobs in or­der to be able to get by fi­nan­cially. Ta­lented effec­tive al­tru­ists los­ing time like this is es­pe­cially tragic given that a lot of cause ar­eas are cur­rently con­strained by the amount of qual­ity di­rect work be­ing done in them.

Build­ings in the seaside holi­day re­sort of Black­pool (UK) are re­ally cheap. I’ve bought a 17 bed­room ho­tel with din­ing room, lounge and bar for £130k. As­sum­ing a 7% rental yield (which is rea­son­ably high), this works out at about £45 per per­son per month rent. Fac­tor­ing in bills, cater­ing, and a mod­est stipend/​en­ter­tain­ment bud­get, liv­ing costs could be as low as £5700/​per­son/​year (or lower for peo­ple shar­ing rooms, see bud­get). This is amaz­ing value for ho­tel liv­ing with all ba­sic ser­vices pro­vided.

The idea is to in­vite peo­ple to live there, with all their ex­penses cov­ered by donors, for up to two years. Fund­ing is already in place (via me) for the first year of op­er­a­tions. The pro­ject will be man­aged by some­one who lives on site and deals with all the ad­min/​fi­nances, shop­ping/​cook­ing/​clean­ing/​laun­dry, so­cials/​events and morale—they will also have free liv­ing ex­penses, and be paid a mod­est salary. Note that this should be con­sid­ered as a po­ten­tial high im­pact, high pres­tige sup­port­ing role, for those ex­cited to be in­volved in such a ca­pac­ity on an EA mis­sion. Guests will be free from con­cerns of ma­te­rial sur­vival, and be able to have pro­longed and un­in­ter­rupted fo­cus on what­ever pro­jects they are work­ing on. Ob­vi­ously these will be largely limited to purely desk-based, or re­mote work. Po­ten­tial peo­ple suited to be­ing guests are those who want to:

- self-study eco­nomics, philos­o­phy, sci­ence, AI, pro­gram­ming etc, in or­der to con­tribute to fu­ture efforts in var­i­ous EA cause ar­eas, with­out hav­ing to worry about—or waste time—get­ting (mostly ir­rele­vant or me­nial) paid work in or­der to fund themselves

- im­merse them­selves in desk-based re­search/​writ­ing on an EA-re­lated sub­ject with­out hav­ing to worry about fundrais­ing, a job, grants, teach­ing, tenure etc.

- work on a mostly soft­ware or re­search based start-up/​pro­ject(/​phase of a pro­ject) with­out hav­ing to worry about run­way (be it for profit with a view to earn­ing to give, or an EA-al­igned/​ad­ja­cent char­ity)

A some­what out of the way lo­ca­tion with lit­tle in the way of out­side dis­trac­tions most of the year, com­bined with fully catered liv­ing ar­range­ments, should be ideal for en­gag­ing in deep work (usual fo­cus­ing hacks ap­plied)

Whilst a sim­ple pledge to work on use­ful EA things is re­quired (“I pledge to work to­wards do­ing the most good that I can whilst stay­ing at the ho­tel”), there is no for­mal re­quire­ment for peo­ple to pro­duce more value than is put in to help them. Given the low costs in­volved, it’s a fairly low bar to clear though—pro­duce more value in a year than £5700 donated to the most effec­tive char­ity (+ op­por­tu­nity cost; for more see Value Propo­si­tion sec­tion). And just one ma­jor suc­cess will make the whole thing worth it. A use­ful anal­ogy might be a catered and man­aged uni­ver­sity res­i­dence (paid for by a grant), or per­haps even New York’s Chel­sea Ho­tel, which pro­duced billions of dol­lars worth of art, de­spite many guests not pay­ing rent (and was an in­spira­tion for this pro­ject). The slightly an­ar­chic feel to it with a lack of for­mal struc­ture does ap­peal too. Not re­quiring peo­ple to sub­mit de­tailed grant ap­pli­ca­tions is also some­thing that may seem like a breath of fresh air to peo­ple used to con­tinu­ally chas­ing fund­ing against the odds. This can only be fairly done though if as­sum­ing a rel­a­tively low de­mand. Which I ex­pect from it be­ing in Black­pool and not some­where more hap­pen­ing like Lon­don or the Bay Area. Of course if de­mand lev­els in­crease to be sig­nifi­cantly higher than sup­ply, then this can be re­assessed.

Read on for more de­tail, or skip to the fi­nal sec­tion (Next Steps) to get in­volved!


Life at the EA Hotel

Why Black­pool?

The Ho­tel Manager

Other uses for the hotel

Value proposition

Sus­tain­abil­ity, risks and organisation

Next Steps



Ap­pendix: Budget

Life at the EA Hotel

Fa­cil­ities at the ho­tel will be fairly ba­sic, but ad­e­quate for any­one ac­cus­tomed to liv­ing in an av­er­age house-share or stu­dent res­i­dence in a Euro­pean coun­try. In fact, given that two thirds of the rooms have en suite bath­rooms, it might be a cut above what many peo­ple are used to, at least in that re­gard. A sense of com­mu­nity will be en­couraged through shared meals and reg­u­lar so­cial­is­ing, but given the work fo­cus, per­haps the vibe will lean more to­ward “monastery” than “party house”, al­though cap­tur­ing some of the spirit of EA Globals would be nice.

Given the lo­ca­tion and the aes­thet­ics, liv­ing at the ho­tel will per­haps have more ap­peal for peo­ple who don’t care about sta­tus so much, at least in the usual ways (shiny things, cool lo­ca­tions, lux­u­ries, start-up style offices etc). It will ap­peal more to peo­ple who care about con­crete achieve­ments, and fur­ther­ing their EA goals, over sig­nal­ling (counter-sig­nal­ling of virtue notwith­stand­ing!).

Hav­ing low lev­els of noise and gen­eral dis­rup­tion most of the time will be re­quired for the ho­tel to be a pro­duc­tive en­vi­ron­ment (evening so­cial­is­ing ex­cepted). To this end, clean­ing will be done at set times (say a week­end ro­tor, with each room be­ing done at roughly the same time ev­ery week; week­ends be­ing when peo­ple are least likely to be ei­ther work­ing, sleep­ing, or in their rooms[1]), and sound in­su­la­tion can be put in if need be. The roads ad­ja­cent to the ho­tel are gen­er­ally very quiet, be­ing side streets. In the sum­mer, with many holi­day mak­ers in the area there is more noise com­ing from out­side, but so far (a cou­ple of weeks into holi­day sea­son) this hasn’t been too much of an is­sue. For work­ing, desks will be pro­vided in each room, and fast in­ter­net (broad­band ser­vices are cur­rently limited to “up to 76Mb”. There is rea­son­able 4G cov­er­age though). There are also plenty of ta­bles in both the bar and din­ing room, and seat­ing in the lounge, for peo­ple to work at out­side of meal times. Th­ese ar­eas will also be use­ful for group work, meet­ings, work­shops and events (see be­low). The bar (pic­tured be­low[2]) can seat 40 peo­ple, the din­ing room 32 and the lounge 12. Per­haps a con­ven­tion could be made for one of the large rooms to be for work­ing, and the other so­cial­is­ing, out­side of meal­times, evenings and week­ends. This would help peo­ple who benefit from hav­ing sep­a­rate work-and non-work lo­ca­tions, and help keep groups on task (less dis­trac­tion from other groups hav­ing a break and so­cial­is­ing).

EA Ho­tel bar (I pub­lished this post from the table in the cor­ner)

Given zon­ing reg­u­la­tions in Black­pool, the ho­tel will re­main nom­i­nally as a pub­lic fac­ing ho­tel. Whilst in gen­eral there will be no ad­ver­tis­ing to the pub­lic (on­line or el­se­where), and nor­mal prices will be set at the high end of the mar­ket, the op­tion to host non-EA guests will be there. This will likely only be ex­er­cised in or­der to bolster the fi­nan­cial health of the pro­ject if nec­es­sary; “No Va­can­cies” signs fea­ture in the front win­dows by de­fault. Hav­ing the odd few out­siders stay might be good for the health of the com­mu­nity though, to guard against it be­com­ing too in­su­lar.

Guests will be free to have their own guests stay in their rooms, but if they aren’t EAs do­ing EA work, then they will be re­quired to con­tribute to­ward costs (at some­thing like the cost price - £10-15/​day food and ac­com­mo­da­tion. See bud­get). The same will ap­ply for part­ners (costs for dou­ble rooms will be about 25% less per per­son). Most of the rooms are dou­ble rooms, and some are twin and even fam­ily rooms (three beds). Given this, EA cou­ples, or groups of two or three good friends stay­ing in the same room, could in­crease the ca­pac­ity of the ho­tel and fur­ther boost its ex­pected cost effec­tive­ness. Small groups are also wel­come even if not shar­ing a room, e.g. early stage EA star­tups, or already formed study/​re­search groups (like the AI Safety Camp). If pos­si­ble, they could have their own “wing” of the ho­tel, or even a whole floor.

