-
Notifications
You must be signed in to change notification settings - Fork 0
/
considerations_on_cryonics.html
1667 lines (1636 loc) · 112 KB
/
considerations_on_cryonics.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
<html><head><title>niplav</title>
<link href="./favicon.png" rel="shortcut icon" type="image/png"/>
<link href="main.css" rel="stylesheet" type="text/css"/>
<meta content="text/html; charset=utf-8" http-equiv="Content-Type"/>
<!DOCTYPE HTML>
<style type="text/css">
code.has-jax {font: inherit; font-size: 100%; background: inherit; border: inherit;}
</style>
<script async="" src="./mathjax/latest.js?config=TeX-MML-AM_CHTML" type="text/javascript">
</script>
<script type="text/x-mathjax-config">
MathJax.Hub.Config({
extensions: ["tex2jax.js"],
jax: ["input/TeX", "output/HTML-CSS"],
tex2jax: {
inlineMath: [ ['$','$'], ["\\(","\\)"] ],
displayMath: [ ['$$','$$'], ["\\[","\\]"] ],
processEscapes: true,
skipTags: ['script', 'noscript', 'style', 'textarea', 'pre']
},
"HTML-CSS": { availableFonts: ["TeX"] }
});
</script>
<script>
document.addEventListener('DOMContentLoaded', function () {
// Change the title to the h1 header
var title = document.querySelector('h1')
if(title) {
var title_elem = document.querySelector('title')
title_elem.textContent=title.textContent + " – niplav"
}
});
</script>
</head><body><h2 id="home"><a href="./index.html">home</a></h2>
<p><em>author: niplav, created: 2019-10-18, modified: 2024-09-21, language: english, status: maintenance, importance: 7, confidence: possible</em></p>
<blockquote>
<p><strong>Is <a href="https://en.wikipedia.org/wiki/Cryonics">cryonics</a>
worth it, and if yes, should one
<a href="https://alcor.org/Library/html/cryocrastination.html">cryocrastinate</a>
(i.e. postpone signing up for cryonics to a later date)? <a href="https://en.wikipedia.org/wiki/Betteridge's_law_of_headlines">Betteridge's law
of headlines</a>
only applies partially here: Yes, it is probably worth it (under plausible
assumptions <a href="#Currently_20_years_old">\$850k for a 20 year old</a>,
and <a href="#Currently_40_years_old">more for older people</a>), and no,
cryocrastination is usually irrational (unless one <a href="#Without_Motivation_Drift">doesn't experience
value drift</a>, in which case cryocrastination
is rational until age 55). A cost-benefit analysis written in
<a href="https://en.wikipedia.org/wiki/Lua_(programming_language)">Lua</a>,
considering
<a href="https://www.alcor.org/library/neuropreservation-option/">neurocryopreservation</a>
with <a href="https://www.alcor.org/">Alcor</a>. I also perform a <a href="https://en.wikipedia.org/wiki/Monte_Carlo_method">Monte-Carlo
simulation</a> using
<a href="https://www.getguesstimate.com/" title="Official website">Guesstimate</a>,
and find that signing up for cryonics at age 20 is worth \$18m
in the mean, median -\$100k (90% confidence interval: -\$2.16m,
\$58.25m). It therefore seems recommendable to sign up for cryonics
<a href="#Now_What">immediately</a>.</strong></p>
</blockquote><div class="toc"><div class="toc-title">Contents</div><ul><li><a href="#Other_Cryonics_Texts">Other Cryonics Texts</a><ul></ul></li><li><a href="#Caveats">Caveats</a><ul></ul></li><li><a href="#CostBenefit_Calculation_for_Cryonics">Cost-Benefit Calculation for Cryonics</a><ul></ul></li><li><a href="#The_Disvalue_of_Waiting">The Disvalue of Waiting</a><ul><li><a href="#Motivation_Drift">Motivation Drift</a><ul></ul></li><li><a href="#Dying_Before_Signing_Up">Dying Before Signing Up</a><ul></ul></li><li><a href="#Longevity_Escape_Velocity">Longevity Escape Velocity</a><ul></ul></li></ul></li><li><a href="#Calculating_the_Cost">Calculating the Cost</a><ul><li><a href="#Membership_Fees">Membership Fees</a><ul><li><a href="#Direct_Fees">Direct Fees</a><ul></ul></li><li><a href="#Comprehensive_Member_Standby">Comprehensive Member Standby</a><ul></ul></li></ul></li><li><a href="#Preservation_Cost">Preservation Cost</a><ul></ul></li><li><a href="#Other_Possible_Costs">Other Possible Costs</a><ul></ul></li></ul></li><li><a href="#Calculating_the_Benefit">Calculating the Benefit</a><ul><li><a href="#Value_of_a_Lifeyear_in_the_Future">Value of a Lifeyear in the Future</a><ul><li><a href="#Caveats_on_Future_Life_Years">Caveats on Future Life Years</a><ul></ul></li><li><a href="#QALYs_and_VSL">QALYs and VSL</a><ul></ul></li><li><a href="#Negative_Scenarios">Negative Scenarios</a><ul><li><a href="#Ascended_Economy">Ascended Economy</a><ul></ul></li><li><a href="#Malevolent_Future_Actors">Malevolent Future Actors</a><ul></ul></li><li><a href="#Information_from_the_Past_is_Valuable">Information from the Past is Valuable</a><ul></ul></li><li><a href="#Being_Revived_as_an_Emulation_and_Forced_to_Work">Being Revived as an Emulation and Forced to Work</a><ul></ul></li></ul></li><li><a href="#Steps_for_Reducing_the_Risk_from_such_Scenarios">Steps for Reducing the Risk from such Scenarios</a><ul></ul></li><li><a href="#Other_Thoughts">Other Thoughts</a><ul></ul></li></ul></li><li><a href="#Probability_of_Revival">Probability of Revival</a><ul><li><a href="#Other_Relevant_Questions">Other Relevant Questions</a><ul></ul></li></ul></li><li><a href="#Years_Gained">Years Gained</a><ul><li><a href="#Existential_Risk">Existential Risk</a><ul></ul></li></ul></li><li><a href="#Probability_of_Being_Preserved">Probability of Being Preserved</a><ul></ul></li><li><a href="#Quality_of_Preservation">Quality of Preservation</a><ul></ul></li><li><a href="#Surviving_Until_LEV">Surviving Until LEV</a><ul></ul></li><li><a href="#Extinction_Risk_After_Revival">Extinction Risk After Revival</a><ul></ul></li></ul></li><li><a href="#Putting_it_All_Together">Putting it All Together</a><ul></ul></li><li><a href="#Results">Results</a><ul><li><a href="#Standard_Parameters">Standard Parameters</a><ul><li><a href="#Currently_20_years_old">Currently 20 years old</a><ul></ul></li><li><a href="#Currently_40_years_old">Currently 40 years old</a><ul></ul></li></ul></li><li><a href="#Without_Motivation_Drift">Without Motivation Drift</a><ul></ul></li><li><a href="#The_Critics_Scenario">The Critic's Scenario</a><ul></ul></li><li><a href="#Other_Modifications">Other Modifications</a><ul></ul></li></ul></li><li><a href="#Appendix_A_A_Guesstimate_Model">Appendix A: A Guesstimate Model</a><ul><li><a href="#Variables">Variables</a><ul><li><a href="#Year_for_Longevity_Escape_Velocity">Year for Longevity Escape Velocity</a><ul></ul></li><li><a href="#Age_at_Death">Age at Death</a><ul></ul></li><li><a href="#Years_Lived_After_Revival">Years Lived After Revival</a><ul></ul></li><li><a href="#Value_of_Lifeyears_After_Revival">Value of Lifeyears After Revival</a><ul></ul></li><li><a href="#Provider_Cost_per_Year">Provider Cost per Year</a><ul></ul></li></ul></li><li><a href="#Value">Value</a><ul></ul></li><li><a href="#Conclusion">Conclusion</a><ul></ul></li></ul></li><li><a href="#Now_What">Now What?</a><ul></ul></li><li><a href="#See_Also">See Also</a><ul></ul></li></ul></div>
<h1 id="Cryonics_CostBenefit_Analysis"><a class="hanchor" href="#Cryonics_CostBenefit_Analysis">Cryonics Cost-Benefit Analysis</a></h1>
<blockquote>
<p>If I died, would I be rid of my senses?<br/>
Or will it retain, trapped within my corpse, in stasis?<br/>
If I died, would I be a woman in heaven?<br/>
Or would I fall asleep, not knowing what it’s like to feel alive?</p>
<p>If I died, would I begin with a new life?<br/>
Or would I be gone as quickly as the breath I give last?<br/>
If I died, would I be a woman in heaven?<br/>
Or would I fall asleep, not knowing what it’s like to feel alive?</p>
</blockquote>
