Do the people attacking Anduril get mistreated here? (or even by the SC?)
5 Likes
I still don’t even know why people are talking about Anduril and politics in general in a software community. Instead of identifying as being pro/anti politics we should identify as being pro/anti features.
Why are we assuming that Anduril or any other company wants to do harm to Nix? Shouldn’t we focus on what people want to contribute and discussing if that’s a good or bad thing to have in the software, regardless of where it’s coming from?
As I’ve stated in another post lately, where showing support for ukraine was the topic (which was sadly banned), I’m all for keeping politics out of the community as good as possible. As someone who refused to do military service, or to even touch a gun, I’m still interested in every possible contribution to Nix. As long as the community has the power to discuss & express their feelings about those, I tend to welcome absolutely anyone that’s treating people respectfully.
14 Likes
You’d think so, you’d really think so…but, that line of questioning leads us to an uncomfortable moment of reflection around how DetSys and Anduril have been treated here. We have not been a good community for collaboration, full stop.
At my most charitable, it’s more important for some folks to get their activism fix (observation: has Anduril stopped or even slowed down in the face of all this toxic bluster? has DetSys? If this is some grand plan to fix capitalism and kneecap the MIC, it don’t appear to be working!) than to support or engage productively (or again, even civilly) with companies and developers doing genuinely useful and helpful work for the community.
Is this a good message to send to potential sponsors and supporters? No. Is this a good signal about how we treat people who volunteer for things like the SC? No. Is this a good message to send to potential trolls about decorum? No.
Last year, we had so much drama and argument and even a big fork–and even that wasn’t enough, because a lot of these people stayed here to raise hell, brigade on social media, and dox people when they didn’t get their way (and why the mods allowed this, and why the SC didn’t put the hammer to the mods, we’ll never know).
Edit: removed a “don’t delete this” bit.
17 Likes
This was an interesting analysis. Thank you for taking the time to write it. I do have some thoughts, which I have separated into parts. Sorry for the length, and I’m sure I made typos and editing errors so point them out if they are confusing and I can clarify.
A. I do not believe the contents of the first letter can be used to support any conclusion about the community being more generally “against MIC”
The specific points of disagreement I had in my first reply to you were 1) framing that important voice as the almost unanimous voice of the community and thus disregarding or erasing those who do not hold the stated view (addressed in your reply) and 2) misrepresenting the specific statement made by the signatories.
Thank you for reducing your previous statement of “99% of the community” to “one in four maintainers”. However, I must reiterate: The letter states that signatories only affirm that they “do not want to see the NixOS community become [a] vehicle for advertising the Military Industrial Complex”, and “does not mean you are personally against the Military Industrial Complex”. This is not the same as “against MIC” and I would not reduce it as such.
Since you have opened the door to more mathematical notation, I can describe my point more formally with the hope that it is clearer.
Term definitions (please click to expand)
P(totalNixUsers) = 1 = entire Nix community
(including maintainers and non-maintainers,
respondents and non-respondents
P(X) = probability of some event X, bounded [0,1]
P(X) = Count(X) / Count( totalNixUsers )
Count(X) = number of X, bounded [0,inf)
_respondedCommunitySurvey_ = Nix community who responded to the 2024 survey
_readLetter1_ = Nix community who read the first letter
_readLetter2_ = Nix community who read the second letter
_signatory1_ = Nix community who signed the first letter
_signatory2_ = Nix community who signed the second letter
_maintainer_ = Nix community who are maintainers
_agrees1_ = Nix community that agrees with first open letter
_agrees2_ = Nix community that agrees with second open letter
_againstMIC_ = Nix community that is against MIC
P(A | B) = probability of A given B
P(A * B) = probability of A and B
P(A + B) = probability of A or B
Count( maintainer * signatory1 ) = 92
Count( signatory1 ) = 232
Count( maintainer * signatory2 ) = 1
Count( signatory2 ) = 2
Count( maintainer * respondedCommunitySurvey ) = 330
Count( respondedCommunitySurvey ) = 2260
Count( maintainer ) = 4619
Your formal claim might be translated as:
P( againstMIC | (maintainer * readLetter1) ) ≈ 1/4
Plain english: within the population of maintainers who have also
read the first open letter, about 1 in 4 are against MIC
However, even before discussing specific elements of the exact math, your specific conclusion relies logically on signatory1 and/or agrees1 implying againstMIC, which I again believe is directly contradicted by the contents of open letter 1.
