How can Cognito Mentoring do the most good?

In late De­cem­ber 2013, I an­nounced that Vipul Naik and I had launched Cog­nito Men­tor­ing, an ad­vis­ing ser­vice for in­tel­lec­tu­ally cu­ri­ous young peo­ple.

Vipul Naik and I are as­piring effec­tive al­tru­ists, and we started Cog­nito with a view to­ward do­ing the most good. We’ve learned a lot over the past 3 months, and are work­ing on plan­ning what to do next. We’d be very grate­ful for any feed­back on cur­rent think­ing, which I’ve de­scribed be­low.

Our Mission

Hu­man cap­i­tal is one of so­ciety’s most valuable re­sources, and school years (ages 5 through 22) are a cru­cial time pe­riod for build­ing hu­man cap­i­tal. Ed­u­ca­tion is a ~$1 trillion dol­lar sec­tor, but schools are of­ten dys­func­tional in­sti­tu­tions, and very lit­tle effort goes into helping young peo­ple de­velop as much as pos­si­ble and to al­lo­cate their hu­man cap­i­tal as well as pos­si­ble. We want to help op­ti­mize young peo­ple’s life tra­jec­to­ries. For the time be­ing, we’ve cho­sen to fo­cus on helping highly in­tel­lec­tu­ally ca­pa­ble young peo­ple. Some rea­sons for this are:

  • In­tel­lec­tu­ally ca­pa­ble peo­ple con­tribute dis­pro­por­tionate so­cial value (e.g. Bill Gates solved an un­solved math­e­mat­ics re­search prob­lem as a sopho­more in col­lege; the Google co-founders were com­puter sci­ence grad­u­ate stu­dents at Stan­ford), and helping them de­velop is cor­re­spond­ingly more lev­er­aged.

  • We have deep knowl­edge of the pop­u­la­tion.

  • The ed­u­ca­tional in­fras­truc­ture is de­signed for the av­er­age stu­dent, and the gap be­tween how things are and what would be op­ti­mal is great­est for the out­liers.

  • By fo­cus­ing on a sub­pop­u­la­tion, we can offer more tar­geted recom­men­da­tions.

Some ways in which we aim to help them im­prove their life tra­jec­to­ries are:

  • En­courag­ing re­flec­tive de­ci­sion mak­ing and meta-cog­ni­tion: we get them think­ing about what they want out of life, and how best to at­tain it. In this re­spect, we over­lap with CFAR.

  • High­light­ing the ad­van­tages of learn­ing differ­ent sub­jects to help them de­cide which ones are most im­por­tant to learn.

  • Point­ing them to the best learn­ing re­sources available.

  • Helping them find high value ex­tracur­ricu­lar ac­tivi­ties to en­gage in.

  • In­form­ing them of the ad­van­tages and dis­ad­van­tages of differ­ent ca­reer choice. In this re­spect, we over­lap with 80,000 Hours (while differ­ing in that our fo­cus is on peo­ple who won’t be en­ter­ing the job mar­ket for sev­eral years).

  • Con­nect­ing them with peo­ple who have sub­ject mat­ter knowl­edge in their aca­demic or pro­fes­sional ar­eas of in­ter­est.


We’re pri­mar­ily tar­get­ing high school and col­lege stu­dents within the range of in­tel­lec­tual abil­ity of Less Wrongers.

About 75% of re­spon­dents to the 2013 Less Wrong Sur­vey who re­ported SAT scores out of 2400 gave a score of 2130+: this is at the 98th per­centile of SAT tak­ers. There are ~40,000 peo­ple per grade in that score range in the United States na­tion­wide, so ~320,000 Amer­i­cans. When one ac­counts for peo­ple at lower per­centiles who would benefit, as well as stu­dents from other coun­tries, the rele­vant pop­u­la­tion is ~1 mil­lion.

We’re also well equipped to serve peo­ple of younger ages who are highly gifted, and are at a de­vel­op­men­tal stage where they’re ca­pa­ble of en­gag­ing in metacog­ni­tion and learn­ing high school and col­lege level ma­te­rial. There are per­haps ~200,000 such peo­ple wor­ld­wide.

Our activities

At the time when we posted in De­cem­ber 2013, we were think­ing of fo­cus­ing on per­son­al­ized ad­vis­ing, per­haps with a view to­ward be­com­ing a fran­chise. Since then, we’ve shifted in the di­rec­tion of fo­cus­ing on pro­duc­ing writ­ten con­tent. There are two rea­sons for this:

  • Our ad­visees have de­rived most of the value from our generic writ­ten con­tent.