There is a li­censed bar at the ho­tel. Main­tain­ing this could help with sus­tain­able rev­enue, al­though hav­ing some­one work on the bar could make it net nega­tive given the rel­a­tively lit­tle that EAs tend to drink. Per­haps the bar could be open in­fre­quently—say Fri­day nights and spe­cial oc­ca­sions. In any case, it will be best to re­strict its open­ing to a few hours in the evening, to min­imise noise and dis­trac­tion. How­ever, af­ter din­ner events will likely be a reg­u­lar fea­ture. Ex­pect peo­ple giv­ing talks about their work, struc­tured group dis­cus­sions, and games (I have got a cou­ple of old CEA/​FHI favourites—Re­sis­tance and Pan­demic, plus Pan­demic Le­gacy for a twist).

The food will be all ve­gan, with plenty of va­ri­ety and healthy yet tasty meals be­ing pro­vided (at a con­sis­tent qual­ity that will likely be sig­nifi­cantly higher than most who cook for them­selves are used to—think some­thing like this), all with­out guests hav­ing to plan, shop, cook, or clean up! Ve­ganism is fairly main­stream in EA now, given the pres­ence of an­i­mals in most peo­ple’s moral cir­cles, so hope­fully this won’t put off too many po­ten­tial guests[3]. Spe­cial dietary re­quire­ments (e.g. food aller­gies) will be catered for. Clean­ing and laun­dry will also be done for guests; sheets and tow­els changed once a week. Ba­sic concierge ser­vices will be pro­vided like tak­ing de­liv­er­ies of mail and pack­ages. Hav­ing not only free ac­com­mo­da­tion, but also free meals, and clean­ing and laun­dry done for them, will en­able guests to in­crease their pro­duc­tivity by giv­ing them not only 1-2 hours more time per day, but also a lighter cog­ni­tive load, and less stress. Whilst not the lap of lux­ury—no room ser­vice, or daily fresh sheets—this is still a pretty high stan­dard of liv­ing, for a very low cost (and free to guests).

With the all in­clu­sive ser­vices of set meal­times and menus, laun­dry and clean­ing, the ho­tel will be more suited to peo­ple who are flex­ible and easy go­ing. Op­tions for self-ser­vice laun­dry, and ba­sic food and drink prepa­ra­tion fa­cil­ities (ket­tle in room, microwave in com­mu­nal area) will be made available how­ever, for peo­ple who are more par­tic­u­lar with their eat­ing and clothes wash­ing. There is also cost-price meal-re­place­ment pow­der available via Queal, as per a gen­er­ous offer from founder Floris Wolswijk (with thanks to Rem­melt Ellen). Com­mu­nal meal­times will be en­couraged though, as a way for ca­ma­raderie and a sense of com­mu­nity to de­velop amongst guests. Friend­ships will be formed, prob­lems, and their solu­tions, dis­cussed, and some amount of fun and laugh­ter will be had too hope­fully! Also, some peo­ple may find that the in­evitable dis­cus­sion of work that will arise will lead to some amount of peer pres­sure that will aid in keep­ing them pro­duc­tive.

Such peer pres­sure could also work in re­verse too though, caus­ing peo­ple to dread hav­ing to talk to peo­ple, and hide away in their rooms when they are pro­cras­ti­nat­ing a lot. Or maybe it will have lit­tle effect and un­pro­duc­tive peo­ple will end up be­ing the most so­cial. Which could be good for morale, but per­haps dent oth­ers’ pro­duc­tivity (this was me, and some other Grad Bar reg­u­lars, dur­ing most of my PhD). Peo­ple slack­ing off/​pro­cras­ti­nat­ing too much is a po­ten­tial prob­lem in terms of how much value the ho­tel cre­ates. I have mixed feel­ings about this as some­one who pro­cras­ti­nated for the ma­jor­ity of my PhD. It was dur­ing that pro­cras­ti­na­tion that I read LessWrong, read about proto-EA stuff, re­searched 3D print­ing and open source hard­ware, which be­came a busi­ness for me af­ter, etc. Some­times hav­ing that kind of flex­i­bil­ity can lead to things that wouldn’t have hap­pened in more stringent en­vi­ron­ments. Giv­ing peo­ple free­dom to do what they want can be a good thing, as long as their mo­ti­va­tions are al­igned (i.e. are EA; see e.g.).

Others may re­spond to a more struc­tured set­ting though. Hav­ing at­tended the AI Safety Camp in Gran Ca­naria, where there is some over­lap be­tween the at­ten­dees and po­ten­tial guests at the ho­tel, I have up­dated in terms of po­ten­tial guests be­ing more con­scien­tious. Camp at­ten­dees were very dili­gent and pro­duc­tive, but some of this could be at­tributed to the ex­ter­nal de­mand of hav­ing to pre­sent their work at the end of the 10 days. To provide some struc­ture at the ho­tel, hav­ing some­one fulfil a role of “work co­or­di­na­tor” could be use­ful (H/​T Maia Pasek). Such a co­or­di­na­tor would be available for peo­ple to meet with to dis­cuss their work, and help re­solve any is­sues they may have (con­nect them to helpful peo­ple, sug­gest strate­gies for in­creas­ing out­put etc). I think it would be more fit­ting with the stress-min­imis­ing free liv­ing na­ture of the ho­tel to not re­quire manda­tory meet­ings with such a co­or­di­na­tor, but if peo­ple feel they need a mo­ti­va­tion kick, they could ask for reg­u­lar meet­ings. The ho­tel man­ager could per­haps fulfil this role dur­ing the 3 hours of office hours timetabled for giv­ing ad­vice/​men­tor­ing. In terms of guard­ing against peo­ple be­com­ing freeload­ers (or effec­tively squat­ters), per­haps there could be a manda­tory min­i­mum of very brief (one page) re­ports on progress to be filed ev­ery three months. There would also be some in­di­rect self-gen­er­ated pres­sure from guests wor­ry­ing about donors get­ting their money’s worth (al­though per­haps this will only be ap­pli­ca­ble for the more highly scrupu­lous). But then there would have to be a mechanism for chuck­ing peo­ple out if they are achiev­ing very lit­tle. The hard cap of two years on length of free stays per per­son will be an ul­ti­mate limiter, but it might be pru­dent to pro­ceed with cau­tion re­gard­ing tak­ing on long term res­i­dents (per­haps a three month pro­ba­tion pe­riod would be use­ful).

Truly strong rea­sons for eject­ing peo­ple from the ho­tel would be se­vere con­flict be­tween guests or abuse. To this end, a code of con­duct, and con­flict re­s­olu­tion pro­ce­dures would be put in place. Th­ese can be adapted from what EA Global has in place, for ex­am­ple. Guests and staff would be re­quired to agree to abide by them at the start of their ten­an­cies/​con­tracts. When it comes to the prac­ti­cal­ities of im­ple­ment­ing such poli­cies, hav­ing peo­ple skil­led and ex­pe­rienced in such mat­ters (es­pe­cially the Ho­tel Man­ager) would be ideal.

Guests who have prob­lems with ex­ec­u­tive func­tion and mo­ti­va­tion might benefit from var­i­ous meth­ods of self-im­prove­ment. It is ex­pected that these will largely be prac­ticed in a self-or­ganised man­ner amongst guests (with the en­courage­ment of the man­ager[4]), al­though the odd work­shop (such as CFAR) wouldn’t go amiss for those who are keen. Guests with prob­lems that ex­tend be­yond the fi­nan­cial and those which can be hacked away at with self-im­prove­ment meth­ods would benefit from hav­ing un­der­stand­ing peo­ple around. Per­haps a coun­sel­lor even. The Ho­tel Man­ager could dou­ble as a “morale officer”, but may not have enough time for many peo­ple need­ing help. Hav­ing an­other mem­ber of staff as a full time morale officer/​men­tor/​work co­or­di­na­tor would sig­nifi­cantly add to costs, but might be worth it in terms of po­ten­tial pay­off. Although hav­ing said this, I think just hav­ing a strong com­mu­nity will be enough to keep most peo­ple mo­ti­vated, in good spirits and pro­duc­tive.

Some guests might ap­pre­ci­ate hav­ing com­pan­ion an­i­mals around (e.g. cats or dogs), for com­fort, stress re­lief, or maybe even to help viscer­ally ex­pand their moral cir­cle of con­cern. The down sides are feed­ing them non-ve­gan diets, in­creas­ing wild an­i­mal suffer­ing (e.g. cats catch­ing birds) and food hy­giene is­sues (e.g. dogs eat­ing hu­man food in kitchen). On bal­ance I think the de­fault of no non-hu­man an­i­mals might be best, but if many guests want to live with one, and there are no strong ob­jec­tions from other guests, then it’s a pos­si­bil­ity.

Why Black­pool?