<p><em>— <a href="https://www.youtube.com/channel/UCOjLfFE0TABS9bk05oH-kgw">Patricia Taxxon</a>, <a href="https://patriciataxxon.bandcamp.com/track/deconstruct">“Deconstruct”</a> from <a href="https://patriciataxxon.bandcamp.com/album/foley-artist">“Foley Artist”</a>, 2019</em></p>
<blockquote>
<p>One day they woke me up<br/>
So I could live forever<br/>
It’s such a shame the same<br/>
Will never happen to you</p>
</blockquote>
<p><em>— <a href="https://en.wikipedia.org/wiki/Jonathan_Coulton">Jonathan Coulton</a>, <a href="https://genius.com/Aperture-science-psychoacoustic-laboratories-want-you-gone-lyrics">“Want You Gone”</a> from <a href="https://en.wikipedia.org/wiki/Music_of_Portal_2">“Portal 2: Songs to Test By (Volume 3)”</a>, 2011</em></p>
<p>Many would-be cryonicists cryocrastinate, i.e. they put off signing
up for cryonics until a later point in their life. This has often
been explained by the fact that signing up for cryonics <a href="https://www.lesswrong.com/posts/hiDkhLyN5S2MEjrSE/normal-cryonics" title="Normal Cryonics">requires high
conscientiousness</a>
and can be easily be delayed until another point in life.</p>
<blockquote>
<p>I'll get around to doing it eventually</p>
</blockquote>
<p><em>— person who was cremated</em></p>
<p>However, it hasn't yet been explored whether this procrastination might be
rational if eventually followed through with: Many cryonics organisations
have high membership fees, which might be avoided by waiting.</p>
<p>To find this out, I first present a point-estimate model of
whether (and if yes, when) to sign up for cryonics. The model is
written in <a href="https://www.lua.org/">Lua</a>. I then proceed and create a
<a href="https://www.getguesstimate.com/">Guesstimate</a> model to determine the
distribution of the expected value of signing up.</p>
<h2 id="Other_Cryonics_Texts"><a class="hanchor" href="#Other_Cryonics_Texts">Other Cryonics Texts</a></h2>
<p>This write-up is not intended as an introduction to the concept
of cryonics. For a popular introduction to the topic that
clarifies many common misconceptions about the practice, see <a href="https://waitbutwhy.com/2016/03/cryonics.html" title="Why Cryonics Makes Sense">Urban
2016</a>.</p>
<p>For more basic information about the topic, there are the <a href="http://www.benbest.com/cryonics/CryoFAQ.html">Cryonics
FAQ</a> by Ben Best,
a former director of the Cryonics Institute, as well as <a href="https://alcor.org/FAQs/index.html">Alcor's
Cryonics FAQ</a>, which should answer
most questions people usually have about cryonics.</p>
<p>If you have decided that you indeed do want to sign up for
cryonics, the best existing resource is the <a href="https://www.lesswrong.com/s/weBHYgBXg9thEQNEe">Cryonics Signup
Guide</a>.</p>
<h2 id="Caveats"><a class="hanchor" href="#Caveats">Caveats</a></h2>
<p>This cost-benefit calculation depends on various factors that carry
significant uncertainty with them, and necessarily contains numerous
simplifications and inadequacies. The calculated values ought to be
equipped with gigantic error bars. I encourage others to make their
own calculations and models, compare results and bet on the relevant
probabilities.</p>
<p>Specifically, this analysis attempts to be relatively
conservative, think of the lower range of a 50% <a href="https://en.wikipedia.org/wiki/Confidence_interval">confidence
interval</a>. For example,
this leads to excluding singularity scenarios with lifespans of billions
or trillions of years at enormous quality. This is a balancing act,
some people might criticize that the number of worlds with indefinite
life extension and cryonics revival but without cosmic endowment is
very small. These concerns might be correct, but bare resemblance to
<a href="ttps://en.wikipedia.org/wiki/Pascal's_mugging">Pascal's mugging</a>-like
scenarios. In order to avoid such concerns, I focus on <em>relatively</em>
unspectacular visions of the future. If cosmic-endowment scenarios are
more likely than I suspect, this only strengthens the case for signing up.</p>
<h2 id="CostBenefit_Calculation_for_Cryonics"><a class="hanchor" href="#CostBenefit_Calculation_for_Cryonics">Cost-Benefit Calculation for Cryonics</a></h2>
<blockquote>
<p>If you make 50k\$/yr now, and value life-years at twice your income,
and discount future years at 2% from the moment you are revived for a
long life, but only discount that future life based on the chance it
will happen, times a factor of 1/2 because you only half identify with
this future creature, then the present value of a 5% chance of revival
is \$125,000, which is about the most expensive cryonics price now.</p>
</blockquote>
<p><em>— <a href="https://en.wikipedia.org/wiki/Robin_Hanson">Robin Hanson</a>, <a href="http://www.overcomingbias.com/2009/03/break-cryonics-down.html">“Break Cryonics Down”</a>, 2009</em></p>
<p>To find out whether to sign up for cryonics at all, one needs
to make a cost-benefit calculation. This has been <a href="http://www.overcomingbias.com/2009/03/break-cryonics-down.html" title="Break Cryonics Down">attempted
before</a>,
but that analysis has been rather short
(disregarding several important factors) and it might be productive to
approach the topic independently.</p>
<p>The costs of cryonics are comparatively easy to calculate and contain
little uncertainty: The price of cryopreservation and life-insurance
are widely known, and can be easily added together. The benefits of
cryopreservation, however, contain a lot more uncertainty: It is not at
all clear that the technology for resuscitation will be developed, cryonics
organizations and humanity survive to develop such technology, or that the
future will be interested in resuscitating people from cryopreservation.</p>
<p>The model presented makes the assumption that a person has a given age
and has the option of waiting for signing up for cryonics every year
up to their expected year of death. So, for example, a person that is
20 years old now is able to plan signing up when they are 20 years old,
21 years, 22 years and so on up to 78 years. The value of cryonics is
calculated, and the value of a regular death is tacitly assumed to be
\$0.</p>
<pre><code>curage=20
actval={78.36, 78.64, 78.66, 78.67, 78.68, 78.69, 78.69, 78.70, 78.71, 78.71, 78.72, 78.72, 78.73, 78.73, 78.74, 78.75, 78.75, 78.77, 78.79, 78.81, 78.83, 78.86, 78.88, 78.91, 78.93, 78.96, 78.98, 79.01, 79.03, 79.06, 79.09, 79.12, 79.15, 79.18, 79.21, 79.25, 79.29, 79.32, 79.37, 79.41, 79.45, 79.50, 79.55, 79.61, 79.66, 79.73, 79.80, 79.87, 79.95, 80.03, 80.13, 80.23, 80.34, 80.46, 80.59, 80.73, 80.88, 81.05, 81.22, 81.42, 81.62, 81.83, 82.05, 82.29, 82.54, 82.80, 83.07, 83.35, 83.64, 83.94, 84.25, 84.57, 84.89, 85.23, 85.58, 85.93, 86.30, 86.68, 87.08, 87.49, 87.92, 88.38, 88.86, 89.38, 89.91, 90.47, 91.07, 91.69, 92.34, 93.01, 93.70, 94.42, 95.16, 95.94, 96.72, 97.55, 98.40, 99.27, 100.14, 101.02, 101.91}
for age=curage,math.floor(actval[curage]) do
print(value(age) .. ": " .. age)
end
</code></pre>
<p><code>curage</code> contains the current age of the user of the
program. <code>actval</code> is an actuarial table that contains at the
nth position the average life expectancy of a person that is
n years old at the moment for a western nation (in this case
<a href="https://www.destatis.de/DE/Themen/Gesellschaft-Umwelt/Bevoelkerung/Sterbefaelle-Lebenserwartung/_inhalt.html">Germany</a>).</p>
<h2 id="The_Disvalue_of_Waiting"><a class="hanchor" href="#The_Disvalue_of_Waiting">The Disvalue of Waiting</a></h2>
<p>Two important factors play into the value (or disvalue) of waiting
to sign up for cryonics: <a href="./notes.html#Value_Drift">Motivation drift</a>
and the possibility of dying before signing up.</p>