  • While some of our ad­visees have benefited very sub­stan­tially, the av­er­age benefits per per­son don’t seem to be out­sized.

Based on the first point and the size of the tar­get pop­u­la­tion, if we can pro­duce high qual­ity writ­ten con­tent and dis­sem­i­nate it widely, in prin­ci­ple 100k+ peo­ple could get a large frac­tion of the benefit of per­son­al­ized ad­vis­ing for free.

So far ~70 peo­ple have con­tacted us, in­clud­ing ~40 from Less Wrong (c.f. What we learned about Less Wrong from Cog­nito Men­tor­ing ad­vis­ing). We cor­re­sponded at length with a sub­stan­tial frac­tion of them. We’ve taken the ad­vice that we’ve gen­er­ated and con­verted it into dozens of ar­ti­cles on our ad­vice wiki, at our Quora blog, on Less Wrong and at the David­son In­sti­tute Gifted Is­sues Dis­cus­sion Fo­rum. (We’ll be con­soli­dat­ing ev­ery­thing into the wiki even­tu­ally: the rea­son that we’re post­ing to mul­ti­ple fo­rums is for out­reach pur­poses and to get feed­back.)


Our front page has been get­ting ~400 page views per week, and our wiki has been get­ting ~400 page views a week. Our Quora blog has 27 fol­low­ers. We would like our visi­bil­ity to in­crease by 1000x.

We’ve strug­gled to find av­enues by which to dis­sem­i­nate our ad­vice. There seem to be few fo­rums where smart high school stu­dents con­gre­gate. Those fo­rums and mailing lists that do ex­ist of­ten have strict guidelines against posters pro­mot­ing their own blogs. We’re grate­ful that Less Wrong has been wel­com­ing.

We’d ap­pre­ci­ate any sug­ges­tions for how we might be able to reach more peo­ple.

Where will the so­cial value come from?

The main av­enues through which peo­ple gen­er­ate so­cial value and dis­value are

  1. Career

  2. Side pro­jects /​ volunteering

  3. Donat­ing to charity

  4. En­joy­ing recre­ational ac­tivity, health and wealth

  5. Re­la­tions with fam­ily and friends

  6. Hav­ing children

We have to offer our ad­visees ad­vice that im­proves their lives for them to find it worth­while, but we think that our so­cial im­pact will be me­di­ated pri­mar­ily through the im­pacts of #1 and #2 on oth­ers.

It may be sur­pris­ing that we high­light #2. One rea­son that we high­light it is that high school and col­lege stu­dents tend to have free time out­side of school, that they can spend more pro­duc­tively on side pro­jects than on the rel­a­tively low-skil­led part time jobs that are available to them with­out the cre­den­tial of a col­lege de­gree. Another is that it can be hard to find fund­ing to work on some­thing of high so­cial value full time. Some ex­am­ples of suc­cess­ful side pro­jects cre­ated by mem­bers of the effec­tive al­tru­ist /​ Less Wrong com­mu­ni­ties are:

Why don’t we ex­pect our im­pact to be through #3 (donat­ing to char­ity)?

  • A lot of the peo­ple well-suited to mak­ing money already do it by de­fault: while there are in­di­vi­d­u­als who would do more good tak­ing a higher pay­ing job and earn­ing to give, we wouldn’t ex­pect to be able to boost peo­ple’s salaries a lot on av­er­age, given the con­straints that they op­er­ate un­der, both with re­spect to skills and with re­spect to the sorts of work they’d be will­ing to do.

  • Our ad­visees won’t be mak­ing a lot of money for a long time — by the time they do, they may have had a lot of ex­po­sure to the ideas of effec­tive al­tru­ism through other chan­nels (whether through ex­ist­ing or­ga­ni­za­tions such as GiveWell and 80,000 Hours, or through fu­ture or­ga­ni­za­tions).

  • For effec­tive al­tru­ist types, 80,000 Hours Ex­ec­u­tive Direc­tor Ben­jamin Todd has said that he doesn’t think that it’s plau­si­ble that earn­ing to give is likely to be the path to­ward do­ing the most good. I gave more points against earn­ing to give as op­ti­mal effec­tive al­tru­ism in Earn­ing to Give vs. Altru­is­tic Ca­reer Choice Re­vis­ited.

