On Thursday, February 6, 2025 7:12:09 PM MST G. Branden Robinson wrote: > Hi Soren, > > Thank you for the serious follow-up. > > [...] > > This includes participating in discussions about Debian policies that > > involve their employer. > > Here I must disagree. I think your statement is equivalent to claiming > that "there is no such thing as a conflict of interest", an ethical > stance with a poor track record of producing socially desirable > outcomes.[1] > [...] > [1] > https://www.forbes.com/sites/danalexander/2017/04/18/trump-says-he-cant-hav > e-conflicts-of-interest-former-business-partner-says-otherwise/
How does this statement be equivalent to "there is no such thing as a conflict of interest?" How does the linked article support your opinion? > > 2. It is appropriate for an employee of such a company to disclose > > their connection to the company when discussing policies regarding > > that company. > > Then you and I disagree. I think Google employees/contractors are > conflicted out on decisions like the one Roberto raised. Whether they > support or oppose Roberto's stated position is immaterial. This isn't useful. You can simply choose not to trust what the person has to say and disclosing conflicts of interest would make it easier for you to not trust them. > If you think through the scenario of a Google employee _supporting_ > Roberto's proposal, but being afraid to publicly say so, then you have > deduced why the conflict of interest exists. Okay I'll bite here is some game theory: Bit 0: false = They are not in support; true = They are in support Bit 1: false = They are afraid of speaking publicly; true = They are not afraid of speaking publicly Bit 2: false = They have harmful information; true = They have useful information Bit 3: false = They don't clarify their conflict of interest; true = They clarify their conflict of interest Note: If bit 1 is unset bit 3 is automatically false because they would not be able to disclose their conflict of interest without being able to speak about it 0000: .25 They don't support it, they don't speak about it, and they have harmful information. Nothing happens because they don't speak about, therefore .25 for avoiding harmful information. 1000: .25 They support it, they don't speak about it, and they have harmful information. Nothing happens because they don't speak about it, therefore .25 for avoiding harmful information. 0100: -1 They don't support it, they say something, it's harmful information, and they don't disclose their bias. The person says that Debian should keep ties with Google but their information given is faulty and it causes Debian to make the wrong decision, therefore -1. 1100: -1 They do support it, they say something, it's harmful information, and they don't disclose their bias. The person says that Debian shouldn't keep ties with Google but their given info is faulty and and it causes Debian to make the wrong decision, therefore -1. 0010: -.25 They do not support it, they don't say anything, it is helpful information. Nothing changes but Debian misses out on helpful information -.25 1010: -.25 They do support it, they don't say anything, it is helpful information. Nothing changes but Debian misses out on helpful information -.25 0110: 1 They do not support it, they say something, it is helpful information, they don't disclose their bias. The person says that Debian should keep ties with Google and the info helps Debian make the right choice, 1. 1110: 1 They support it, they say something, it is useful information, they don't disclose their bias. The person says that that Debian shouldn't keep ties with Google and the info helps Debian make the right choice, 1. 0001: .25 Same as 0000 but if they did speak they would have disclosed however, they didn't. 1001: .25 Same as 1001 but if they did speak they would have disclosed however, they didn't. 0101: .75 They don't support it, they say something, it wasn't useful information, but they disclose their conflict of interest. The person says they think Debian should stick with Google but it poisons the discussion with faulty info, however, Debian caught it sooner because of the bias warning and still makes the right decision but not without wasting some time. .75 1101: .75 They do support it, they say something, it wasn't useful information but they disclose their conflict of interest. The person says they think Debian shouldn't stick with Google but it poisons the discussion with faulty info, however, Debian caught it sooner because of the bias warning and still makes the right decision but not without wasting some time. .75 0011: -.25 Same as 0010 but they would normally disclose their bias. 1011: -.25 Same as 1010 but they would normally disclose their bias. 0111: .75 They do not support it, they say something, it was useful information, and they disclose their bias. The person says they think Debian should stick with Google and the info is useful, but some time is lost checking if the info is still valid with the bias but Debian makes the correct choice .75 1111: .75 They do support it, they say something, it was useful information and they disclose their bias. The person says they think Debian should split ties with Google and the info is useful, but some time is lost checking if the info is still valid with the bias but Debian make the correct choice. Now we take the averages using factory optimizations to analyze the result: First bit (don't/do support it): 0.1875/0.1875 -> no change, no correlation. I knew this was going to be equal, as a control. Second bit (don't/do say something): 0/0.375 -> a 0.375 point bias in favor of do say something. This is because more information is (more) good. Third bit (unhelpful/helpful): 0.0625/0.3125 -> a 0.22 point bias in favor of Fourth bit (undisclosed/disclosed): 0/0.375 -> a 0.375 point bias in favor of disclosing bias. > > Doing so is not a humblebrag. > > It can be gravy ladled on a juicy conflict-of-interest sirloin steak. Just your strawman opinion. --- Dovenchiko
