SilasBarta comments on Help Fund Lukeprog at SIAI - Less Wrong
You are viewing a comment permalink. View the original post to see all comments and the full post content.
You are viewing a comment permalink. View the original post to see all comments and the full post content.
Comments (276)
I agree with 'more testing and evidence, please', but you often come across as adversarial and I think that generally makes it harder for you to convince the people you want to convince.
As an aside, remember that the minicamp was a relatively unplanned event; it came about because SIAI had extra time and space. I will be more concerned if the megacamp has a similar lack of testing.
Well, I hope they're not relying on "Silas is a meanie" as their intellectual "covering fire" for not substantiating this claim. And it's not that I want more testing and evidence, I just want to see what they think proves its success.
True, but I wouldn't be asking for any of this if leaders didn't try to paint it afterwards as a major success. If they want to take a risky venture, fine. If they want to play, "I meant to do that", let's see what it accomplished.
I'm not talking about 'covering fire'. If your goal is to win an argument or appear righteous, then your strategy is alright. If your goal is to actually get SIAI to change their behavior, then your language is hurting your cause. You want to make it as easy as possible for them to change their behavior, and it's psychologically much easier to do something because an ally asks than because an adversary asks.
You have seen evidence: both Guy (link) and I (link) posted 'lessons learned' for the minicamp. You are right to say this is not especially strong evidence, but it is evidence. I think it would have been good to video tape some of the sessions and post them and post the exit surveys (they took testimonials too).
No, it clearly isn't. He left himself wide open to this sort of attack.
More fundamentally he justified and explained himself. He did it reasonably well and a good justification can work but it is almost never the optimal strategy.