Prop­erty prices in the north of England are very low com­pared to the pop­u­la­tion cen­tres of the South-East. In par­tic­u­lar, towns that have seen their economies de­cline, such as Black­pool, are per­haps up­wards of ten times cheaper than Lon­don. Note that Black­pool is fairly close to Manch­ester (1hr drive, 1.5hr train/​bus), where the newly ac­tive EA Manch­ester is. It has the benefits of be­ing in a rich an­glo­phone coun­try with sta­ble and rel­a­tively easy to nav­i­gate in­sti­tu­tions. Although visas may be an is­sue for peo­ple out­side of the EU (and with Brexit, per­haps in­creas­ingly for peo­ple within the EU too). Alter­na­tives with similar low costs and rel­a­tive sta­bil­ity (and a bet­ter cli­mate!) can be found in South-East Asia (some work has been done be­fore re­search­ing low cost lo­ca­tions for EA hubs). How­ever, I have strong per­sonal ties to the UK, and as such, as fun­der and in­sti­ga­tor of the pro­ject, pre­fer it to be within rea­son­able reach.

Re­gard­ing qual­ity of life, Black­pool may be fairly dull most of the year (out­side holi­day sea­son), but there is a cer­tain charm to the bleak­ness of the beach in off sea­son, with its wide open vis­tas largely empty of peo­ple, the sea, and wheel­ing mur­mu­ra­tions of star­lings (see pic­ture be­low[5]). Walks fea­tur­ing long hori­zons are good for think­ing. The beach is just two min­utes’ walk from the ho­tel. There is also the pic­turesque and award win­ning Stan­ley Park just a mile in­land. The na­tional parks of the Lake District, and York­shire Dales, and sev­eral Areas of Out­stand­ing Nat­u­ral Beauty are also rel­a­tively nearby. And Scot­land isn’t that far. Any­way, given 10-20 EAs liv­ing to­gether, and reg­u­lar vis­i­tors, a vibrant com­mu­nity can be built around the ho­tel it­self. The lo­cal econ­omy may be in de­cline, but that shouldn’t mat­ter too much as we’re largely in­de­pen­dent of it in terms of the work we are do­ing. It’s nice to be in a place with lots go­ing on; a big city like Lon­don, or Manch­ester, for ex­am­ple. How­ever, when it comes down to it, many peo­ple who live in such vibrant places of­ten fail to take full ad­van­tage of them. One of the rea­sons my house­mates and I moved to Manch­ester last year was that it’s a big city with a lot go­ing on. How­ever, in prac­tice we rarely left the house, so as I said a few times—we might as well have been in Black­pool! Manch­ester is also only an hour to an hour-and-a-half away, should peo­ple stay­ing at the Black­pool ho­tel feel the need for more en­ter­tain­ment than Black­pool can provide.

Black­pool beach in December

Re­gard­ing Manch­ester, liv­ing costs are about dou­ble what they are in Black­pool, so still fairly cheap. If the ho­tel idea proves pop­u­lar, and there are peo­ple who would live at some­thing like that in Manch­ester, but not in Black­pool, then it may be worth a try do­ing some­thing similar in Manch­ester. Lon­don is prob­a­bly dou­ble the costs again, so at that point the bar is sig­nifi­cantly higher in terms of pay­ing for liv­ing costs be­ing pos­i­tive EV when com­pared with dona­tions to highly effec­tive char­i­ties. As­sum­ing most peo­ple work­ing at highly effec­tive char­i­ties have similar “big city” liv­ing costs, it’s then more of a level play­ing field in terms of “is this EA pro­duc­ing as much value on a day to day ba­sis as some­one work­ing at a highly effec­tive EA char­ity” (al­though you can take off the char­ity’s over­heads in terms of em­ploy­ing some­one, and em­ploy­ment taxes. More on this in the Value Propo­si­tion sec­tion).

The Ho­tel Manager

The Ho­tel Man­ager role is suited to some­one who’s hap­piest when as­sist­ing oth­ers in achiev­ing as much good as pos­si­ble; some­one who’s in­ter­ested in play­ing a vi­tal role on an im­por­tant mis­sion. If the pro­ject is suc­cess­ful, it will be­come a high im­pact po­si­tion.

The Man­ager will live on site and deal with all the ad­min, fi­nances, shop­ping, cook­ing, dish­wash­ing, clean­ing, laun­dry, so­cials, events and morale—they will also have free liv­ing ex­penses, and be paid a mod­est salary (£20,000/​year[6]). Ideally they will be a com­pe­tent cook. All the cook­ing, clean­ing and laun­dry for 17+ peo­ple is too much for one per­son to take on in ad­di­tion to man­age­ment tasks, so part time staff and ex­ter­nal ser­vices will be hired to to do some of these tasks. The laun­der­ing of bed linen and tow­els is con­tracted out to a ser­vice that costs ~£2/​guest/​week (which is eco­nom­i­cal when com­pared with hiring staff to do it and man­ag­ing them). Clean­ing will be done by ex­ter­nal con­trac­tors on week­ends. Per­haps one or two guests could vol­un­teer to help out in the kitchen for the evening meal (break­fast will be a sim­ple buf­fet of ce­real, toast, and fruit most days, per­haps a cooked break­fast at the week­end; and lunch will be sal­ads, DIY sand­wiches, crisps, fruit and other snacks, and lef­tovers from din­ner the night be­fore.) With part time staff and some small amount of vol­un­teer­ing and self-ser­vice amongst guests, there should be enough slack in the sys­tem for the Ho­tel Man­ager to have week­ends off and a stan­dard five weeks of holi­day per year. One or two guests can also be paid to take on crit­i­cal roles if nec­es­sary when the man­ager is away.

When the ho­tel was at full ca­pac­ity in the hey­days of the 1980s, there were four or five staff. How­ever, most of the work was restau­rant re­lated. Hav­ing serve-your­self buf­fet break­fasts and lunches, and a choice of say two op­tions for each din­ner, makes the food prepa­ra­tion a lot eas­ier when com­pared to an av­er­age ho­tel that has many op­tions on the menu and is mak­ing meals in­di­vi­d­u­ally or in small batches to or­der. So the job will not be on the level of full time chef and full time ho­tel man­ager in one. All nec­es­sary train­ing, such as at­tend­ing food hy­giene and li­censed bar man­age­ment courses will be pro­vided (the bar will prob­a­bly only be open one night a week).

Ideally, it will be good to have the Ho­tel Man­ager start as soon as pos­si­ble[7], so they can help get the ho­tel fixed up (some amount of re­pairs need do­ing) and meet­ing reg­u­la­tions (make sure all safety cer­tifi­cates and li­censes are in or­der). There is also work to be done in set­ting up the or­gani­sa­tional struc­ture—a Char­i­ta­ble In­cor­po­rated Or­gani­sa­tion, for which trustees need to be found[7a], a char­ter writ­ten, and in­cor­po­ra­tion doc­u­ments filed (see Sus­tain­abil­ity, risks and or­gani­sa­tion sec­tion be­low) - and get­ting it in a po­si­tion to spon­sor visas for over­seas guests. And back­end sys­tems for ad­min, like a web­site, book­ing sys­tem and ac­counts spread­sheets need to be built.

The role of Ho­tel Man­ager will re­quire a highly mo­ti­vated self-starter with a gen­er­al­ist skil­lset and good peo­ple skills. Hiring a com­pe­tent man­ager could end up as a “chicken and egg” situ­a­tion, whereby the pro­ject might need to be suc­cess­ful already in or­der to at­tract the nec­es­sary tal­ent. I have hopes that some­one from within the EA move­ment who sees the po­ten­tial in the pro­ject can be found though. In the event that there is no one suit­able can­di­date, the role could be split into two part time roles, per­haps along the lines of hav­ing a pub­lic fac­ing “com­mu­nity man­ager” and a back room “op­er­a­tions man­ager”.

Other uses for the hotel

With two large com­mon rooms—the bar and the din­ing room—and a smaller lounge, the ho­tel is also well-suited to host­ing work­shops, con­fer­ences, and re­search re­treats, such as the AI Safety Camp, or ra­tio­nal­ity work­shops. Maybe even cod­ing boot­camps. The bar has seat­ing for 40 peo­ple, the din­ing room 32, and the lounge 12. I’d say the ho­tel should be com­fortable up to about 35 day-time par­ti­ci­pants at a work­shop/​con­fer­ence, and 70-100 for a party. Th­ese could be run at very low cost, with the in­fras­truc­ture already in place. Of course, if the ho­tel is full with res­i­dents, al­ter­na­tive ac­com­mo­da­tion will need to be sup­plied for some (or all) of the event guests. It should be fairly easy to have ex­cess guests stay in nearby B&Bs if need be—the im­me­di­ate area (within 50m) is full of them, and the prices are gen­er­ally quite low (and the amount of ex­cess ca­pac­ity high). And even if the ho­tel is host­ing an event, it should still be pos­si­ble to feed the non-at­tend­ing guests at meal­times too if peo­ple sit in both the din­ing room and the bar.

The ho­tel was ad­ver­tised as 18 bed­rooms, but there are in fact 17 as there is no num­ber 13![8] There are also three rooms in the base­ment, and two in the at­tic, but these do not meet reg­u­la­tions so are for stor­age/​oc­ca­sional/​emer­gency use (I—and/​or other Trus­tees[9] - can stay in the at­tic if the ho­tel is oth­er­wise full, and the base­ment can be the nu­clear bunker). Most of the rooms are dou­ble or twin rooms, with a few sin­gles and triples (“fam­ily” rooms); 35 peo­ple could stay in to­tal if the ho­tel is filled to its cur­rent bed ca­pac­ity with most peo­ple shar­ing (see “Rooms” sheet here). Even more could stay were some of the larger rooms made into 4- or 6-bed dorm rooms with the ad­di­tion of bunkbeds. In this sce­nario, there are still more than enough bath­rooms (12 rooms have en suite shower rooms, and there is a stand alone shower on each floor).