B. An open source letter in the NixOS community which can gather 232 signatories is noteworthy
Having said that, I do believe that in my original post I should have stated more explicitly: 232 signatories to the first open letter does represent a substantial voice and it is reasonable to consider that voice in the design and execution of policy.
It is a useful data point that the first open letter did attract more signatories than did the second letter. I do disagree with a framework in which these two letters seem to represent an axis of “for MIC” vs. “against MIC”; I do not believe the two letters are fundamentally complementary (in the mathematical sense of any statement consistent with the views endorsed in the first letter must therefore be inconsistent with the views expressed in the second letter, and vice versa). My own beliefs around MIC are more complicated than can be represented through a binary variable. If they could be so reduced, I further do not believe that these two letters are a good proxy for that binary variable.
C. Disagreements with regards to math and data analysis/interpretation
The math is fun to explore, and while I may disagree on some of the specific choices in your analysis, I think focusing too much on the exact number distracts from Part A and Part B, and we don't have to agree on the math to agree on the importance of engaging in discussion and considering different opinions (Click if you really like the math.)
This was not my statement. My statement was that the “open letter signatory process is poorly equipped to provide reliable data on proportionality”. In contrast, depending on study design, surveys are a major research tool that are frequently used for descriptive statistics including estimates of proportion.
I think you might have typo-ed? I believe you agree with me, since the proportion of survey respondents who are also maintainers is fundamental to your chain of inference leading to your conclusions regarding proportionality within the open letter signatory process. It would have been better for me to state that an open letter signatory process in isolation cannot capture proportionality, and your ideas do point out that these the letter signatories and the community survey do not have to be analyzed in isolation.
There are interesting results from applying your basic approach. For example, the straightforward application of Bayes in this case would suggest that the proportion of the total Nix population that signed the letter is less than 1%, although with some important caveats in the analysis. Such an analysis also does also not imply that 0.7% of the Nix community agrees with the view in open letter 1; that would be a very different statement. My calculation is closer to a very, very, very conservative lower bound on the percentage agreeing.
Term definitions repeated for convenience (click to expand)
P(totalNixUsers) = 1 = entire Nix community
(including maintainers and non-maintainers,
respondents and non-respondents
P(X) = probability of some event X, bounded [0,1]
P(X) = Count(X) / Count( totalNixUsers )
Count(X) = number of X, bounded [0,inf)
_respondedCommunitySurvey_ = Nix community who responded to the 2024 survey
_readLetter1_ = Nix community who read the first letter
_readLetter2_ = Nix community who read the second letter
_signatory1_ = Nix community who signed the first letter
_signatory2_ = Nix community who signed the second letter
_maintainer_ = Nix community who are maintainers
_agrees1_ = Nix community that agrees with first open letter
_agrees2_ = Nix community that agrees with second open letter
_againstMIC_ = Nix community that is against MIC
P(A | B) = probability of A given B
P(A * B) = probability of A and B
P(A + B) = probability of A or B
Count( maintainer * signatory1 ) = 92
Count( signatory1 ) = 232
Count( maintainer * signatory2 ) = 1
Count( signatory2 ) = 2
Count( maintainer * respondedCommunitySurvey ) = 330
Count( respondedCommunitySurvey ) = 2260
Count( maintainer ) = 4619
Worked example approximating signatories to letter 1 as <1% of Nix population (click to expand)
Manipulation of conditional probability will be used to solve for P( signatory1 ), an estimate of the proportion of the total Nix population that signed open letter 1.