<pre><code>function value(age)
return prob_signup(age)*prob_liveto(age)*(benefit(age)-cost(age))
end
</code></pre>
<h3 id="Motivation_Drift"><a class="hanchor" href="#Motivation_Drift">Motivation Drift</a></h3>
<!--TODO: find out whether there has been any research into the concrete shape
of motivation drift-->
<p><code>prob_signup</code> is a function that calculates the probability of
signing up for cryonics after having waited up to having a certain
age. It seems clear that people loose motivation to finish plans over
time, especially if they are unpleasant or complex.</p>
<p>A good example for this is people being motivated at the start of the
year to do regular exercise: How many of those actually keep their
promises to themselves? They might start off exercising, but after the
first few weeks the first people drop out, and and a couple of months
there is nearly nobody left still going to the gym except the ones who
already did it before. It seems like there is a strong <a href="https://en.wikipedia.org/wiki/Regression_toward_the_mean">regression to
the mean</a> in
regards to action: Most regular actions are replaced by inaction, most
strong values are replaced by apathy over time. A similar phenomenon
seems likely for signing up for cryonics: At first, people are very
enthusiastic about signing up, but then loose interest as time progresses.</p>
<p>It isn't obvious to me how strong motivation drift is and how
it develops over time (some people might regain motivation
after some time), but intuitively it seems like a <a href="https://en.wikipedia.org/wiki/Geometric_distribution">geometric
distribution</a>. The
reasoning is as follows: Imagine that a thousand people have the
motivation to perform a given action <code>$n$</code> years into the future. Every year,
a certain percentage <code>$p$</code> of the people who were still motivated loses
interest in performing that action and drop out. After <code>$n$</code> years, the
number of people who perform the action is <code>$1000*p^n$</code> (the percentage
of people still motivated is <code>$p^n$</code>). This also works if some people
get motivated again to do something after a number of years, as long as
their decisions are independently and identically distributed.</p>
<p>When trying to find out what the value of <code>$p$</code> is for oneself, one can
imagine a thousand independent identical copies of oneself planning on
executing a complex plan one year ahead. How many of those would actually
follow through on that plan? Intuitively, I'd say that it can't be much
higher than 95%, possibly much lower, especially for something as complex
and time-consuming as signing up for cryonics.</p>
<pre><code>decay=0.95
function prob_signup(age)
return decay^(age-curage)
end
</code></pre>
<p>Interestingly, this does not mean that the decision of whether to be
cryogenically preserved or not is then set in stone as soon as possible:
<em>Cryonics memberships are very easy to cancel</em>, very often a simple
email and a cessation of paying membership fees suffices. Signing up for
cryonics earlier protects against regression to the mean, which means
apathy or lack of motivation towards cryonics, but does not protect
against changing ones mind about cryonics: If one becomes convinced it's
bullshit later, one can easily get out (much more easily than getting in).
On the other hand, there might be a feeling of considerable sunk cost
due to already paid membership fees and the acquired life insurance.</p>
<p>In this analysis, it will be assumed that once one is signed up for
cryonics, one stays signed up for it.</p>
<!--TODO: Image here-->
<h3 id="Dying_Before_Signing_Up"><a class="hanchor" href="#Dying_Before_Signing_Up">Dying Before Signing Up</a></h3>
<p>If you die before signing up, all possible value (or disvalue) of cryonics
gets lost. So we want to calculate the probability of dying before having
a certain age given being currently <code>curage</code> years old.</p>
<p>Mortality rates are usually calculated using a <a href="https://en.wikipedia.org/wiki/Gompertz_distribution">Gompertz
distribution</a>. I
determined the b and eta values by eyeballing <a href="https://www.wolframalpha.com/input/?i=life+expectancy+of+a+0+year+old+german">Wolfram
Alpha</a>
and using a calculator in <a href="https://reducing-suffering.org/estimating-aggregate-wild-animal-suffering-from-reproductive-age-and-births-per-female/#Choosing_a_distribution" title="Estimating Aggregate Wild-Animal Suffering from Reproductive Age and Births per Female">Tomasik
2016</a>.</p>
<!--TODO: find out which exact values statisticians
use, then use them: Maybe take cue from here:
https://www.gwern.net/Longevity#population-survival-curve-->
<pre><code>b=0.108
eta=0.0001
function gompertz(age)
return math.exp(-eta*(math.exp(b*age)-1))
end
</code></pre>
<p><code>gompertz</code> returns the probability of reaching <code>age</code>
starting from birth, but I need the probability of reaching
<code>age</code> given one is already <code>curage</code> years old. With <a href="https://en.wikipedia.org/wiki/Bayes'_theorem">Bayes
theorem</a> one can calculate
that</p>
<div>
$$Pr[X \ge age|X \ge curage]\\
=\frac{Pr[X \ge curage \cap X \ge age]}{Pr[X \ge curage]}\\
=\frac{Pr[X \ge age]}{Pr[X \ge curage]}$$
</div>
<p><code>$Pr[X \ge curage \cap X \ge age]$</code> is equal to <code>$Pr[X \ge age]$</code> because
being older than <code>age</code> is (in this calculation) a subset of being older
<code>curage</code>, and <code>$A \subset B \Rightarrow A \cap B=A$</code>. Some precautions
have to apply in the case that the probabilities of reaching <code>age</code>
is not completely independent of the probability of reaching <code>curage</code>,
but those are difficult to estimate and will not be implemented here.</p>
<p>This way, one can implement the probability of living until <code>age</code> given
<code>curage</code> the following way:</p>
<pre><code>function prob_liveto(age)
return gompertz(age)/gompertz(curage)
end
</code></pre>
<h3 id="Longevity_Escape_Velocity"><a class="hanchor" href="#Longevity_Escape_Velocity">Longevity Escape Velocity</a></h3>
<p><a href="https://en.wikipedia.org/wiki/Longevity_escape_velocity">Longevity Escape
Velocity</a> (short
LEV) is the name for the possible year when anti-aging technology becomes
so good that people can be rejuvenated faster than they age. Although the
concept is considered idle speculation in many circles, many futurists
justify not signing up for cryonics because they expect that LEV will
arrive during their lifetime, and see no reason to sign up for a cryonics
membership they are probably not going to need anyway. In this text,
I will consider LEV by assuming there will be a certain year after which
the probability of death from aging is practically zero.</p>
<p>I somewhat arbitrarily set this year to 2080<!--TODO: find some
estimates of LEV, maybe Metaculus?-->, though many futurists seem more
optimistic<!--TODO source-->:</p>
<pre><code>levyear=2080
</code></pre>
<p>However, there are some reasons why one might want to stay signed up
to cryonics even after LEV—there might be types of accidents where
large parts of the body are destroyed in ways that both lead to death
and that regenerative medicine at the time can't handle, and so one will
want to be preserved until the missing bodyparts can be re-created and
resuscitation is possible. Another risk to life even after LEV is new
and for some time deadly diseases. This analysis does not include such
considerations (yet).</p>
<!--
TODO: make probability of extinction explicit, not just implicit in
the expected future life years.