In or­der to help keep the ho­tel both well oc­cu­pied and on a good fi­nan­cial foot­ing, rooms could also be offered for rent to vis­i­tors from the EA com­mu­nity who may or may not be work­ing on im­por­tant things dur­ing their stay. Peo­ple pledg­ing to do EA work will have prece­dence (and of those, peo­ple re­quiring a free stay). More on this in the fol­low­ing “Value propo­si­tion” sec­tion.

Another idea for empty rooms is offer­ing out­siders the chance to pur­chase a kind of “catas­trophic risk in­surance”; pay­ing, say, £1/​day to re­serve the right to live at the ho­tel in the event of a global (or re­gional) catas­tro­phe. When they make it to the ho­tel, they would then pay their share of the costs of liv­ing[10]. In the case of a nu­clear ex­change el­se­where in the world, peo­ple may want to leave big cities like Lon­don, and move to stay in a safer place with known al­lies. They would be pay­ing to re­serve their right to do so. Also, grow­ing the EA com­mu­nity in North­ern England in gen­eral could be seen as a hedge against x-risk, i.e lessen the num­ber of eggs (EAs) in the same, higher risk bas­kets (Lon­don and the South East). Black­pool might be hard to get to in the event of a catas­tro­phe, but the flip side of this is that there would be a lower risk from hos­tile ac­tors (mer­ce­nar­ies, mil­ita), as well as lower di­rect dam­age from nukes and fal­lout. How­ever, as a safe house it has the dis­ad­van­tage of it’s ad­dress be­ing pub­lic knowl­edge.

The cel­lars could serve as a nu­clear bunker of mod­er­ate pro­tec­tion. It will be rel­a­tively low cost to keep a stock­pile of long last­ing food down there, which could be slowly used and re­plen­ished by the kitchen over a 2-5 year cy­cle. There is already bath­room plumb­ing down there, and other es­sen­tials could be eas­ily added.

Value proposition

In terms of a value propo­si­tion, it will be hard to as­sign quan­ti­ta­tive value to the out­puts of this pro­ject with any pre­ci­sion. It’s es­sen­tially spec­u­la­tive in­vest­ing in the EA po­ten­tial of the guests. There will be com­pound­ing effects from those who suc­cess­fully “level up” in their abil­ities (such as study and re­search), and from the work they pro­duce. Fu­ture effects of the work done will also have a high var­i­ance, and in­volve many un­knowns. This is es­sen­tially a form of “hits based giv­ing”. And as with any hits based giv­ing, the baseline that it has to beat is us­ing the same re­sources on rel­a­tively well quan­tified low-risk in­ter­ven­tions, such as those pro­moted by GiveWell. In this case £5700[11]/​guest/​year donated to the most effec­tive char­ity, plus the op­por­tu­nity cost of what guests would oth­er­wise be do­ing, which pre­sum­ably would be fairly low in the situ­a­tion where stay­ing at the ho­tel is the best op­tion. Although per­haps the op­por­tu­nity cost could be of the same or­der—a cou­ple of thou­sand pounds a year given to EA char­i­ties (10% of a me­dian salary). Again, as with all “hits based” things, it’ll only take one ma­jor suc­cess for the whole pro­ject to be worth­while.

If there is not enough de­mand from EAs want­ing to stay for free to fill the ho­tel, empty rooms could be filled with pay­ing guests to help cover costs. Peo­ple pay­ing cost price (~£210-700/​per­son/​month, de­pen­dent on oc­cu­pancy, see bud­get) would keep costs the same re­gard­less of num­ber of pay­ing and free guests. Peo­ple pay­ing more would re­duce the cost of host­ing guests for free. In the event of empty rooms be­ing available, EAs would be wel­come to stay and pay cost price, whether work­ing or not, in or­der to boost com­mu­nity; peo­ple not com­mit­ting to EA stuff are wel­come to stay and pay at a markup. In the case of half the guests pay­ing at a markup and half be­ing hosted for free, costs for host­ing the free guests would be re­duced by the markup. As an ex­am­ple, as­sum­ing all 16 guest bed­rooms are filled by a sin­gle oc­cu­pant, cost price (ex­clud­ing stipend) is £330/​month per guest. In the case of eight free/​cost price guests, and eight guests pay­ing at a 50% markup, costs will be £165/​month per free guest, or £3500/​year fac­tor­ing in the £30/​week stipend. At the ex­treme, charg­ing a 100% markup + £30/​week (£825/​month[11]) for half the rooms would make the ho­tel rev­enue neu­tral, and there would be zero cost for host­ing eight free EA guests and the Ho­tel Man­ager. This (£780/​month) com­pares favourably to mar­ket rates of be­tween £15 and £60 a night for ho­tels and similar for Airbnbs, es­pe­cially con­sid­er­ing those prices don’t in­clude food. There may be scope to roll this out as a sus­tain­able (non-profit) busi­ness model for EA ho­tels around the world.

Fac­tor­ing in shorter stays with gaps, the av­er­age yield from rooms with pay­ing guests would be re­duced, so even with half the rooms set aside for pay­ing guests, if charg­ing a 50% markup, costs for the free guests might re­al­is­ti­cally be re­duced by 10-35%. Costs could per­haps be fur­ther offset by charg­ing for meals sep­a­rately at a higher markup. Per­haps £2.50 for break­fast, £2.50 for lunch, and £5 for din­ner, v.s. £2.75/​day cost price for food. Separat­ing out the room and board costs might lead to more peo­ple eat­ing out though, and over­all less rev­enue com­ing in.

For many guests, it is an­ti­ci­pated that the ho­tel will act as a step­ping stone to big­ger things. A free stay of up to two years should be suffi­cient to com­plete a large enough amount of work to en­able good fu­ture op­por­tu­ni­ties; ones that would have not been read­ily available oth­er­wise, be they jobs at EA or­gani­sa­tions earned af­ter study­ing rele­vant fields, or in­vest­ment op­por­tu­ni­ties earned af­ter the cre­ation of a suc­cess­ful start up. The ra­tio­nale for hav­ing two years as a de­fault timescale is that it’s long enough to get deep into work, with­out hav­ing to plan next moves for a while, but not so long as to overly en­courage freeload­ing or de­pen­dence. How­ever, in ex­cep­tional cir­cum­stances the limit will be open to be­ing ex­tended (say a ma­jor piece of work is not quite finished, or high value in­de­pen­dent re­search is be­ing con­tinu­ally pro­duced).

In or­der to help max­imise the flour­ish­ing of guests’ po­ten­tial, hav­ing an ad­di­tional mem­ber of staff as a full time morale officer/​men­tor/​work co­or­di­na­tor could well be worth it, not only for the guests them­selves, but in terms of po­ten­tial pay­off in value cre­ated.

Some of you may be think­ing: in­stead of go­ing to all the trou­ble buy­ing a ho­tel and start­ing a new or­gani­sa­tion, why not just give peo­ple the money in­stead? Well, for it to be equiv­a­lent, they would have to be liv­ing (or go to live) some­where where the costs of liv­ing are com­pa­rable, oth­er­wise I would effec­tively be buy­ing them time at a much higher cost. For ex­am­ple, some­one fru­gal liv­ing in Lon­don might be able to get by on £15,000 a year. So for the same costs they would get about a third of the time; and this is be­fore fac­tor­ing in the free time-sav­ing ser­vices (cook­ing, laun­dry, clean­ing etc). So un­less they were ~4x as pro­duc­tive as the av­er­age ho­tel guest, this would be a bad deal for me as a donor[12]. Also, the com­mu­nity as­pect of hav­ing a sig­nifi­cant num­ber of EAs in the same place is prob­a­bly worth some­thing too in terms of in­creased fo­cus, col­lab­o­ra­tion and morale boost­ing pro­duc­tivity. Com­mu­nity build­ing via deeper in-per­son ties is be­com­ing in­creas­ingly im­por­tant to the fur­ther de­vel­op­ment of the EA move­ment (see here and here).

If peo­ple were to live some­where like South-East Asia, they could prob­a­bly get the equiv­a­lent bang for their buck in terms of liv­ing ex­penses. In fact, I know of a few EAs who are do­ing this already. To be con­sis­tent I, or the or­gani­sa­tion this be­comes, should also offer grants for bud­ding EA digi­tal no­mads. It will be even bet­ter if a group of digi­tal no­mads formed a hub so there was some kind of fo­cal point for peo­ple want­ing to do this.

When CEA was start­ing out, they man­aged group houses in Oxford for their work­ers and vol­un­teers. After a while, once Oxford was more es­tab­lished as an EA hub, they de­cided it best to free up their valuable time and let their staff self-or­ganise their own ac­com­mo­da­tion and shared houses. The same model could be ap­plied for EAs form­ing hubs in cheap lo­ca­tions. Per­haps this pro­ject could act as an ini­tial fo­cal point, and man­age ini­tial ac­com­mo­da­tion in each lo­ca­tion. Maybe the um­brella or­gani­sa­tion could be called “EA Ho­tels and Low-cost Liv­ing”?