Bayes theorem:
P(A | B) = P(B | A) * P(A) / P(B)
P( maintainer | signatory1 ) = P( signatory1 | maintainer ) * P( maintainer ) / P( signatory1 )
Definition of Conditional Probability:
P( A | B ) = P( A * B ) / P(B)
Count( maintainer * Signatory1 ) / Count( signatory ) = Count( signatory1 * maintainer ) / Count( maintainer ) * P ( maintainer ) / P( signatory1 )
Assume Count( maintainer ) === the count from maintainers.nix
92 / 232 = 92 / 4619 * P( maintainer ) / P( signatory1 )
Rearrange:
P( signatory1 ) = 232 / 4619 * P( maintainer )
Approximate P( maintainer ) === P( maintainer | respondedCommunitySurvey ) = 330 / 2290, or in other words approximate that P( maintainer ) and P( respondedCommunitySurvey ) are independent variables.
P ( signatory1 ) = 232 / 4619 * 330 / 2290
P( signatory1 ) = 0.0072
Explicitly:
Count( signatory1 ) / Count( totalNixUsers ) = 0.0072
A mathematically simpler derivation can be had via:
Assume that the ratio of maintainers:users is roughly the same within the community survey
and the overall population. If maintainers are are over-represented in the survey, then
this calculation will underestimate the total population size; if maintainers are under-
represented, then the calculation will overestimate the total population size.
Count( maintainers-list ) / Count( totalNixUsers ) = Count( maintainer * respondedCommunitySurvey ) / Count( respondedCommunitySurvey )
Count( maintainers-list ) / Count( totalNixUsers ) = P( maintainer | respondedCommunitySurvey )
Count( totalNixUsers ) = Count( maintainer-list ) / P( maintainer | respondedCommunitySurvey )
Count( totalNixUsers ) = 4619 / (330/2290)
≈ 32053
P( signatory1 ) = Count( signatory1 ) / Count( totalNixUsers)
= 232 / 32053
= 0.0072
Overestimating totalNixUsers would reuduce P( signatory1 ), underestimating would increase.
The inference is that ~0.7% of the Nix community in total signed the first open letter. This analysis is gated on the number of maintainers in maintainer-list.nix being a decent count (yeah I know, the likely overestimate probably drags the calculated P down), and on the proportion of maintainers who responded to the Community Survey being a decent proxy for the overall percentage of maintainers within the Nix Community (yeah, maintainers are probably more likely to respond, thus dragging the calculated P up). These are both not great estimates, so this analysis should be taken with heavy caveats. At the least, it is nice that the likely bias of these two estimates are more likely to cancel rather than aggregate.
You came to the conclusion that P( against MIC | maintainer ) ≈ 1/4. I have already said why I don’t believe the letters allow you to make conclusions about this term, but rather only about P( agrees1 | maintainer ). I also have concerns about how reliable we can make our estimate of this number, as in my analysis the conclusion is extremely sensitive to response rate.
I don't think this is easily estimated (expand for details)
The basic goal is to identify
P( agrees1 | maintainer )
- which is not the same as P( againstMIC | maintainer )
- difficult to answer directly
Lacking a way to directly measure P( agrees1 ), we might try to approximate it by
P( agrees1 | maintainer ) ≈ P( signatory1 | (maintainer * readLetter1) )
By Bayes Law, and then
§ By logical deduction, P( (maintainer * readLetter1) | signatory1 ) should be equal to P( maintainer | signatory1 ), unless maintainers were signing without reading.
P( signatory1 | (maintainer * readLetter1) )
= P( (maintainer * readLetter1) | signatory1 )
* P( signatory1 ) / P( maintainer * readLetter1 )
= P( maintainer | signatory1) * P( signatory1 ) / P( maintainer * readLetter1 )
We already derived P( signatory1 ) in previous worked example
= (92 / 232) * ( 232 / 32053 ) / P( maintainer * readLetter1 )
= (92 / 32053) / P( maintainer * readLetter1 )
It’s actually quite difficult to formulate a reasonable proxy for
P( maintainer * readLetter1 ).