"If one believes that a singularity-like scenario will occur relatively soon,
and estimates the risk of human extinction from such a scenario to be high,"
-->
<!--https://www.fightaging.org/archives/2018/02/aubrey-de-grey-on-progress-and-timescales-in-rejuvenation-research/-->
<!--
Metaculus questions related to longevity/LEV:
https://www.metaculus.com/questions/4125/what-will-be-the-first-year-during-which-most-deaths-are-from-non-disease-related-causes-and-fewer-than-1-of-humans-die/
https://www.metaculus.com/questions/4110/assuming-some-humans-are-alive-what-will-be-the-birthdate-of-the-oldest-living-human-on-january-1st-2200/
https://www.metaculus.com/questions/353/will-someone-born-before-2001-live-to-be-150/
https://www.metaculus.com/questions/1628/what-will-be-the-longest-verified-human-lifespan-on-record-on-january-1-2050/
-->
<h2 id="Calculating_the_Cost"><a class="hanchor" href="#Calculating_the_Cost">Calculating the Cost</a></h2>
<p>Calculating the cost is comparatively straightforward, but there are
some hidden variables (like opportunity costs and social costs) that
have to be considered (not all of these in this text).</p>
<p>The raw cost for cryonics depends heavily on the organisation choosen for
preservation, the price ranges from ~\$20000 to ~\$250000<!--TODO: Sources
for these, maybe more exact numbers for different organisations.-->. In
this case, I choose the costs for neurocryopreservation at Alcor, though
this analysis should be extended to other organisations.</p>
<p>Raw cryonics cost can be split into three different parts:
membership fees, comprehensive member standby costs and the cost for
cryopreservation.</p>
<pre><code>function cost(age)
return membership_fees(age)+pres_cost(age)+cms_fees(age)
end
</code></pre>
<h3 id="Membership_Fees"><a class="hanchor" href="#Membership_Fees">Membership Fees</a></h3>
<!--TODO: fees have changed! Update-->
<p>Membership fees for Alcor are calculated using the age of the member
and the length of their membership.</p>
<h4 id="Direct_Fees"><a class="hanchor" href="#Direct_Fees">Direct Fees</a></h4>
<blockquote>
<p>Current Membership Dues, net of applicable discounts, are:<br/>
1. First family member: \$525 annually or \$267 semi-annually or \$134 quarterly.<br/>
2. Each additional family member aged 18 and over, and full-time students aged 25 and under: \$310 annually or \$156 semi-annually or \$78 quarterly.<br/>
3. Each minor family member under age 18 for the first two children (no membership dues are required for any additional minor children): \$80 annually or \$40 semi-annually or \$20 quarterly<br/>
4. Full-time student aged 26 to 30: \$460 annually or \$230 semi-annually or \$115 quarterly.<br/>
5. Long-term member (total membership of 20 - 24 years): \$430 annually or \$216 semi-annually or \$108 quarterly.<br/>
6. Long-term member (total membership of 25 - 29 years): \$368 annually or \$186 semi-annually or \$93 quarterly.<br/>
7. Long-term member (total membership of 30 years or longer): \$305 annually or \$154 semi-annually or \$77 quarterly.<br/>
8. Long-term member (total membership of 40 years or longer): \$60.00 annually or \$30.00 semi-annually or \$15.00 quarterly</p>
</blockquote>
<p><em>— <a href="https://alcor.org/">Alcor Life Extension Foundation</a>, <a href="https://alcor.org/BecomeMember/scheduleA.html">“Alcor Cryopreservation Agreement - Schedule A”</a>, 2016</em></p>
<p>The following assumptions will be made in the implementation:</p>
<ol>
<li>The person considering signing up for cryonics is over 18 years old.</li>
<li> If the person is under 25 years old, they are a
student. Considering the fact that cryonics members seem to
be more likely to be rich and educated, this seems likely,
though maybe a bit classist. The code can be changed if personal
need arises.</li>
<li>If the person is over 25 years old, they are not a student.</li>
<li>The person stays a member until their death (otherwise the cryonics
arrangement doesn't apply).</li>
<li> The membership fees will not be changed drastically over time. In
fact, inflation adjusted prices for cryonics have mostly stayed
constant (see chart), so this is a reasonable assumption.</li>
<li> The cryonicist will know when LEV has occurred, and will cancel
their membership starting from that year.</li>
</ol>
<p><img alt="Inflation adjusted historical cryonics prices" src="https://timelines.issarice.com/images/7/79/Historical_cost_of_cryonics%2C_inflation_adjusted.png" title="Inflation adjusted historical cryonics prices"/></p>
<!--TODO: make this image local-->
<p>The implementation is quite straightforward:</p>
<pre><code>function alcor_fees(age)
local left=math.min(math.floor(actval[age])-age, levyear-curyear)
local cost=0
if age<25 then
newage=25
cost=(newage-age)*310
end
if left>=30 then
cost=cost+(left-30)*305
left=30
end
if left>=25 then
cost=cost+(left-25)*368
left=24
end
if left>=20 then
cost=cost+(left-20)*430
left=20
end
if age<=25 then
cost=cost+(left-(25-age))*525
else
cost=cost+left*525
end
return 300+cost
end
</code></pre>
<h4 id="Comprehensive_Member_Standby"><a class="hanchor" href="#Comprehensive_Member_Standby">Comprehensive Member Standby</a></h4>
<!--TODO: find out how CMS really works!-->
<blockquote>
<p>For Members residing in the continental U.S. and Canada: Alcor will
provide Comprehensive Member Standby (CMS) to all Members (standby
in Canada may be subject to delays due to customs and immigration
requirements), which includes all rescue activities up through the time
the legally pronounced Member is delivered to the Alcor operating room
for cryoprotection. <strong>This charge is waived for full-time students under
age 25 and minors (under age 18).</strong></p>
</blockquote>
<p><em>— <a href="https://alcor.org/">Alcor Life Extension Foundation</a>, <a href="https://alcor.org/BecomeMember/scheduleA.html">“Alcor Cryopreservation Agreement - Schedule A”</a>, 2016</em></p>
<p>Emphasis mine.</p>
<blockquote>
<p>Current CMS charges are:<br/>
\$180 annually, \$90 semi-annually, or \$45 quarterly</p>
</blockquote>
<p><em>— <a href="https://alcor.org/">Alcor Life Extension Foundation</a>, <a href="https://alcor.org/BecomeMember/scheduleA.html">“Alcor Cryopreservation Agreement - Schedule A”</a>, 2016</em></p>
<p>I will assume that the cryonics member starts paying a CMS fee starting
10 years before their actuarial age of death.</p>
<pre><code>cms=180
function cms_age(age)
return actval[age]-10
end
function cms_fees(age)
return cms*(actval[age]-cms_age(age))
end
</code></pre>
<h3 id="Preservation_Cost"><a class="hanchor" href="#Preservation_Cost">Preservation Cost</a></h3>
<p>There are several different methods of funding cryonics, the most
popular of which seems to be life insurance. I haven't spent much time
investigating the price finding mechanisms of life insurance companies,
so I make the assumption that the insurance companies price their products
adequately, so one doesn't have much of a financial advantage by choosing
life insurance as opposed to simply saving money & paying the cryonics
membership in cash. I also assume that life insurance companies can
accurately price in the arrival date of LEV.</p>
<blockquote>
<p>Minimum Cryopreservation Funding:<br/>
• \$200,000.00 Whole Body Cryopreservation […].<br/>
• \$80,000.00 Neurocryopreservation […].<br/>
[…]<br/>
Surcharges:<br/>
• \$10,000 Surcharge for cases outside the U.S. and Canada other than China.<br/>
• \$50,000 Surcharge for cases in China.<br/>
[…]</p>
</blockquote>
<p><em>— <a href="https://alcor.org/">Alcor Life Extension Foundation</a>, <a href="https://alcor.org/BecomeMember/scheduleA.html">“Alcor Cryopreservation Agreement - Schedule A”</a>, 2016</em></p>
<!--TODO: image that shows that these have been fairly consistent over time-->
<p>I assume that the person considering signing up lives outside of the
U.S (but not in China), since a lot more people live outside the U.S
than inside of it. I also assume that the person wants to sign up for
neurocryopreservation. With these assumptions, the function that returns
preservation costs becomes quite simple:</p>
<pre><code>function pres_cost(age)
return 90000
end
</code></pre>
<h3 id="Other_Possible_Costs"><a class="hanchor" href="#Other_Possible_Costs">Other Possible Costs</a></h3>
<p>There is a number of different additional costs that have not been
considered here because of their (perceived) small scale or difficult
tractability.</p>
<p>These include opportunity costs for the time spent informing oneself
about cryonics (tens of hours spent), opportunity costs for the time
spent signing up (tens of hours spent), social costs by seeming weird
(though cryonics is easy to hide, and most cryonicists seem to be rather
vocal about it anyways), and alienating family members who necessarily
come into contact with cryonics (considering the <a href="https://www.biostasis.com/is-that-what-love-is-the-hostile-wife-phenomenon-in-cryonics/" title="Is That What Love is? The Hostile Wife Phenomenon in Cryonics">"Hostile Wife
Phenomenon"</a>).</p>
<h2 id="Calculating_the_Benefit"><a class="hanchor" href="#Calculating_the_Benefit">Calculating the Benefit</a></h2>
<p>Calculating the benefit of cryonics carries a great uncertainty,
but basically it can be divided into seven distinct components: The
probability of being preserved, the probability of revival, the amount
of years gained by cryonics, the value of one lifeyear, the probability
of living to the year when one will sign up, the probability of then
dying before LEV, and the expected quality of preservation.</p>
<pre><code>function benefit(age)
return prob_pres*prob_succ*years_gain*val_year*prob_liveto(age)*prob_diebeforelev(age)*avg_pres_quality(age)
end
</code></pre>
<p>Here, I will only take <a href="https://en.wikipedia.org/wiki/Point_estimation">point
estimates</a> of these
values.</p>
<h3 id="Value_of_a_Lifeyear_in_the_Future"><a class="hanchor" href="#Value_of_a_Lifeyear_in_the_Future">Value of a Lifeyear in the Future</a></h3>
<!--TODO: Dangers from incompetent revival -->
<p>Much ink and pixels have been spilled on the question of the quality of
the future, very little of it trying to make accurate or even resolvable
predictions. Anders Sandberg summarizes the current approaches:</p>
<blockquote>
<p>Some of this discourse relates to the issue of whether history is
cyclical, spiralling, progressive or regressive. Is there a particular
shape to history? If there is, then this makes a strong long-term
prediction about the future. While this could be a teleological arc
towards completion, it could also be an unavoidable thermodynamic decline,
the environmentally biased random walk of evolution or an emergent
trajectory set by a mixture of forces, events and individuals. In
a sense a sufficiently large-scale macrohistorical theory will tend
to become a prospective universal history. Still, even the idea of a
prospective history is tied to a particular (forward-looking, linear)
concept of time: many macrohistorians have been concerned with very
different concepts of time and its structure.<br/>
Another key issue is how the patterns of history are caused and shaped.