For con­sis­tency, grants should be given in gen­eral to any EA whose liv­ing costs are low enough. Would this cause per­verse in­cen­tives though? All else be­ing equal, it’s much bet­ter to give promis­ing peo­ple an amount that al­lows them a lot of free­dom and slack. How­ever, this makes the value propo­si­tion pro­por­tion­ally lower. Pro­vid­ing grants only at low lev­els of liv­ing costs is a bet­ter in­vest­ment for the donor, but it also leaves a gap for peo­ple who don’t want to—or aren’t in a po­si­tion to—rad­i­cally al­ter their lifestyles but still need their liv­ing costs cov­ered in or­der to do (more) use­ful EA things. Given that EA (as a move­ment) seems to have plenty of cash at the mo­ment, per­haps there is room for this space to grow. I can en­visage a hi­er­ar­chy where the big­ger grants have more stringent de­mands and more com­pe­ti­tion. I guess this is just the cur­rent non-profit (and for profit!) land­scape, but for in­di­vi­d­u­als in­stead of or­gani­sa­tions. There are already many such grants in the forms of prizes and schol­ar­ships at the high end, but not much at the low end, so this is per­haps where the op­por­tu­nity is.

There has been some de­bate re­cently about fund­ing peo­ple/​pro­jects at the low end and whether donors are be­ing too risk averse. I think we’re prob­a­bly at the point where there are more false nega­tives than false pos­i­tives, so more chances can be taken on peo­ple at the low end. Scam­mers can be avoided, at least in the first in­stance, by re­quiring a prior his­tory of in­volve­ment in EA and refer­ences from peo­ple with stand­ing in the com­mu­nity. The big­ger risk is well in­ten­tioned but in­effec­tive grant re­cip­i­ents (/​ho­tel guests) just not de­liv­er­ing. How­ever, we are likely already in a situ­a­tion where EAs fre­quently go off on their own to work on in­di­vi­d­ual pro­jects that don’t de­liver, but we don’t hear much about it be­cause of se­lec­tion effects/​so­cial de­sir­a­bil­ity bias (though there are wel­come ex­cep­tions that come to mind[13]). The ho­tel scheme makes this more trans­par­ent if it is in­deed oc­cur­ring, and it gives us a testbed to try out in­ter­ven­tions (H/​T John Maxwell). Such failure to de­liver will be miti­gated by the low costs in­volved (mean­ing that there is a lower bar to what level of pro­duc­tivity could be con­sid­ered effec­tive), but is nev­er­the­less some­thing to keep in mind.

It’s worth not­ing that in a situ­a­tion where a po­ten­tial in­vest­ment has high risk, but also has high up­side po­ten­tial, the an­swer to the ques­tion of how much to in­vest is of­ten “a small amount” as op­posed to “zero” (Pas­cal’s mug­ging does not ap­ply when there aren’t that many similar op­por­tu­ni­ties, and cer­tainly not enough to come close to drain­ing all your cap­i­tal. This is cer­tainly the case with the cur­rent EA move­ment[14]). There is also still the risk of “ide­olog­i­cal nepo­tism” (“re­ward­ing peo­ple solely for as­cribing to the right val­ues and know­ing the right memes”) (H/​T David Krueger) to con­tend with. To a cer­tain ex­tent this plays a role in any com­mu­nity that has shared be­liefs and cul­ture; nev­er­the­less, be­ing aware of it should help re­duce it. The dy­nam­ics of sta­tus and pres­tige in the non-profit world seem to be geared to­ward be­ing averse to risk-of-failure to a much greater ex­tent than in the for-profit world (see e.g. the high rate of failure for VC funded start-ups). Per­haps we need to close this gap, con­sid­er­ing that the bot­tom line re­sults of EA ac­tivity are of­ten con­sid­ered in terms ex­pected util­ity.

In line with the “hits-based giv­ing” an­gle, there are large po­ten­tial up­sides to this pro­ject. If the pro­ject is suc­cess­ful and ho­tel res­i­dents are able to gen­er­ate value in a way that’s clear to other EAs, then we’ve got a work­able pi­lot for de­creas­ing EA liv­ing ex­penses by a large fac­tor, mean­ing donor dol­lars can go that much fur­ther. There are also huge po­ten­tial up­sides in terms of com­mu­nity build­ing (imag­ine a year-round EA Global), fund­ing pro­jects that are less leg­ible to donors, and de­ploy­ing donor dol­lars in a way that re­duces per­verse in­cen­tives (the pos­si­bil­ity of per­verse in­cen­tives might be a large part of what’s hold­ing back EAs from giv­ing money to each other) (H/​T John Maxwell).

Why a Ho­tel? Ho­tels offer a high stan­dard of liv­ing (cater­ing, clean­ing, laun­dry ser­vices), and are effi­cient in terms of re­source use (cater­ing for many peo­ple at once, buy­ing in bulk[15]). Whilst nor­mally as­so­ci­ated with holi­days, they ac­tu­ally can be viewed as en­abling high pro­duc­tivity in gen­eral—whether that is a high pro­duc­tivity of leisure, or work. This list de­scribes the benefits, al­though ob­vi­ously the Black­pool EA Ho­tel will not in­clude ev­ery­thing on it. Here is a list of no­table peo­ple who have lived in ho­tels, and some dis­cus­sion. Quite apart from mat­ters of effi­cient liv­ing, there is a cer­tain ro­man­ti­cism about the idea of liv­ing in a ho­tel, evoca­tive of mid-20th cen­tury cin­ema. The EA Ho­tel: Black­pool will be at the low end, but may in­spire other pro­jects that take it to an­other level (see x-risk hub dis­cus­sion be­low).

Hav­ing a set-up where there is plenty of com­mu­nal space, but also pri­vate rooms for work­ing, gives the best of both wor­lds in terms of deep solo work and serendipi­tous col­lab­o­ra­tion, as illus­trated in Cal New­port’s hub-and-spoke model.

Other op­tions for low cost liv­ing in­clude large shared houses, and con­verted ware­houses. Some houses with many bed­rooms in Black­pool have similar low costs per bed­room as the cheap ho­tels; see this list from a pop­u­lar prop­erty web­site (it was search­ing on this web­site with a wide range and sort­ing by price low-to-high, and read­ing about the Chel­sea Ho­tel, that gave me this idea. Along with ini­tial prim­ing from re­search into ho­tels for com­mu­nal liv­ing by John Beshir). Large shared houses can benefit from economies of scale if ten­ants are highly com­mu­nal in their shop­ping and cook­ing. Without ex­plicit or­gani­sa­tion from the out­set, it is gen­er­ally rare for such high lev­els of co­or­di­na­tion to hap­pen though (I have not lived in a shared house where it has, de­spite the best in­ten­tions of many house­mates in sev­eral houses). Con­verted ware­houses are similar to large shared houses, but have the po­ten­tial ad­van­tage of sig­nifi­cantly lower rents. There are build­ings in the Black­pool area where it might be pos­si­ble (given plan­ning per­mis­sion) to make rooms for per­haps as low as £3000/​room (com­pared to £7500/​room for the ho­tel). How­ever, already once you are get­ting as low as £7500/​room, effec­tive rent costs start to be­come a small frac­tion of liv­ing costs (<20% See bud­get), so even re­duc­ing them to zero can only help so much.

Liv­ing costs in a shared house could be lower still than the ho­tel, but not by much more than ap­prox­i­mately 35%, which is the frac­tion of the above costs taken up by staff wages (Staff costs are ap­prox £30k out of an £87k an­nual bud­get[16]). So per­haps £3500/​year. Go­ing down to ab­solute min­i­mum lev­els of food spend­ing (£1/​day is pos­si­ble with bread and mar­garine, peanut but­ter, por­ridge, ba­nanas, car­rots and rice), costs could per­haps be as low as £2800 a year. EAs who want more in­de­pen­dent liv­ing can form group houses nearby, which would be open to be­ing funded by this pro­ject. Of course liv­ing in a shared house and do­ing all meals and clean­ing your­self might re­duce pro­duc­tive EA hours by a few a week, so fund­ing would prob­a­bly be given at a slightly re­duced rate com­pared to the ho­tel to fac­tor this in.

I haven’t found many good ex­am­ples of already ex­ist­ing pro­jects that are similar to this one. Newspeak House in Lon­don pro­vides fel­low­ships to peo­ple who share a com­mon cause (poli­ti­cal tech­nol­o­gists fo­cus­ing on digi­tal democ­racy); whilst fel­lows have ac­com­mo­da­tion pro­vided, they largely need to provide for them­selves. His­tor­i­cally, there has been the Chel­sea Ho­tel. There has been much dis­cus­sion in the EA com­mu­nity of hous­ing pro­jects, and form­ing new (low cost) hubs, but thus far, progress has been rel­a­tively slow, with a few group houses with EA, or part-EA mis­sions com­ing to fruition (e.g.). The hard part of all pro­jects of this kind is get­ting peo­ple to ac­tu­ally move![17] In the ad­ja­cent ra­tio­nal­ist com­mu­nity there are many more group houses, many with a pro­por­tion of EA res­i­dents. No real low cost hubs have formed as of yet though.