Interestingly, the smaller the number we approximate for “the number of maintainers who read open letter 1”, the greater the resulting estimation of the importance of each signature. So everyone who is arguing about ignoring the drama causing a reduction of eyes on the letters is theoretically increasing the value of the signatures in the resulting inference to how many maintainers would agree with letter 1.
In your analysis, you essentially say
P( maintainer * readLetter1 ) ≈ P( maintainer * respondedCommunitySurvey )
= 330 / 32053
[ your analysis ]
= (92 / 32053) / (330 / 32053 )
= (92 / 330)
= 0.279
≈ 28%
Unfortunately, this is extremely reliant on a stable response rate of maintainers in two very different conditions. I think 28% feels like a “good” number and that makes it a very attractive estimate. However, it would be mathematically possible that an open letter in the NixOS community might attract 750 signatures, of which 350 are maintainer signatures. In that case, it would not be reasonable to conclude that P( maintainer | signatory1) * P( signatory1 ) / P( maintainer * respondedCmmunitySurvey ) = (350 / 750) * (750 / 32053) / (330 / 32053) = 350 / 330 =106% of maintainers would agree with the position in that open letter, as probability must be mathematically bounded within 0-100%. That using P( maintainer * respondedCommunitySurvey ) as a fixed approximation cannot make this guarantee on bounds [0-100%] is a sign that it is not mathematically appropriate to assume this substitution. We might be able to overlook this if the overall analysis is not very sensitive to the estimate of P( maintainer * readLetter1 ). Unfortunately, the analysis is sensitive to this.
With 92 maintainer signatories to letter 1, the lower bound of P( maintainer * readLetter1 ) is 92 / 32053 (maintainers who read the letter decided to sign, but not many read the letter) and the upper bound is 4716 / 32053 (all maintainers read the letter, only 92 signed). From this, we see that the bounds of our estimate for P( agrees1 | maintainer ) ≈ P( signatory1 | (maintainer * readLetter1) ) are
# Bounds for indirect estimate of P( agrees1 | maintainer )
Lower Bound (92 / 32053) / (4716 / 32053) = 0.00195 = 2%
Upper bound (92 / 32053) / (92 / 32053) = 1 = 100%
We can also model different values for the number of maintainers having read the letter:
# How different numbers of maintains reading the first open letter affects my
# indirect estimate of P( agrees1 | maintainer )
Count( maintainer * readLetter1 ) = 100, then (92 / 32053) / (100 / 32053) = 92%
Count( maintainer * readLetter1 ) = 250, then (92 / 32053) / (250 / 32053) = 36.8%
Count( maintainer * readLetter1 ) = 500, then (92 / 32053) / (500 / 32053) = 18.4%
Count( maintainer * readLetter1 ) = 1000, then (92 / 32053) / (1000 / 32053) = 9.2%
Count( maintainer * readLetter1 ) = 1500, then (92 / 32053) / (1500 / 32053) = 6.13%
You can see that the value swings wildly depending on how many maintainers you think read the letter. I don’t necessary think that your estimate of 330 maintainers is terrible, but when the analysis is this sensitive to the choice, I think it’s an important limitation to how much confidence we can place in the resultant figure of 28% that you get. This contributes to my statement on how an open letter signatory process is poorly equipped to answer questions on proportionality. It can at best provide you definitively a lower bound (in this case 2%).
Again, I think it’s fascinating how people who argue that few people paid attention seem to believe that it weakens the significance of N=232 signatories. I also thinks it’s fascinating how people are arguing the vice versa. It’s not immediately intuitive, but with a fixed number of signatories (which we can directly observe), a lower estimate for the amount of reading/engagement with the letters would mean that proportionally a higher percentage of those who read chose to sign.