Beside any transcendental or environmental forces shaping history,
who (if any) are the vanguard? This might for example be enlightened
leaders (Ssu-Ma Ch’ien, Sarkar), creative elites or intellectuals
(Toynbee, Gramsci), people seizing power (Khaldun), or the proletariat
(Marx). An issue that will be relevant in this book is to what degree we
can expect deliberate shaping of future trajectories to be possible. Many
macrohistorical theories deny this, since the shape is largely determined
by general patterns rather than individual decisions, yet some are also
compatible with contingency and freedom along some dimensions, or that
certain stages and patterns can be broken under the right conditions.</p>
</blockquote>
<p><em>—Anders Sandberg, “Grand Futures”, 2023</em></p>
<p>One way to look at the question could be to find clear
metrics that encapsulate the most important human values and
then fund a prediction market to bet on these metrics. This
could include the power of humanity to make most important
decisions regarding its development and resource management,
diversity among human beings, average happiness and lifespans and
other variables such as inequality regarding resources (see also <a href="https://reflectivedisequilibrium.blogspot.com/2013/12/what-proxies-to-use-for-flow-through.html" title="What proxies to use for flow-through effects?">Shulman
2013</a> for a more
extensive list of metrics for flow-through effects, which could be
used to evaluate the general quality of life in the future). <a href="http://lukemuehlhauser.com/three-wild-speculations-from-amateur-quantitative-macrohistory/" title="Three wild speculations from amateur quantitative macrohistory">Muehlhauser
2017a</a>
and <a href="http://lukemuehlhauser.com/industrial-revolution/" title="How big a deal was the Industrial Revolution?">Muehlhauser
2017b</a>
find that along 5 different metrics, human well-being has been improving
increasingly rapidly since the industrial revolution.</p>
<p>But a much simpler way of approaching the topic could be the following:
One takes arguments from both sides (proclaiming positive futures
and negative futures) and prematurely concludes that the future is on
average going to be neutral, with a high variance in the result. But
some problems present themselves: In different value systems, "neutral"
means very different things. Strictly speaking, a utilitarian would
see human extinction as neutral, but not net neutral (the utility of
a world without any sentient beings is exactly 0, which is presumably
lower than the current value of the world), anti-natalists consider an
empty world to be a positive thing, and most people working on preventing
human extinction would consider such a world to be a gigantic loss of
opportunity, and therefore net negative.</p>
<p>There seems to be no simple way to resolve these conflicts, otherwise it
would have been written down up to now. But it seems like most people
would take the current state of affairs as neutral, with improvements
in happiness, meaning and wealth to be positive, and decreases in
those to be negative. Also, they don't see dying tomorrow as a neutral
event.</p>
<h4 id="Caveats_on_Future_Life_Years"><a class="hanchor" href="#Caveats_on_Future_Life_Years">Caveats on Future Life Years</a></h4>
<p>Here I will assume that</p>
<ul>
<li> One can exclude future life years with astronomically high values.
Although these deserve further consideration,
they are avoided to in order to prevent <a href="https://en.wikipedia.org/wiki/Pascal's_mugging">Pascal's
mugging</a>-like
arguments. Also, they might have quite small probabilities, and
may be even harder to forecast than naive success from cryonics</li>
<li> Future life years can be averaged in their quality, and that
average has monetary value</li>
<li> Future lifeyears are not <a href="https://en.wikipedia.org/wiki/Time_preference#Temporal_discounting">temporally
discounted</a>,
i.e a lifeyear 1000 years in the future is as valuable as the
next lifeyear</li>
<li> There are no <a href="https://en.wikipedia.org/wiki/Diminishing_returns">diminishing marginal
returns</a>
to lifeyears, i.e 1000 additional life years are 1000 times as
valuable as one additional lifeyear</li>
</ul>
<p><a href="./doc/philosophy/ethics/discounting_future_health_greaves_2017.pdf" title="Discounting future health">Greaves
2017</a>
argues against pure positive discounting for health (i.e QALYs):</p>
<blockquote>
<p>A discount rate should be applied to future health for two
reasons. First, to correct for the possibility that the beneficial
future effects of a proposed health intervention, in the relevant
sense of 'effect', are smaller than the model's naïve calculation
suggests. Second, to take into account the fact that earlier health
benefits are generally more instrumentally valuable than otherwise-equal
later ones.</p>
<p>In both cases, the use of a discount rate for future health is not
the only way to model the effects in question. Alternatively, one could
(i) explicitely represent the various possible actual effects of one's
intervention and their probabilities, and perform a corresponding explicit
expected-value calculation, eschewing the use of the 'naïve calculation'
as an estimaate of effects on future health, and (ii) explicitely model
the instrumental as well as the intrinsic benefits of health improvements
one is considering. If one did both of those things, there would be no
reasons of the kinds we have accepted for discounting future health.</p>
</blockquote>
<p><em>— <a href="https://en.wikipedia.org/wiki/Hilary_Greaves">Hilary Greaves</a>, <a href="./doc/philosophy/ethics/discounting_future_health_greaves_2017.pdf">“Discounting future health”</a> p. 7, 2017</em></p>
<p>She also argues against applying considerations from diminishing marginal
returns to health (although it must be noted that this analysis does
not explicitely use QALY numbers for cryonics, since they have not been
collected by healthcare departments yet):</p>
<blockquote>
<p>this does not give rise to any phenomenon of diminishing
marginal returns of health to well-being when units of health are
measured in QALYs, because the QALY scale has already more-or-less
been constructed to be linear in well-being. There is therefore
no reason based in considerations of diminishing marginal returns
to discount future QALYs.</p>
</blockquote>
<p><em>— <a href="https://en.wikipedia.org/wiki/Hilary_Greaves">Hilary Greaves</a>, <a href="./doc/philosophy/ethics/discounting_future_health_greaves_2017.pdf">“Discounting future health”</a> p. 3, 2017</em></p>
<p>Additionally, problems with interpersonal utility comparison do not apply
in this case, since the externalities from signing up for cryonics are
disregarded in this analysis.</p>
<h4 id="QALYs_and_VSL"><a class="hanchor" href="#QALYs_and_VSL">QALYs and VSL</a></h4>
<p>There are two different methods of putting a value on human life:
the <a href="https://en.wikipedia.org/wiki/Value_of_life">VSL</a> and the
<a href="https://en.wikipedia.org/wiki/Quality-adjusted_life_year">QALY</a>.