Black­pool Tower and tram­lines along the Promenade

X-risk fo­cused EAs in the Geneva area have un­der con­sid­er­a­tion, as part of a wider strat­egy, an x-risk hub, that would both be a fo­cal point for dis­cus­sion of x-risk with wider so­ciety (plugged in to the global diplo­macy scene there), and an in­cu­ba­tor for new tal­ent. The grand vi­sion is for some­thing much larger and more lux­u­ri­ous; a large com­plex for 100+ res­i­dents with games room, mu­sic room, sports fa­cil­ities, kids play area etc (al­though it would have apart­ments rather than a ho­tel for­mat). It would provide a pipeline for tal­ent to level up and move on to work­ing at x-risk orgs, as well as host­ing its own x-risk org fo­cused on global co­or­di­na­tion and im­ple­ment­ing the fruits of re­search. This might sound like a rather ex­trav­a­gant use of EA fund­ing, but given the na­ture of x-risk the re­turn in ex­pected value could eas­ily be well worth it (con­cerns about Pas­cal’s mug­ging notwith­stand­ing). Also, given that OpenPhil are likely to spend a lot more on x-risk in the near fu­ture, it is a dis­tinct pos­si­bil­ity that some­thing like this could get funded. Although to be clear, the peo­ple be­hind this 100+ res­i­dent x-risk hub idea feel that there is in­suffi­cient ev­i­dence to jus­tify ded­i­cat­ing re­sources to move ahead with the pro­ject at this time.

The Black­pool EA Ho­tel could serve as a use­ful test bed for such a larger hub (if suc­cess­ful, it could in­crease the chances of some­thing like the Geneva x-risk hub get­ting fund­ing). With a low cost hub and a high cost hub, there could be a pro­gres­sion of peo­ple lev­el­ling up at the ho­tel, to en­rol­ling in a struc­tured in­tern pro­gram at the Geneva hub, to re­search­ing and im­ple­ment­ing things at the fore­front in Geneva, or go­ing on to FHI, MIRI, CSER etc)[18].

Sus­tain­abil­ity, risks and organisation

Cur­rently this pro­ject has fund­ing for the first year of op­er­a­tions (and the ho­tel has already been bought and paid for). As­sum­ing the cryp­tocur­rency mar­kets main­tain at least a third of their cur­rent value[19] go­ing for­ward, the next few years are also cov­ered. For longer term sus­tain­abil­ity it is de­sir­able for fund­ing to be cov­ered by more than one per­son, not least due to the greater sta­bil­ity and ac­countabil­ity re­sul­tant from hav­ing mul­ti­ple fi­nan­cial stake­hold­ers. To this end, it makes sense to run the ho­tel as a non-profit char­i­ta­ble or­gani­sa­tion. Given the busi­ness-like na­ture of a ho­tel, a foun­da­tion Char­i­ta­ble In­cor­po­rated Or­gani­sa­tion (CIO) is the best op­tion (this is a cor­po­rate body where the only mem­bers are the trustees). This has an added benefit of sav­ing hav­ing to pay tax on the cap­i­tal gains used for fund­ing the pro­ject (I bought the ho­tel with my cap­i­tal gains and plan on gift­ing it to the CIO).

Best prac­tice says that CIO’s should have a min­i­mum of three Trus­tees. Any­one in­ter­ested in be­ing a Trus­tee? My ini­tial thought was that in or­der to avoid con­flict of in­ter­est, Trus­tees should not be guests stay­ing for free, or the man­ager. But per­haps that will lead the gov­er­nance of the CIO to be out of touch. Bet­ter or­gani­sa­tion might re­sult if the Man­ager, and a rep­re­sen­ta­tive of the guests, are both on the Board of Trus­tees. Per­haps then add two in­de­pen­dents for bal­ance. So the board would be five—me (as Founder), the Man­ager (as day-to-day or­ganiser), the EA Guest Rep (re­lay­ing the guest ex­pe­rience first hand, and two in­de­pen­dent Trus­tees (for an out­side per­spec­tive).

Once the or­gani­sa­tion is set up, peo­ple who think in­vest­ing in EAs at ~£5.4k/​year will pro­duce more value than the same given to top EA char­i­ties will be free to donate to it. Op­tions to spon­sor in­di­vi­d­ual stays can also be pro­vided. To give non-fee-pay­ing guests (FreEAs?) se­cu­rity and peace of mind that the rug won’t be pul­led from un­der them, there will be a con­trac­tual obli­ga­tion pro­vided by the ho­tel to provide them with ac­com­mo­da­tion and food. On the part of the non-fee-pay­ing guests there will be a coun­ter­part to this obli­ga­tion of abid­ing by the Code of Con­duct[20], and pledg­ing to work on EA things. Ac­cep­tance of long-term (over a month) non-fee-pay­ing guests, and events, will be sub­ject to at least four of five trustees agree­ing, in­clud­ing the EA Guest Rep and Man­ager. Short-term (less than a month) non-fee-pay­ing guests will be ac­cepted at the dis­cre­tion of the Man­ager (I am the de facto man­ager un­til the full-time slaried po­si­tion is filled).

In terms of over­all pro­duc­tivity, it will prob­a­bly be best to limit the ho­tel solely to free EA places. How­ever, In the event of empty rooms, it might be good to en­courage donors and other pro­mot­ers of the pro­ject to pay for rooms and stay at the ho­tel, for com­mu­nity build­ing pur­poses. In gen­eral how­ever, the em­pha­sis will be on filling as many rooms as pos­si­ble with EA guests work­ing on im­por­tant things and pay­ing noth­ing (the goal of the pro­ject is host­ing such peo­ple!). Allo­cat­ing rooms to pay­ing guests at the ex­pense of such free guests will only be done if it be­comes a ne­ces­sity for the fis­cal health of the pro­ject (say if re­serves drop be­low 12 months’ run­way). But in the ex­treme, charg­ing a 100% markup on cost price—plus and ad­di­tional £30/​week to cover stipends—for half the rooms would make the ho­tel rev­enue neu­tral, and there would be zero cost for host­ing eight free EA guests and the Ho­tel Man­ager.

Re­gard­ing the stipend, is £30/​week enough for phone, lap­top, clothes, travel, en­ter­tain­ment, and all other non-food-or-hous­ing ex­pen­di­ture? It seems so with the calcu­la­tions done in the bud­get, but per­haps there could be a sys­tem where peo­ple can re­quest ad­di­tional funds for emer­gen­cies. Also a sep­a­rate “ca­reer pro­gres­sion” fund could be use­ful for peo­ple to at­tend key con­fer­ences likely to help fur­ther their work (these funds be­ing sep­a­rate so as not to in­crease baseline costs of liv­ing at the ho­tel).

Alumni will be wel­come to give back if the ho­tel serves them well in their en­deav­ours. Per­haps a start-up launched from the ho­tel might give some equity, which in fu­ture be­comes worth a lot. This would be analo­gous to painters giv­ing paint­ings to the Chel­sea Ho­tel be­fore they be­came fa­mous, and then the paint­ings be­com­ing worth mil­lions.

What would the cost of failure of this pro­ject be? Con­ser­va­tively, per­haps sel­l­ing the ho­tel at a 15% loss (£20k), and six months of run­ning costs (<£33k (bud­get) as­sum­ing no more than half oc­cu­pancy if it’s failing). Round to £50k. Plus, say 250 hours of my time, and the time spent by the ho­tel staff and vol­un­teers. This would need coun­ter­fac­tual ad­just­ment though, which given that the pro­ject is high ex­pected value to be­gin with, could per­haps be largely de­pre­ci­ated. Fi­nan­cially, one way of look­ing at this is that it’s no more than a bad trade on the crypto mar­kets, of which I made more than one of similar mag­ni­tude in the pro­cess of mak­ing the money to fund this pro­ject (I made oth­ers that were of much larger mag­ni­tude in the op­po­site di­rec­tion though!) In fact, even sel­l­ing the ETH to fund the pur­chase of the ho­tel was in hind­sight not ideal, given that I could’ve got dou­ble the money if I’d waited three weeks! Such “what if”s are nei­ther here nor there though. £50k is still a lot of cash that could be spent on a lot of other good EA pro­jects (or more es­tab­lished giv­ing op­por­tu­ni­ties). This per­haps be­hooves more anal­y­sis for this pro­ject. Then again, it’s some­times good to just try things and avoid anal­y­sis paral­y­sis. Per­haps a happy medium could be struck though—any­one want to make a Guessti­mate model? Fund­ing is available to do so.

What is the risk of failure? Even at 50%, the down­side (~£50k + coun­ter­fac­tual time) is much more limited than the up­side, so it seems like some­thing that should be done. And The Value of In­for­ma­tion is high.

