Checking Kurzweil’s track record

Pre­dic­tions are cheap and easy; ver­ifi­ca­tion is hard, es­sen­tial, and rare. For things like AI, we seem to be re­stricted to noth­ing but ex­pert pre­dic­tions—but ex­pert pre­dic­tions on AI are not very good, ei­ther in the­ory or in prac­tice. If we are some ex­perts who stand out, we would re­ally want to iden­tify them—and there is noth­ing bet­ter than a track record for iden­ti­fy­ing true ex­perts.

So we’re ask­ing for help to ver­ify the pre­dic­tions of one of the most promi­nent fu­tur­ists of this cen­tury: Ray Kurzweil, from his book “The Age of Spiritual Machines”. By ex­am­in­ing his pre­dic­tions for times that have already come and gone, we’ll be able to more ap­pro­pri­ately weight his pre­dic­tions for times still to come. By tak­ing part, by lend­ing your time to this, you will be di­rectly helping us un­der­stand and pre­dict the fu­ture, and will get show­ered in grat­i­tude and ku­dos and maybe even karma.

I’ve already made an at­tempt at this (if you are in­ter­ested in tak­ing part in this pro­ject, avoid click­ing on that link for now!). But you can­not trust a sin­gle per­son’s opinions, and that was from a small (albeit ran­dom) sam­ple of the pre­dic­tions. For this pro­ject, I’ve tran­scribed his pre­dic­tions into 172 sep­a­rate (short) state­ments, and any vol­un­teers would be pre­sented with a ran­dom se­lec­tion among these. The vol­un­teers would then do some Google re­search (or other) to es­tab­lish whether the pre­dic­tion had come to pass, and then in­di­cate their ver­dict. More de­tails on what ex­actly will be mea­sured, and how to in­ter­pret am­bigu­ous state­ments, will be given to the vol­un­teers once the pro­ject starts.

If you are in­ter­ested, please let me know at stu­art.arm­strong@philos­o­phy.ox.ac.uk (or in the com­ment thread here), in­di­cat­ing how many of the 172 ques­tions you would like to at­tempt. The ex­er­cise will prob­a­bly hap­pen in late Novem­ber or early De­cem­ber.

This will be done un­blinded, be­cause Kurzweil’s pre­dic­tions are so well known that it would be in­fea­si­ble to find large num­bers of peo­ple who are tech­nolog­i­cally aware but ig­no­rant of them. Please avoid shar­ing your ver­dicts with oth­ers; it is en­tirely your own in­di­vi­d­ual as­sess­ment that we are in­ter­ested in hav­ing.