The Wikipedia page on VSL
<a href="https://en.wikipedia.org/wiki/Value_of_life#Estimates_of_the_value_of_life">lists</a>
\$181893 for the value of a year of life in Australia, and \$50000 as the
"de facto international standard most private and government-run health
insurance plans worldwide use to determine whether to cover a new medical
procedure". This number seems like a good conservative estimate.</p>
<p>Interestingly, this approximately equals a year of waking hours worth
the minimum wage in some countries (<code>$\$10*16*7*52=\$58240$</code>)<!--TODO:
add https://en.wikipedia.org/wiki/List_of_minimum_wages_by_country-->.</p>
<p>Intuitively, the probability distribution over the value of a year of
life in the future should then look like this:</p>
<pre><code>.l("nplot")
.l("nstat")
grid([-20000 120000 20000];[0 0.00004 0.000004])
xtitle("Dollar value of a future life year")
ytitle("Probability")
plot({n.pdf(x;50000;500000000)})
draw()
</code></pre>
<p><img alt="Probability distribution over the value of a lifeyear in the future" src="./img/considerations_on_cryonics/avg_fut_year_val.png" title="The probability distribution over the value of a lifeyear in the future. The distribution is a normal distribution with an expected value of 50000 and a standard deviation of ~22000."/></p>
<p><em>this graph is not based on real data and only here for illustrative purposes</em></p>
<p>But one can take another factor into account: Most negative future
scenarios don't lead to resuscitation (civilisational collapse, stable
totalitarianism, existential catastrophes like AI failure, nuclear
war, biotechnological disaster and natural catastrophe all reduce
human capabilities or keep them constant, preventing the development
of resuscitation technology). In most of the negative futures, there
are either no more humans around or people don't have time, energy or
resources to bring people back from cryonic preservation (if indeed they
still <em>are</em> in preservation by that point), and for malicious actors, in
most scenarios it is easier to create new people than to bring preserved
people back.</p>
<p>This effect might be dampened by the consideration that most possible
futures have net-negative value, but on the other hand, nearly all of
those futures don't lead to resuscitation.</p>
<p>Furthermore, <a href="https://forum.effectivealtruism.org/posts/gMzWWtdsArkwqNJEK/why-might-the-future-be-good" title="Why might the future be good?">Christiano 2013</a> outlines two reasons that indicate that the future will be good:</p>
<blockquote>
<ol>
<li>Decisions will be made by people whose lives are morally valuable
and who want the best for themselves. They will bargain amongst each
other and create a world that is good to live in. Because my values are
roughly aligned with their aggregate preferences, I expect them to create
a rich and valuable world (by my lights as well as theirs).<br/></li>
<li>Some people in the future will have altruistic values broadly similar
to my own, and will use their influence to create a rich and valuable
world (by my lights as well as theirs).</li>
</ol>
</blockquote>
<p><em>— Paul Christiano, <a href="https://forum.effectivealtruism.org/posts/gMzWWtdsArkwqNJEK/why-might-the-future-be-good">“Why might the future be good?”</a>, 2013</em></p>
<p>This would mean that the probability distribution over the value of
a lifeyear in the future conditional on being resuscitated could look
like this:</p>
<pre><code>.l("nplot")
.l("nstat")
grid([-20000 120000 20000];[0 0.00004 0.000004])
xtitle("Dollar value of a future life year")
ytitle("Probability")
plot({:[x>50000;n.pdf(x;50000;500000000);0.4472*n.pdf(x;50000;100000000)]})
draw()
</code></pre>
<p><img alt="Probability distribution over the value of a lifeyear in the future conditional on being resuscitated" src="./img/considerations_on_cryonics/cryo_fut_year_val.png" title="The probability distribution over the value of a lifeyear in the future conditional on being resuscitated. The distribution is two halves of two different normal distributions meeting at around $50000, the left half has much lower variance than the right half."/></p>
<!--TODO: wouldn't that just shift the normal curve rightwards? Think about it.-->
<p><em>this graph is also not based on real data and only here for illustrative purposes</em></p>
<h4 id="Negative_Scenarios"><a class="hanchor" href="#Negative_Scenarios">Negative Scenarios</a></h4>
<p>However, I can think of 3 very specific (and thereby highly unlikely)
scenarios where people could be resuscitated into a (for them) net-negative
world.</p>
<!--See also:
https://www.lesswrong.com/posts/MzmMZmSzyeLPjLQjD/survey-what-s-the-most-negative-plausible-cryonics-works-->
<h5 id="Ascended_Economy"><a class="hanchor" href="#Ascended_Economy">Ascended Economy</a></h5>
<p>The <a href="http://slatestarcodex.com/2016/05/30/ascended-economy/">ascended
economy</a> is a
scenario where the development of capitalism diverges significantly from
the desires of humans, leading to most (if not all) of humanity becoming
extinct. It seems highly unlikely, but possible that cryopreserved
humans are placed into the hands of an algorithm that invests the
money in the relevant funds to resuscitate the cryopreserved humans at a
certain point. This algorithm could receive little (or no) information
on what to do with the resuscitated humans afterwards, leading either
to these humans quickly dying again because of an economy where they
are worthless, or being kept alive solely for fulfilling the contract
that is embedded in the algorithm. This might lead to insanity-inducing
boredom as the humans are kept alive as long as algorithm manages to,
possibly hundreds or thousands of years. This would have net-negative
value for the people resuscitated.</p>
<h5 id="Malevolent_Future_Actors"><a class="hanchor" href="#Malevolent_Future_Actors">Malevolent Future Actors</a></h5>
<p>A superintelligence becomes a
<a href="https://en.wikipedia.org/wiki/Singleton_(global_governance)">singleton</a>
and starts behaving malevolently because of a <a href="https://reducing-suffering.org/near-miss/" title="Astronomical
suffering from slightly misaligned artificial intelligence">near
miss</a>
in its implementation or or because it has been set up by a
malevolent human. This would lead to cryopreserved people being
resuscitated, having their brains scanned and executed as a <a href="https://en.wikipedia.org/wiki/Mind_uploading">brain
emulation</a>, copied and
put into very painful conditions.</p>
<h5 id="Information_from_the_Past_is_Valuable"><a class="hanchor" href="#Information_from_the_Past_is_Valuable">Information from the Past is Valuable</a></h5>
<p>In a future where agents that don't care about humans find the
cryopreserved remains of humans, they might be interested in extracting
information from those brains. If it is not possible to extract this
information without reviving the cryopreserved people, they might
resuscitate them and then interrogate these revived people for a very
long time, with little regard for their well-being.</p>
<h5 id="Being_Revived_as_an_Emulation_and_Forced_to_Work"><a class="hanchor" href="#Being_Revived_as_an_Emulation_and_Forced_to_Work">Being Revived as an Emulation and Forced to Work</a></h5>
<p>A relatively common worry that prevents people from signing up for
cryonics is a scenario in which their brain is scanned and they are
transformed into a whole-brain emulation, which will then be used
for slave labour or worse. This is often accompanied by linking <a href="https://qntm.org/mmacevedo" title="Lena">qntm
2021</a>.</p>
<p>For this to occur, several things need to happen:</p>
<ol>
<li>The technology for whole-brain emulations is developed</li>
<li>The brain of the person is scanned</li>
<li>The scanned person is forced to work</li>
<li>Optionally, the preserved person is copied many times over and the copies are all used for work</li>
</ol>
<p>For 2 to occur, somebody needs to be interested in emulating
the person in question. This could be because they are part
of the cryonics organisation, and want to fulfill the contract
that the cryonics organisation and its members have signed,
or because they want to use the person for slave labour.