A Face­book poll shows 1-3 month stays as the most pop­u­lar op­tions. How­ever, the most benefit might per­haps be pro­vided to those that stay over a longer term, i.e. 12-24 months. Peo­ple stay­ing shorter times might well be more likely to have coun­ter­fac­tu­ally man­aged to work on the same stuff el­se­where, with­out the free liv­ing pro­vided by the ho­tel.

An­ti­so­cial be­havi­our and petty crime are is­sues in Black­pool, but hope­fully this won’t be too much of an is­sue for ho­tel guests if they stick to­gether when out at night. Statis­ti­cally, it’s no worse than Manch­ester, and I have been fine so far liv­ing there. Given the pub­lic na­ture of a ho­tel, peo­ple not con­tent with hav­ing their ad­dress read­ily dis­cov­er­able should prob­a­bly not stay at the ho­tel (at least not long term).

The Ho­tel Man­ager will need to make sure that ho­tel does not fall foul of any of the many reg­u­la­tions that ap­ply. Failing in this re­gard could have sig­nifi­cant con­se­quences, and in a worse case sce­nario lead to the ho­tel be­ing shut down, and/​or sued. The stairs (and some floors) are a bit crooked from some his­toric sub­si­dence. This will ei­ther need to be fixed (prob­a­bly overly costly), or per­haps guests can sign a dis­claimer ac­cept­ing the (min­i­mal) risk of us­ing them (the lean is about 5 de­grees, so it feels a lit­tle like be­ing on a ship ini­tially. One quickly gets used to it though).

For guests com­ing from out­side the EU (or even in­side soon, given Brexit), there will be the is­sue of visas. It should be pos­si­ble to get a Tier 5 (Tem­po­rary Worker—Char­ity Worker) visa for up to 12 months if the guest is spon­sored by the CIO. The CIO will first need to ap­ply to be­come a spon­sor.

What about the name and brand­ing of the ho­tel? Some­thing straight for­ward like “The EA Ho­tel”, or “The EA Ho­tel: Black­pool”? Or maybe “The Ben­tham Ho­tel”? Or some­thing with more in-group ap­peal? “The Phoenix’ Nest” (H/​T Ryan Christo­pher Au­gus­tine Thomas) has as­so­ci­a­tions with in­cu­ba­tion, HPMoR, al­tru­ism, im­mor­tal­ity, and also the en­glish pub aes­thetic, and the trope of ad­ven­tur­ers meet­ing at an inn. How­ever, ex­plic­itly as­so­ci­at­ing it with EA might not be ideal when fac­tor­ing rep­u­ta­tional con­cerns; given that the ini­tial fund­ing source comes from risky crypto in­vest­ments, and the pos­si­bil­ity of a failure to de­liver in terms of im­pact. A re­cent poll on the Face­book group for this pro­ject now has “The EA Ho­tel: Black­pool” in sec­ond place, with “Athena” Ho­tel (what it’s cur­rently called) in the lead. This sug­gests a com­mu­nity prefer­ence for straight for­ward nam­ing, and some cau­tion around at­tach­ing the moniker “EA” to pro­jects.

The ho­tel, with view of Black­pool Tower around the cor­ner.

Per­haps the name could be kept as it is, but the logo could have the “E” and the sec­ond “A” of “Athena” a differ­ent colour (EA turquoise?). Some­thing like this:
or this:

There is a dan­ger if the com­mu­nity is too in­su­lar and in­ward look­ing that it will de­velop un­healthy “cult-like” ten­den­cies. This should be miti­gated by hav­ing a reg­u­lar in­flux of tem­po­rary vis­i­tors from within EA (and oc­ca­sion­ally from out­side EA).

As with any pro­ject that in­volves mov­ing to a new city, get­ting peo­ple to com­mit to move, and then ac­tu­ally fol­low through and move, is the hard part. There have been over 30 peo­ple ex­press­ing in­ter­est so far, but ideally there needs to be an­other 3-5x this amount to en­sure the ho­tel is well pop­u­lated. For some, the top con­sid­er­a­tion is who they will be liv­ing with; for them, it will be a case of see­ing how the ho­tel is pop­u­lated be­fore de­cid­ing on mov­ing.

In or­der to miti­gate the risk of hiring an un­suit­able man­ager and then be­ing stuck with them, the job will in­volve a six month trial pe­riod to be­gin with. The job will also not con­tinue af­ter this amount of time if the ho­tel proves to be un­vi­able in terms of not at­tract­ing enough guests. If there aren’t many guests, the man­ager can use their work­ing time to fo­cus on de­vel­op­ing the wider idea of “EA Ho­tels and Low-cost Liv­ing” (which may lead to a con­tinu­a­tion of the job in other cir­cum­stances).

For the ho­tel to be suc­cess­ful, the at­mo­sphere needs to be invit­ing. To a large part this will be de­ter­mined by the guests and how well they get on with each other. Also im­por­tant is the qual­ity of the pro­jects that guests are in­volved with. Po­ten­tial new guests will be more likely to join the higher the pal­pable sense of ex­cite­ment is re­gard­ing the peo­ple they will be liv­ing with, and they pro­jects they are work­ing on. For ex­am­ple, a strong group of peo­ple study­ing/​re­search­ing AI Safety should at­tract oth­ers with high po­ten­tial for AI Safety work. To some ex­tent this is a bit of a chicken and egg situ­a­tion, but this can be over­come if a strong ini­tial group is com­mit­ted to mov­ing at the start. Per­haps a pledge where peo­ple com­mit in groups to move could help. Also host­ing work­shops whereby par­ti­ci­pants would be wel­come to stay on at the ho­tel af­ter­ward for free.

Men­tal health is an im­por­tant is­sue to ad­dress. Per­haps there is an in­creased risk that guests will suffer from men­tal ill health, cor­re­lated with the cir­cum­stances they find them­selves in (i.e. be­ing in need of fund­ing and/​or com­mu­nity), or per­haps the wider com­mu­nity suffers dis­pro­por­tionately in tread­ing the “fine line be­tween ge­nius and in­san­ity”. Fos­ter­ing a strong sense of com­mu­nity, ca­ma­raderie and mu­tual sup­port should help. And as pre­vi­ously men­tioned, a case could be made for hiring a full time coun­sel­lor.

For fund­ing peo­ple’s liv­ing costs el­se­where to the ho­tel (e.g. South-East Asia), there would be the is­sue of ver­ify­ing whether they are ac­tu­ally spend­ing that amount/​year, and aren’t spend­ing more, so do­ing less work with the grant—i.e. if they are get­ting fund­ing at £5k/​yr to live but are spend­ing £10k, then the £5k is effec­tively only fund­ing half as much time. To some ex­tent this will have to rely on trust. Other­wise, per­haps bank state­ments and re­ceipts would need to be col­lected, but this would both add ad­min costs and be a po­ten­tial vi­o­la­tion of pri­vacy.

Next Steps

UPDATE 09 Au­gust 2018: The web­site is now up at ea­ho­tel.org. It’s pos­si­ble to book via the Book­ing page, and the Wiki page has in­for­ma­tion for guests and an FAQ. Peo­ple can also book video calls to talk about pro­pos­als for free stays work­ing on EA stuff via the Con­tact page.

Gather­ing feed­back from the EA com­mu­nity—if any­one has strong rea­sons to be­lieve that this is not a good idea (or that other uses of the money are much more likely to be higher ex­pected value), please let me know in the com­ments be­low (or oth­er­wise).

Cru­cial con­sid­er­a­tions aside, whether or not this pro­ject is suc­cess­ful largely de­pends on there be­ing suffi­cient de­mand for it, and find­ing some­one ca­pa­ble and will­ing to take on the role of Ho­tel Man­ager. There has been some ini­tial de­mand already (via dis­cus­sion of the idea on Face­book), but per­haps the bulk of de­mand will be take a “build it and they will come” form.

To help gauge in­ter­est in this pro­ject, please state dates that you would like to stay/​host an event at the ho­tel, so they can be added to the Google Cal­en­dar. Com­ment be­low (or PM me) with the dates (and % like­li­hood of them), or your email so you can be given edit ac­cess to the cal­en­dar (cre­ate an event with your name or event name. e.g. I’ve added an event called “Greg [100%]” from 14 May 2018 to 14 May 2020 − 14th May is the date I gained pos­ses­sion of the ho­tel).

Get­ting an ini­tial crit­i­cal mass of guests is im­por­tant. The ho­tel is already open and stocked up[21]. To reg­ister your gen­eral in­ter­est, please fill out this short form. To pre-reg­ister for a res­i­dency at the ho­tel, please fill out this longer form. Or PM me to ar­range a video call to have a look around and dis­cuss your work plans for your stay. If we are already friends, feel free to just drop in and have a look around/​stay for a few days :)

Re­gard­ing vol­un­teer­ing at the ho­tel, helping it get set up etc, I have had a de­cent amount of in­ter­est in this already and prob­a­bly have enough help for now, thanks for all the sup­port! What’s re­ally needed now is peo­ple to fill the rooms do­ing ob­ject level EA work (see here for in­spira­tion), and peo­ple to fill the for­mal roles of Ho­tel Man­ager and In­de­pen­dent Trus­tees (EAs who aren’t likely to di­rectly benefit from the ho­tel, but are in­ter­ested in helping the pro­ject suc­ceed. The board of trustees of a char­ity fulfils a similar role to the board of di­rec­tors at a com­pany. In this case it would be at­tend­ing fort­nightly meet­ings over video chat dis­cussing the di­rec­tion of the pro­ject and the ac­cep­tance of long term guests).