  • We’re in a differ­ent cul­tural sphere from peo­ple in fi­nance and busi­ness /​ con­sult­ing, and bet­ter suited to help peo­ple who are en­gaged in more in­tel­lec­tual en­deav­ors.

Con­cern­ing #4, the benefits would not be lev­er­aged; con­cern­ing #5, one would ex­pect the benefits would be ~1x the benefits to the in­di­vi­d­ual, which isn’t a large mul­ti­plier; con­cern­ing #6, we wouldn’t ex­pect to have much im­pact on peo­ple’s de­ci­sion to have chil­dren, the sign of the effect would be am­bigu­ous, and our ad­visees are far from the point of ac­tu­ally rais­ing chil­dren.

Here are some ex­am­ples of chan­nels through which we ex­pect to have a pos­i­tive im­pact on #1 and #2:

  • There’s a wide­spread mis­con­cep­tion amongst high school stu­dents that they have to en­gage in par­tic­u­lar ex­tracur­ricu­lar ac­tivi­ties (or many ex­tracur­ricu­lar ac­tivi­ties) to get into good col­leges. By rais­ing aware­ness that this is not the case, we can free stu­dents up to en­gage in sub­stan­tive side pro­jects such as con­tribut­ing to open source soft­ware pro­jects and writ­ing Wikipe­dia ar­ti­cles on im­por­tant top­ics — things that both have di­rect so­cial benefit and that build skills that are use­ful for fu­ture ac­tivi­ties.

  • We’re dis­sem­i­nat­ing in­for­ma­tion about the benefits of com­puter pro­gram­ming, point­ing peo­ple to pro­gram­ming learn­ing re­sources, and point­ing peo­ple to in­for­ma­tion about how to learn pro­gram­ming. By reach­ing high school stu­dents, we can help peo­ple get a head start, prepar­ing them for the op­tion of be­com­ing soft­ware en­g­ineers, which will (in ex­pec­ta­tion) move peo­ple into the tech sec­tor, which has un­usu­ally great pos­i­tive ex­ter­nal­ities.

  • A mod­er­ately large frac­tion of in­tel­lec­tu­ally ca­pa­ble peo­ple go to grad­u­ate school and end up not us­ing their de­grees (e.g. be­cause they’re un­able to get jobs in the aca­demic mar­ket), or end up do­ing re­search of lit­tle prac­ti­cal rele­vance. By dis­sem­i­nat­ing in­for­ma­tion on academia as a ca­reer op­tion and pro­mot­ing an un­bi­ased view of the value of the­o­ret­i­cal re­search, we can di­vert peo­ple into ca­reers where they can make a differ­ence.

  • By ed­u­cat­ing peo­ple about the un­con­ven­tional path of en­trepreneur­ship as a ca­reer op­tion (for ex­am­ple, by point­ing them to en­trepreneur­ship learn­ing re­sources and con­nect­ing them with en­trepreneurs who we know) we can en­able more peo­ple to in­no­vate more than they oth­er­wise would.

We’re very in­ter­ested in fur­ther ideas along these lines, as well as sug­ges­tions for how we can re­al­ize them.


We origi­nally thought in terms of sup­port­ing the op­er­a­tion by charg­ing for per­son­al­ized ad­vis­ing. This could still be an op­tion, but:

  • High school and col­lege stu­dents gen­er­ally don’t have much money.

  • Most of the stu­dents who we ad­vised said that know­ing what they know now, they would have sought ad­vis­ing from us only if it were free. This is true even of those who re­ported to benefit­ing sub­stan­tially, sug­gest­ing that we can’t re­solve the is­sue by im­prov­ing the qual­ity of our ad­vice.

  • Stu­dents only need ~5 hours of ad­vis­ing from us at a given time, so even to the ex­tent that peo­ple are will­ing to pay, there’s sub­stan­tial over­head in­volved per paid hour.

  • While per­son­al­ized ad­vis­ing does feed into our pub­lic con­tent at the cur­rent mar­gin, if we had to fo­cus on it heav­ily, it would dis­tract from pro­duc­ing the more valuable pub­lic con­tent.

At this point, we’re seek­ing philan­thropic fund­ing, and would ap­pre­ci­ate any ideas as to how to se­cure it.