In the latter case, they will usually seek the most profitable
option, and several forces work against the preserved person
being that option:</p>
<ul>
<li>If whole-brain emulation is wide-spread, and copying of whole-brain emulations is easy:
<ul>
<li>Competition among whole-brain emulations is likely fierce, and labor by emulations would be as cheap as slave labour (so there is no particular reason to enslave emulations)</li>
<li>The average emulation is likely very competent and elite, and a much better worker than the average college-educated person (<a href="https://ageofem.com" title="The Age of Em">Hanson 2016 pp. 161-163</a>) estimates that most ems are going to be copies of fewer than a thousand or so original humans, each most likely again selected from several thousand individual childhoods to select the most capable educational lineage)</li>
</ul></li>
<li>If whole-brain emulation is not wide-spread, then the more the more more-obscure whole-brain emulation is, the more unlikely it is that cryonics and whole-brain emulation will "come into contact", since cryonics is also obscure and the number of preserved patients is tiny (<a href="https://docs.google.com/spreadsheets/d/1vBrlNvPkPQbLKxtSUlTF_ZpUS4lwW2W7lENRj9Gdk3I/edit#gid=662993486"><500</a>, with <a href="https://lessdead.com/how-many-humans-will-have-their-brain-preserved-forecasts-and-trends">much more to come</a>)
<ul>
<li>It seems unlikely that whole-brain emulation will stay restricted to a niche, since it will likely require large amounts of computational power and high-tech scanning facilities which are only possible with large amounts of capital</li>
<li>A factor working against this is that whole-brain emulation and cryonics are discussed within in the same subculture</li>
</ul></li>
<li>Uploading a <a href="https://en.wikipedia.org/wiki/Cryopreservation#Vitrification">vitrified</a> brain seems to be at least as difficult as scanning a freshly deceased brain or a living brain, and possibly much more difficult, since one would need to repair structural damage done by outdated cryopreservation methods (why not vitrify a recently deanimated person with the newest cryogenic technology, and then scan & upload them, instead of trying to fix old somewhat broken brains?), so it would probably not be more economically viable</li>
<li>It appears more likely than not that cryonics organisations will attempt to prevent the theft of preserved humans (although the organisations could fail and abandon or sell the preserved humans)</li>
<li>There would have to be no laws protecting cryogenically preserved humans (or the emulations resulting from them), but laws protecting living or recently deanimated people</li>
</ul>
<p>A basic tension is that if whole-brain emulation is wide-spread and the
incentive is economic, then it'll use more competitive brains from living
humans that are easier to scan. If whole-brain emulation is obscure,
it's unlikely that it and cryonics will come into contact in a way that
harms cryonics patients (at the very least, it would be counted as grave
robbery and body snatching).</p>
<p>That said, if scanning living or recently deceased humans is illegal
or very hard, but uploading and using cryogenically preserved humans is
legal or less difficult, then this might be a legitimate worry, because
a preserved human can't take any actions to prevent this scenario.</p>
<h4 id="Steps_for_Reducing_the_Risk_from_such_Scenarios"><a class="hanchor" href="#Steps_for_Reducing_the_Risk_from_such_Scenarios">Steps for Reducing the Risk from such Scenarios</a></h4>
<blockquote>
<p>b) When, in Alcor's best good faith judgement, it is determined
that attempting revival is in the best interests of the Member in
cryopreservation, Alcor shall attempt to revive and rehabilitate the
Member. It is understood by the Member that a careful assessment of
the risks versus the benefits of a revival attempt will be material to
determining when to attempt revival. […]<br/>
d) Where it is possible to do so, Alcor represents that it will be guided
in revival of the cryopreserved Member by the Member's own wishes and
desires as they may have been expressed in a written, audio, or video
<strong>Statement of Revival Preferences and Desires</strong>, which the Member may at
his/her discretion attach to this Agreement.</p>
</blockquote>
<p><em>— <a href="https://www.alcor.org/">Alcor Life Extension Foundation</a>, <a href="./doc/practical/cryopreservation_agreement_alcor_2012.pdf">“Cryopreservation Agreement”</a> p. 15/16, 2012</em></p>
<p>Although not a failsafe measure, steps can be taken to reduce
the risks from hellish scenarios above by making arrangements with
cryonics organisations. This may include not wanting cryopreservation
to continue in an ascended economy, objecting to revival as an
<a href="https://en.wikipedia.org/wiki/Brain_emulation">emulation</a> or revival
after more than a certain number of years (to prevent being resuscitated
in an incomprehensibly strange and alien world).</p>
<h4 id="Other_Thoughts"><a class="hanchor" href="#Other_Thoughts">Other Thoughts</a></h4>
<p>Many people argue that the value of a year of life in the future might
be much lower than in the present, because friends and familiy are not
around, and it is very likely that the future will be extremely alien and
unfamiliar.</p>
<p>These are valid considerations, but can be dampened a bit: Humans have
shown to adapt to very different and varied circumstances, and humans
today feel that modern life in big cities with regular calendars and
highly structured lives without any worries about survival is normal,
while for most humans who ever lived, this would be anything but. One
can speculate that very similar facts will also hold for the future
(becoming increasingly unlikely the further resuscitation lies in the
future). There would certainly be a big culture shock in the future,
but it seems not qualitatively different from the shock people have
when they visit different countries today. It is possible that future
societies might try to help people with this kind of future shock,
but that is of course far from certain.</p>
<p>It is true that most cryonicists will not be able to convince their
friends and family to sign up for it too, so they will be alone in
the future at first. People today sometimes leave their friends and
even families to move to other places, but those people seem to be
the exception rather than the norm. However, people nearly always
move on with their life, even as they get divorced, get estranged
from their friends or see their children less regularly—they
don't seem to prefer death to continuing their lives without
specific people. This consideration doesn't seem to be a <a href="https://www.lesswrong.com/rationality/is-that-your-true-rejection">true
rejection</a>.</p>
<hr/>
<!--
https://en.wikipedia.org/wiki/Value_of_life
-->
<p>After these considerations, I conservatively set the value of a lifeyear
in the future to \$50000.</p>
<pre><code>val_year=50000
</code></pre>
<h3 id="Probability_of_Revival"><a class="hanchor" href="#Probability_of_Revival">Probability of Revival</a></h3>
<blockquote>
<p>Specific equations and values have been
proposed, usually yielding probability of success
0 < x < 10%. For example, Steven Harris in 1989 estimated
<a href="http://www.alcor.org/Library/html/WillCryonicsWork.html">0.2-15%</a>,
R. Mike Perry in the same article runs a different analysis to arrive at
13-77%, <a href="http://www.merkle.com/cryo/probability.html">Ralph Merkle</a>
suggests >85% (conditional on things like good preservation,
no dystopia, and nanotech); Robin Hanson calculated in 2009 a
<a href="http://www.overcomingbias.com/2009/03/break-cryonics-down.html">~6%</a>
chance, Roko gave
<a href="http://lesswrong.com/lw/3j/rationality_cryonics_and_pascals_wager/">23%</a>;
Mike Darwin in 2011 (personal communication)
put the odds at <10%; an <a href="https://docs.google.com/spreadsheet/ccc?key=0Ajn1LpstEUO_dE00ZVVfa3pzX2Y2dk9mWWRKOUVkWlE#gid=0">informal survey of >6
people</a>
(<a href="http://lesswrong.com/lw/fz9/more_cryonics_probability_estimates/">LW
discussion</a>)
averaged ~17% success rate; <a href="http://www.jefftk.com/p/breaking-down-cryonics-probabilities">Jeff
Kaufman</a>
in 2011 provides a calculator with suggested
values yielding 0.2%; The <a href="http://lesswrong.com/lw/fp5/2012_survey_results/">2012 LessWrong
survey</a> yields a mean
estimate of cryonics working of 18% (n=1100) and among ‘veterans’ <a href="https://web.archive.org/web/20131229234145/http://squid314.livejournal.com/349656.html">the
estimate</a>
is a lower 12% (n=59) - but interestingly, they seem to be more likely
to be signed up for cryonics.</p>
</blockquote>
<p><em>— <a href="https://www.gwern.net">Gwern Branwen</a>, <a href="https://www.gwern.net/plastination">“Plastination versus Cryonics”</a>, 2014</em></p>
<p>Besides these estimates, there exist also two related questions
on the prediction website Metaculus.<!--TODO: dig out manifold questions on this--></p>
<ul>
<li><a href="https://www.metaculus.com/questions/455/if-you-die-today-and-get-cryonically-frozen-will-you-wake-up/">"If you die today and get cryonically frozen, will you "wake up"?"</a> receives 5% (n=436).</li>
<li><a href="https://www.metaculus.com/questions/4785/what-percentage-of-people-cryopreserved-between-2020-and-2050-will-be-revived-by-2200/">"What percentage of people cryopreserved between 2020 and 2050 will be revived by 2200?"</a> has a community prediction of 2.138% (n=44).</li>
</ul>
<p>I am not sure where the difference comes from, perhaps either
from worries about the quality of current preservation or from
a low trust in the longevity of cryonics organisations. <a href="https://docs.google.com/spreadsheets/d/1qn7c7rYmYx3KtqvhXTUGiiyuBk5e9kG3sA3jF-4zk6U/">This google
sheet</a>
contains 7 estimates of success: 0.04%, 0.223%, 29%, 6.71%, 14.86%,
0.23% and 22.8%, with various different models underlying these estimates.</p>
