4 ms·

I have a bold vision for the future. Using consensus algorithms, humans are able to build a bucket of knowledge verified by web-of-trust, free of interest corr

I have a bold vision for the future. Using consensus algorithms, humans are able to build a bucket of knowledge verified by web-of-trust, free of interest corruption, and make society-wide decisions in matters of months and weeks rather than decades and years. We can go from "scientists discover X" to "scientists I personally know and trust have reviewed the paper and I now accept it as truth" to trustworthy consensus on the matter to action items with just consensus algorithms and humans voting on blocks of text.

Invictus0 6y ago
The idea that who a scientist is and what they stand for has any bearing on their work is a an affront to science itself. Let the papers speak for themselves; the authors need not even be mentioned.
- gpm 6y ago
  This just isn't how science work. Papers don't stand on their own because it takes far too long for even a subject matter expert to evaluate one paper, and that's even assuming that the paper was written honestly and without fudging the results. Science works on reputation. That reputation is most evident in the form of peer review where trusted members of the community review papers and decide which ones are good enough to get published. It also works behind the scenes in people finding out what papers to read by word of mouth, and people setting there prior on how trustworthy it is by who the authors are. Reputation is built partially (largely even) on being able to reliably publish true and verifiable results, and doing this without ever playing any politics will quickly build a reputation (meanwhile building a reputation by playing politics is next to impossible), but it's still reputation based.
  - waterhouse 6y ago
    You could rely on the reputation of the reviewers, while concealing the identity of the authors from everyone. That may be what Invictus0 was suggesting, and it does sound workable. It raises one question—"How do we get the first reputable scientist to bother reviewing Mr. Anon's paper?" But as long as there wasn't a huge problem of outright spam, or of crackpot submissions that eat up too much time... or if serious submitters had to offer a monetary bounty (on the order of $100-$500) for the first or the first few reviewers, or if the bounty was more like a "security deposit" which would be given back if the paper was confirmed to be neither spam nor crackpot... There seem to be multiple plausible options.
  - chrisco255 6y ago
    Building a reputation by playing politics is next to impossible? What experience do you have in academia? There's plenty of politics in academia.
    - gpm 6y ago
      There's plenty of politics, but you still need results... politics alone is not sufficient.
      - chrisco255 6y ago
        No, you don't: https://en.m.wikipedia.org/wiki/Replication_crisis https://en.m.wikipedia.org/wiki/Replication_crisis
juskrey 6y ago
Science does not work by consensus, simply because single evidence invalidates any consensus with arbitrary number of holders of other opinion, however anonymous. And if it does, it starts to resemble middle ages inquisition committees more than science.
- vbezhenar 6y ago
  Illusionist will conjure you the evidence of rabbit teleportation. Andrea Rossi will show you cold fusion. It does not mean anything. You need experts to verify that evidence. In modern world verifying evidence requires years of education and might require millions of dollars of equipment. So in the end you still need someone you trust to verify or decline that evidence.
dzink 6y ago
I’m thinking more of a consequence engine where you can model and compute second and third order consequences of proposed laws based on economics, past outcomes, and other insights. That said, truth changes and evolves, so consensus while correct today, may be very slow to show you the correct tomorrow. Also if you do have bold visions, do share a way to reach you when others agree.

Invictus0 6y ago

The idea that who a scientist is and what they stand for has any bearing on their work is a an affront to science itself. Let the papers speak for themselves; the authors need not even be mentioned.

gpm 6y ago
This just isn't how science work. Papers don't stand on their own because it takes far too long for even a subject matter expert to evaluate one paper, and that's even assuming that the paper was written honestly and without fudging the results. Science works on reputation. That reputation is most evident in the form of peer review where trusted members of the community review papers and decide which ones are good enough to get published. It also works behind the scenes in people finding out what papers to read by word of mouth, and people setting there prior on how trustworthy it is by who the authors are. Reputation is built partially (largely even) on being able to reliably publish true and verifiable results, and doing this without ever playing any politics will quickly build a reputation (meanwhile building a reputation by playing politics is next to impossible), but it's still reputation based.
- waterhouse 6y ago
  You could rely on the reputation of the reviewers, while concealing the identity of the authors from everyone. That may be what Invictus0 was suggesting, and it does sound workable. It raises one question—"How do we get the first reputable scientist to bother reviewing Mr. Anon's paper?" But as long as there wasn't a huge problem of outright spam, or of crackpot submissions that eat up too much time... or if serious submitters had to offer a monetary bounty (on the order of $100-$500) for the first or the first few reviewers, or if the bounty was more like a "security deposit" which would be given back if the paper was confirmed to be neither spam nor crackpot... There seem to be multiple plausible options.
- chrisco255 6y ago
  Building a reputation by playing politics is next to impossible? What experience do you have in academia? There's plenty of politics in academia.
  - gpm 6y ago
    There's plenty of politics, but you still need results... politics alone is not sufficient.
    - chrisco255 6y ago
      No, you don't: https://en.m.wikipedia.org/wiki/Replication_crisis https://en.m.wikipedia.org/wiki/Replication_crisis

juskrey 6y ago

Science does not work by consensus, simply because single evidence invalidates any consensus with arbitrary number of holders of other opinion, however anonymous. And if it does, it starts to resemble middle ages inquisition committees more than science.

vbezhenar 6y ago
Illusionist will conjure you the evidence of rabbit teleportation. Andrea Rossi will show you cold fusion. It does not mean anything. You need experts to verify that evidence. In modern world verifying evidence requires years of education and might require millions of dollars of equipment. So in the end you still need someone you trust to verify or decline that evidence.

dzink 6y ago

I’m thinking more of a consequence engine where you can model and compute second and third order consequences of proposed laws based on economics, past outcomes, and other insights. That said, truth changes and evolves, so consensus while correct today, may be very slow to show you the correct tomorrow. Also if you do have bold visions, do share a way to reach you when others agree.