Peo­ple in­ter­ested in the Ho­tel Man­ager role can ap­ply here.

Peo­ple in­ter­ested in be­com­ing an In­de­pen­dent Trus­tee can ap­ply here.

Cur­rently we have a team of re­mote vol­un­teers as­sem­bled who are stand in trustees un­til a for­mal re­cruit­ment pro­cess is com­pleted. We are pro­ceed­ing to form a Char­i­ta­ble In­cor­po­rated Or­gani­sa­tion (CIO) to run the pro­ject, with the goal of mov­ing to a five-mem­ber team of trustees made up of the Ho­tel Man­ager, the EA Guest Rep­re­sen­ta­tive, me (as Founder) and two for­mally re­cruited ex­ter­nal trustees, in due course. It would also be use­ful to have peo­ple with knowl­edge/​ex­pe­rience of UK im­mi­gra­tion law, as many peo­ple from both in­side and out­side the EU are in­ter­ested in stay­ing.

Those in­ter­ested in fol­low­ing the pro­ject can join the Face­book group (feel free to ig­nore the sign-up ques­tions if you aren’t con­sid­er­ing stay­ing at the ho­tel and we have mu­tual Face­book friends). Easier than giv­ing spe­cific dates for a stay you can fill in this poll in the group.

Look­ing for­ward to wel­com­ing many of you to the ho­tel :)


For use­ful dis­cus­sion, pro­mo­tion, vol­un­teer­ing and sup­port, both re­mote and in per­son: John Maxwell, Jes­sica Cooper, Rem­melt Ellen, David Krueger, Maia Pasek, John Beshir, Ryan Carey, Linda Linse­fors, Kel­sey Piper, Toon Alfrink, Florent Ber­thet, Chado Nihi, Abra­ham Rowe, Mati Roy, Alex Dick­in­son, Tor Øk­land Barstad, Dony Christie, Ryan Christo­pher Au­gus­tine Thomas, Rochelle Har­ris, Kyle Hub­bard, Simmo Simp­son, Ge­nie Yogini, Floris Wolswijk, David Kristoffers­son, Liam Dono­van, Char­lie Rogers-Smith, Tom Rogers, Iso­bel Colbourn, Chris Colbourn.


[1] Rooms will need to be tidy ready for the cleaner.

[2] More pic­tures of the ho­tel here.

[3] Guests who want to eat non-ve­gan food are wel­come to buy their own. There will be a combi-oven (microwave, steamer, grill, oven) and toaster available for guest use. Fry­ing and pot-based cook­ing by guests will not be fea­si­ble though. And ob­vi­ously there are take­aways, cafes and restau­rants nearby (in­clud­ing a hell of a lot of Fish & Chips).

[4] By ask­ing peo­ple to write ini­ti­a­tives on white­boards, for ex­am­ple.

[5] More pic­tures of the area sur­round­ing the ho­tel, and Black­pool beach, here and here.

[6] This is within the range of mar­ket rates.

[7] Ap­ply here.

[7a] Two In­de­pen­dent Trus­tees are sought. Ap­ply here.

[8] Per­haps we could re­order them and have room 13 as “the LessWrong suite” or some­thing :)

[9] See here for dis­cus­sion about the or­gani­sa­tional struc­ture of the pro­ject.

[10] This may well be in non-mon­e­tary form if lit­tle in the way of sta­ble cur­rency re­mains.

[11] Vari­able, de­pen­dent on oc­cu­pancy of the ho­tel, see bud­get.

[12] Still, given the cur­rent skewed cash to tal­ent ra­tio in EA, per­haps it would be a good idea for there to be a pro­gram (per­haps run by OpenPhil or CEA) where some­one re­searches the EA sphere in de­tail to iden­tify bot­tle­necks in ca­pac­ity, and com­pe­tent peo­ple who are com­mit­ted in those ar­eas as vol­un­teers, and then just gives them money to do the EA stuff full time. This would com­ple­ment the EA Grants pro­gram, and would be valuable to peo­ple who liter­ally don’t have the time or en­ergy to for­mally ap­ply for fund­ing. I think it would have a good chance of be­ing highly cost effec­tive. Ac­tu­ally, it does seem that CEA might be mov­ing to­ward this model (“we plan to move the eval­u­a­tion pro­cesses even fur­ther in the di­rec­tion of mostly eval­u­at­ing the mer­its of the ap­pli­cants them­selves rather than their spe­cific plans”) but not with­out con­tro­versy.

[13] Th­ese ex­am­ples also illus­trate the fact that even gen­er­ally effec­tive peo­ple have pro­jects that fail.

[14] To a rough ap­prox­i­ma­tion OpenPhil has ~$10bn to spend (adding the rest of the net worth of the EA com­mu­nity might in­crease this by a fac­tor of ~2 maybe); There were ~700 ap­pli­ca­tions for the last round of EA grants. In­clud­ing ap­pli­cants for other grants (such as those offered by FLI) round to 1000, say with an av­er­age ask of ~$100k (prob­a­bly on the high side). That’s still <1% of the cash available.

[15] Of course in­di­vi­d­u­als can do this to an ex­tent in terms of batch­ing cook­ing and other tasks and buy­ing in bulk, but it re­quires high up-front costs in terms of plan­ning and money.

[16] As­sum­ing 16 oc­cu­pants plus the Man­ager.

[17] Mov­ing of­ten de­pends a lot on timing—in terms of wait­ing on other op­por­tu­ni­ties, finish­ing pro­jects/​courses/​con­tracts, co­or­di­nat­ing with friends/​part­ners/​fam­ily etc.

[18] Were such a grand hub to also have a ho­tel at­tached (en­abling even higher effi­ciency liv­ing, see above), per­haps it could end up as a kind of cross be­tween Google HQ and the Wal­dorf As­to­ria, but in­stead of tech, celebrity and poli­tics, the guests would be from the wor­lds of x-risk re­search and miti­ga­tion, and global diplo­macy and co­or­di­na­tion; and by rights it should be­come as im­por­tant as those two hubs were in the mid-20th cen­tury and early 21st, go­ing for­ward into the 2020s and ’30s.

[19] As of 2018-06-04T17:14:09+01:00.

[20] A ver­sion of this would also be in the Terms and Con­di­tions fee-pay­ing guests agree to abide by.

[21] I’ve been busy over the last cou­ple of weeks or­der­ing sup­plies and use­ful items, and in­stal­ling en­ergy effi­cient white goods and LED light­ing (high lu­men daylight over­head light­ing for day time, and lower pow­ered warm light­ing on walls /​ in lamps for evenings).

[22] Which could be pos­si­ble with Brexit lead­ing to more UK peo­ple holi­day­ing at home, or maybe even a grow­ing lo­cal frack­ing in­dus­try!?

Ap­pendix: Budget

The spread­sheet here de­tails the bud­get for the ho­tel. In­cluded are sheets show­ing fixed costs and vari­able costs. Vari­able costs de­pend on the oc­cu­pancy of the ho­tel—the num­ber of guests can be ed­ited at the top left. Note that all the figures that are in­puts to the calcu­la­tions are high­lighted in yel­low; to play around with the num­bers, please re­quest ac­cess and make a copy of the sheet. Figures quoted in the main text as­sume an oc­cu­pancy of 16 peo­ple (one per­son per bed­room, ex­clud­ing the Ho­tel Man­ager). £5700/​per­son/​year is ar­rived at with all rooms filled by a sin­gle per­son. If peo­ple were to share, filling the ho­tel to its ca­pac­ity of 35 peo­ple would re­sult in costs of ~£4000/​per­son/​year. Th­ese figures are with­out fac­tor­ing in amor­ti­sa­tion of the cost of the build­ing; this is rea­son­able as the build­ing will re­main an as­set which should re­tain value (or even in­crease in value if Black­pool’s for­tunes turn around[22])

Ex­am­ples of costed meals and stipend spend­ing are given to show that the al­lowances of £2.75/​per­son/​day for food and £30/​per­son/​week for stipend are rea­son­able. Utility bills are es­ti­mated us­age.

Room charges for pay­ing guests are shown on an­other sheet. Shown are cost prices per day, week, month and year (in­clud­ing a £10 ad­min fee per stay), and markups of 25%, 50%, 100% and 100% plus stipend, which would cover the costs of a non-fee-pay­ing guest. Dou­ble and triple room prices are calcu­lated based on the frac­tion of their share of sav­ings from economies of scale (see calcu­la­tions on “Rooms” sheet). In­deed, for pay­ing guests, prices could be ad­justed de­pend­ing on the oc­cu­pancy level of the ho­tel (lower prices the ful­ler the ho­tel is).

Also in­cluded are time es­ti­mates for ad­min, clean­ing and other work, and sketched out timeta­bles for work­ers and guests; a table show­ing de­tails of the rooms; and a list of pur­chases made to date.