<p>For forecast aggregation, the <a href="https://forum.effectivealtruism.org/s/hjiBqAJNKhfJFq7kf/p/sMjcjnnpoAQCcedL2">geometric mean of odds is currently
SOTA</a>,
so we will be using that, which results in a chance of ~13%:</p>
<!--TODO: maybe load individual forecasts from Metaculus private data?-->
<pre><code>import numpy as np
>>> probs=np.array([0.002, 0.15, 0.13, 0.77, 0.85, 0.06, 0.23, 0.09, 0.17, 0.002]+1100*[0.18]+59*[0.12]+436*[0.05]+44*[0.02138]+[0.0004]+[0.00223]+[0.29]+[0.0671]+[0.1486]+[0.23]+[0.228])
>>> odds=probs/(1-probs)
>>> statistics.geometric_mean(odds)
0.1379626857389968
</code></pre>
<p>(If we remove the numbers from the 2012 LessWrong Survey, the aggregated
probability drops to ~5%.)</p>
<p>It would certainly be interesting to set up a prediction market for
this question, or get a team of superforecasters to estimate it,
but basically, it seems like for a young or middle-aged person, the
estimated probability is around 10%. However, the people surveyed are
often sympathetic to cryonics or even signed up. People are in general
overconfident, things generally don't happen, so being conservative by
halving the estimate seems like a good idea.</p>
<pre><code>prob_succ=0.05
</code></pre>
<p>This is quite pessimistic: It assumes that cryobiology will make no
progress <em>whatsoever</em> in the fidelity of the preserved tissue (remember,
the probabilities given are usually for the success of a preservation
given that it happens <em>now</em>).</p>
<h4 id="Other_Relevant_Questions"><a class="hanchor" href="#Other_Relevant_Questions">Other Relevant Questions</a></h4>
<ul>
<li><a href="https://www.metaculus.com/questions/1626/before-1-january-2050-will-any-human-cryonically-preserved-for-at-least-1-year-be-successfully-revived/">"Before 1 January 2050, will any human cryonically preserved for at least 1 year be successfully revived?"</a> has a median probability of 2% (n=260).</li>
<li><a href="https://www.metaculus.com/questions/3330/will-any-person-that-has-been-cryopreserved-for-more-than-1-year-be-resuscitated-or-emulated-before-2200/">"Will any person that has been cryopreserved for more than 1 year be resuscitated or emulated before 2200?"</a>: 57% (n=63).</li>
<li><a href="https://www.metaculus.com/questions/3382/what-will-the-earliest-preservation-date-of-any-resuscitated-cryonics-patient-be/">"What will the earliest preservation date of any resuscitated cryonics patient be?"</a> with a community prediction of ~2060 (n=49).</li>
</ul>
<h3 id="Years_Gained"><a class="hanchor" href="#Years_Gained">Years Gained</a></h3>
<p>Conditional on being revived, what is the average life expectancy?</p>
<p>If revival is achieved, it seems highly likely that aging and most
degenerative diseases have been eradicated (it makes little sense to
revive a person that is going to die again in 10 years). Also, most
revival scenarios hinge upon either the feasibility of very advanced
nanotechnology, which seems to be highly conducive to fixing aging,
or on whole brain emulation scenarios, which would likely make aging
unnecessary (why degrade a digital brain?).</p>
<p>If revival happens, there are still risks from accidents and homicide or
suicide that can kill the resuscitated cryonicist, as well as existential
risks that face all of humanity.</p>
<p>The website Polstats <a href="http://polstats.com/#!/life">illustrates</a>
the risks purely from accidents and homicides
using data from the <a href="http://www.iii.org/fact-statistic/mortality-risk">Information Insurance
Institute</a>. They
arrive at "a much more impressive 8,938 years" average life
expectancy. An <a href="https://math.stackexchange.com/a/3044240">answer</a>
on Mathematics StackExchange to the question
<a href="https://math.stackexchange.com/questions/3044194/whats-the-average-life-expectancy-if-only-dying-from-accidents">"What's the average life expectancy if only dying from
accidents?"</a>
arrives at 2850 years.</p>
<p><a href="https://forum.effectivealtruism.org/s/8QHbqGpXnzcQxiAis/p/jYMdWskbrTWFXG6dH#Life_expectancy_after_LEV_">Ascani
2019</a>
conservatively estimates a life expectancy of ~1000 years for each
individual human after LEV.</p>
<p>The life
expectancy of cryonics is thus is (just taking the average of these three values)
<code>$\frac{8938+2850+1000}{3} \approx 4260$</code>.</p>
<p>My rough guess is that resuscitated cryonicists will on average live ~4200
years, with significant uncertainty and leverage, if the cryonicist is
particularly careful to avoid accidents and (<em>checks notes</em>) suicide.</p>
<pre><code>years_gain=4260
</code></pre>
<hr/>
<p>That number should be qualified further in an <a href="https://en.wikipedia.org/wiki/The_Age_of_Em">"Age of
Em"</a> scenario: that
scenario will contain less natural risks (emulation can be backed up,
they live in a simulated world where homicide risks and car accidents
make no sense), but an em also suffers from the risk of not having
enough money to continue being run, and from the fact that the
em era might not last several subjective millennia. Furthermore,
it might be that multiple copies of the emulated cryonicist are
executed. This scenario deserves further consideration (see also
<a href="http://mason.gmu.edu/~rhanson/uploads.html" title="If Uploads Come First">Hanson 1994</a>).
I will not take into account the possibility of multiple
copies of the same person, and assume that only one emulation is being
run (to avoid tricky problems in aggregation).</p>
<!--
DOI: 10.2307/1973599
-->
<h4 id="Existential_Risk"><a class="hanchor" href="#Existential_Risk">Existential Risk</a></h4>
<p>Existential risk affects three different variables in the cost-benefit analysis:</p>
<ol>
<li>The amount of time spent alive (in expectation) before dying</li>
<li>The probability of being resuscitated (because an existential catastrophe happening while in preservation very likely causes <a href="https://en.wikipedia.org/wiki/Information-theoretic_death">information-theoretic death</a>)</li>
<li>The amount of time spent alive after resuscitation</li>
</ol>
<blockquote>
<p>An upper bound on natural existential risk to humanity
<code>$μ_{\text{nature}}$</code> based on the age of the species has been estimated
as <code>$<7.1429×10^{-5}$</code> per year, and based on the survival of the <em>Homo</em>
lineage below <code>$1/1.1494×10^{-6}$</code> [2697][…]. For the sake of argument,
let us set <code>$μ_{\text{nature}}≈7×10^{-5}$</code> per year.<br/>
Anthropogenic existential risk <code>$μ_{\text{anthro}}$</code>is hard to estimate
[2472], and yet likely dominates the risk profile at present and in the
near future. Claims in the literature are on the order of <code>$10^{-3}$</code>
to <code>$10^{-2}$</code> per year, although they are perhaps mainly rhetorical
[1983]. If one naively extrapolates the Richardson war fatality
distribution to killing 100% of the world population it gives a rate of
<code>$3.8730×10^{-4}$</code> per year (using the tail fit from [1959]). Nuclear
war rates may be on the order of 0.01 per year and have perhaps 10%
existential risk of being existenital (<code>$=10^{-3}$</code> per year), but
arguments can certainly be made for lower numbers [REF]. Climate
change combined with an unexpectedly high climate sensitivity may
have 1% probability [2697]; if we spread this across two centuries of
industrialization that may give <code>$5×10^{-5}$</code> per year. Risks from
emerging technology may be inherently unknowable. Still, we can bound
<code>$μ_{\text{anthro}}$</code>is by noting that even great pessimists claiming
extinction within a century is virtually certain[…] produces a rate of
at most <code>$10^{-2}$</code> per year. A lower bound may be <code>$10^{-5}$</code> per year.</p>
</blockquote>
<p><em>—Anders Sandberg, “Grand Futures” p. 488-489, 2023</em></p>
<p>This, in turn, depends on the development of the probability
of existential risk over the next 10k years. Estimating
existential risk appears to be quite tricky: We can take average
extinction rates across many species, or across species in the
<a href="https://en.wikipedia.org/wiki/Homo">homo</a> (<code>$\approx 7 \times
10^{-5}$</code> per year), but that assumes that <em>homo sapiens</em> is a
typical species or a typical member of the genus homo, which is a bold
statement about the first species to single-handedly cause a <a href="https://en.wikipedia.org/wiki/Holocene_Mass_Extinction">mass
extinction</a>.</p>
<p>(Among other things.)</p>
<p>The probabilities, pathways, causes, risk factors, inhibitors,
badness, varieties and reference classes for existential risk have been
<a href="https://en.wikipedia.org/wiki/The_Precipice:_Existential_Risk_and_the_Future_of_Humanity">extensively</a>
<a href="https://en.wikipedia.org/wiki/What_We_Owe_the_Future">belabored</a>,
so I won't roll up that entire discussion again. Unfortunately, most
estimates of existential risk concern themselves only with this century,
and don't make statements about medium term (i.e., the next 10k years)
probabilities of extinction.</p>
<p>For the 21st century I'll take the number by <a href="https://en.wikipedia.org/wiki/The_Precipice:_Existential_Risk_and_the_Future_of_Humanity">Ord
2020</a>:
~16.5% (equivalent to ~0.25% per year).</p>
<p>For the time after that, I'll lowball the number at <code>$2 \cdot 10^{-5}$</code>
per year. I don't have a great good justification for this,
but I find space colonization fairly plausible, which
reduces x-risk, and I also find it plausible achieving
transformative AI could bring humanity <a href="https://musingsandroughdrafts.wordpress.com/2021/03/24/my-current-high-level-strategic-picture-of-the-world/">out of the "time of
perils"</a>.
See also the <a href="https://www.metaculus.com/questions/2568/ragnar%25C3%25B6k-seriesresults-so-far/">Ragnarök question
series</a>.<!--TODO:
check out and incorporate numbers here?--></p>
<p>So, we can write a simple function that calculates the total extinction risk before a year:</p>