D. Lets refocus on a discussion about how to move forward with better policies
This is a pretty charged thread and it’s evident that many here have strong feelings on the matter. Feelings are valid, and I’m sure there is a lot of blame to be passed around. As I said earlier, I wish I had explicitly acknowledged how I do believe 232 signatories does represent a significant voice. Conciseness and clarity are valuable traits, but I do think that in the context of this thread it would have been beneficial to take that particular step. We disagree on how easily we can use this data point to make conclusions about the broader Nix community, and that’s largely OK. Even if I don’t agree that 1/4 is a very reliable figure, I can easily agree with your statement “In general, a subculture of about 1-2% is a considerable minority; they are worth considering and mentioning in relevant topics.”
I particularly like one of the upthread posts and think it rather cuts to the heart of the matter while getting lost in the noise, so I’m bringing it back here in the hopes of redirecting conversation.
As it currently stands, failure to disclose the new employment status to the community at large is not in violation of the NixOS constitution; ElvishJerricco points out that the current constitution as written doesn’t seem to cover this case well, and that NixOS would benefit crafting policy changes that are aligned with a set of consistent values. I expect that this thread will galvanize some changes, but especially if you want it those policy changes to be amended in the constitutional documents it would probably take some time to enact the procedures necessary to craft a strong and clear policy. From the announcements and comments from SC members upthread, it does seem that there is action being taken to at least create a living Conflict of Interest page, and I think that’s a great first step that would be aligned with “option 2”.
6 Likes
As noted, in this case new elections are coming up soon. In future events, the suggestion to update CoI disclosure may improve governance when those are not imminent.
Even if such a measure were implemented though, this additional transparency would not be particularly actionable yet. Without means to retract one’s vote, there would remain no formal means of recourse (until new elections) in the event a future candidate were to lose voters’ trust, presuming remaining SC members were not to vote them out.
4 Likes
We’ve added a team page to the website, which includes CoI info.
(This has been live for >12 hours, I just could not post about it due to slow mode.)
25 Likes
Is the fact that only 2/6 of the SC is currently the remainder of the term they were elected for a mistake? Because only John and Roberth is staying until 2026, until their term expires according to this page.
Not an official announcement yet but Gabby and I will be resigning.
6 Likes
This particular situation happens to exclude the middle, particularly given that the original situation under contention also excluded the middle. If you want to sign the pro-MIC letter, then you can; it’s still open to edits. A lack of signature doesn’t mean a lack of commitment in our existing pro-MIC world, but an acknowledgement of status quo.
I do find it quite interesting that you are willing to lay so much snow on the topic. I’d hope that you’d notice that the maintainers are a little more focused in their approaches than that!
Thank you for doing this. I hope that SC continues to increase their transparency.
2 Likes
Perhaps I am unclear on what you are referring to by “this particular situation”. In my reading of the text, the letter seems to very explicitly make room for a middle. Signatories to that letter affirm that “they do not want to see the NixOS community become [a] vehicle for advertising the Military Industrial complex”, but that “does not mean [they] are personally against the Military Industrial Complex”.
If you would like to make an argument that some count of maintainers are specifically against MIC, you are equally open to create a new open letter from the Nix community which asks signatories for more specific affirmation that they are against MIC. If we want to talk more strongly about proportion specifically, it would strengthen the evidence if a survey or other research instrument was used instead of relying on indirect inference.
If I have misread you, I apologize. However:
I think you might be bringing a long-standing context, history and anger into this discussion that I simply don’t have, and it is making your responses unkind. I don’t believe that I am your enemy, even if I disagree with some of the things you say. I thought the new NixOS mediation service that was discussed at this years NixCon was an intriguing idea. If you think it would be helpful, I am open to participate. Just PM me here and I can look up the details on how to get the ball rolling.
This is interesting. I’m not aware of a voting system that includes mechanisms to retroactively retract votes after an election is called, but 1) I haven’t made a study of voting systems and 2) that doesn’t mean NixOS governance couldn’t experiment with such a system. I’m not really sure how such a thing would work out, as the mid-cycle winners created by retroactive vote modification might no longer have time for SC duties. In that case, would the idea be to go down the list of runner-ups until you reach someone who still has the time to commit to SC duties?
