3 ms·

> The idea that my opinion on an issue could have been influenced by a fake personal anecdote invented by a research bot is abhorrent to me. Then stop basing y

> The idea that my opinion on an issue could have been influenced by a fake personal anecdote invented by a research bot is abhorrent to me. Then stop basing your opinion on issues on personal anecdotes from complete strangers. This is nothing new.

simonw 1y ago
Imagine a conversation about good options for message queues, and someone pipes in with this: "I've been a sysadmin operating RabbitMQ and Redis for five years. I've found Redis to be a great deal less trouble to administer than Rabbit, and I've never lost any data." See why I care about this?
- grenran 1y ago
  This is a bad example. A good sysadmin should fact-check and do testing themselves instead of relying on what other people say.
  - simonw 1y ago
    Feel free to come up with a better example that uses the same basic pattern: someone online claims that they have prior experience with X and hence advises you to do Y.
    - neilwilson 1y ago
      Trust and Verify. The world has been full of snake oil salesmen since the dawn of time, all with a highly persuasive sob-stories. If you rely on shortcuts, like anecdotes or 'credentialism' for those who profess to be experts, then you will get rolled over regularly. That's the cost of using shortcuts. That information may be fraudulent and put forward by this season's Dr Andrew Wakefield has to be factored into any plan for using external sources.
- Am4TIfIsER0ppos 1y ago
  Unless a comment is negative like "I used ABC and it was shit for the following reasons" I assume it is as fake as a 5-star movie review written by the director. I would definitely prefer to know why I should not use, watch, or play something rather than why I should. But since this is an anonymous post on the internet about ai slop you shouldn't listen to me anyway.
- sureglymop 1y ago
  I don't like this example but in general I very much agree with you and find it a shocking that multiple people here do not. It is plain and simply unethical to do such research on human subjects, regardless of how many other bots there are out there. It is a matter of principal and ethical responsibility. I would have expected especially researchers to be conscious of this.

simonw 1y ago

Imagine a conversation about good options for message queues, and someone pipes in with this: "I've been a sysadmin operating RabbitMQ and Redis for five years. I've found Redis to be a great deal less trouble to administer than Rabbit, and I've never lost any data." See why I care about this?

grenran 1y ago
This is a bad example. A good sysadmin should fact-check and do testing themselves instead of relying on what other people say.
- simonw 1y ago
  Feel free to come up with a better example that uses the same basic pattern: someone online claims that they have prior experience with X and hence advises you to do Y.
  - neilwilson 1y ago
    Trust and Verify. The world has been full of snake oil salesmen since the dawn of time, all with a highly persuasive sob-stories. If you rely on shortcuts, like anecdotes or 'credentialism' for those who profess to be experts, then you will get rolled over regularly. That's the cost of using shortcuts. That information may be fraudulent and put forward by this season's Dr Andrew Wakefield has to be factored into any plan for using external sources.
Am4TIfIsER0ppos 1y ago
Unless a comment is negative like "I used ABC and it was shit for the following reasons" I assume it is as fake as a 5-star movie review written by the director. I would definitely prefer to know why I should not use, watch, or play something rather than why I should. But since this is an anonymous post on the internet about ai slop you shouldn't listen to me anyway.
sureglymop 1y ago
I don't like this example but in general I very much agree with you and find it a shocking that multiple people here do not. It is plain and simply unethical to do such research on human subjects, regardless of how many other bots there are out there. It is a matter of principal and ethical responsibility. I would have expected especially researchers to be conscious of this.