Perhaps a recall or special election mechanism might be easier to implement on? The constitution as written seems to have rules for Special Elections, Full Reelections, and Removal for Conduct (which may be used to justify Special Elections. However, as I understand these mechanisms are currently structured as an indirect/representative democracy – for action to be taken in response to “a future candidate… losi[ng] the voters’ trust”, the current constitution implicitly assumes that the SC as a body can and/or will act as a representative proxy for the voters’ wishes.
Is the idea that perhaps the constitution could be amended to have a directly democratic mechanism for either Full Reelection or Removal for Conduct?
2 Likes
I don’t think we do differ. You want diversity with a common goal. You can only have a diversity of view points if you allow for disagreements. How can you allow for disagreements? By narrowing what you feel must be agreed on, because opposite obviously doesn’t hold: forcing agreement on a wider selection of topics must by necessity reduce the diversity of viewpoints acceptable in the community.
1 Like
while recall elections were the main viable option in the analog scenario, compared to the digital alternative of amending votes they might bring along more logistics in terms of gauging when such a recall voting round might be warranted (edit: on top of making the electorate reiterate their stance in such a recall round).
that seems like one option - given candidates with both sufficient mandate as well as sufficient time remain available.
another might be to settle for simply un-electing from the representative body candidates that lose the required share of votes.
either way, at some point a new election round (allowing new candidates to participate) may be warranted (edit: e.g. either once the term expires or once the number of representatives drops below some threshold, whichever comes earliest).
while popular digital voting systems tend to finalize results, one area that seems more likely to have implemented such mechanisms may be liquid democracy ones, tho i’m not really finding a lot of popular implementations out there. to be fair, i think the main reason the traditional digital voting systems would finalize results was their decryption process involving trustees (as used in e.g. Helios).
2 Likes
Is there any reason why? The last 2 members of the SC being from the Nix team doesn’t really paint a pretty picture.
2 Likes
Being employed in a high level position at a MIC enterprise is not compatible with Nix’s stated values: People come first.
I feel it’s problematic that Tomberek didn’t disclose what his employer and thus his employment as a full-time “Principal Software Engineer” in Washington DC entails. I’ll do it since I feel this is more important to mention than “nix adoption.”
- “Trump’s Big Beautiful Gift to Anduril” The Intercept July 9th, 2025 https://archive.is/e2NcN
Anduril is now the country’s only approved border tower vendor.
That reads like a description of Anduril’s product — because it might as well be. A CBP spokesperson confirmed to The Intercept that under the new law, Anduril is now the country’s only approved border tower vendor. Although CBP’s plans for border surveillance tend to be in flux, Homeland Security presentation documents have cited the need for hundreds of new towers in the near future, money that for the time being will only be available to Anduril.
Anduril did not respond to a request for comment.
- " Tech’s Most Controversial Startup Now Makes Drone-Killing Robots" Bloomberg Oct. 3rd, 2019 https://archive.is/dpwGF
Anduril presents itself as immune to such angst. Its founder, Palmer Luckey, is one of Silicon Valley’s most famous Trump partisans. The 27-year-old has gleefully trolled the Valley’s liberals since he left Facebook Inc. in 2017 under controversial circumstances. Founders Fund, one of Anduril’s first big investors, was started by another Trump stalwart, Peter Thiel. Trae Stephens, Anduril’s chairman, is also a Founders Fund partner and took part in Trump’s transition team. The company recently began working on Maven, the project Google dropped.
- " ‘Never Sleeps, Never Even Blinks’: The Hi-Tech Anduril Towers Spreading Along the US Border" RSN Sept. 17th, 2022 https://archive.is/EzeG9
As well as being deadly, increased surveillance infringes on the civil liberties of people living near the border, Rios believes, because they are often not consulted before new tech is deployed. He grew up in south San Diego, and identifies as a “fronterizo”, or borderlander, and has family on both sides of the frontier, but notes that CBP did not consult nearby residents before installing the infrared tower overlooking the beach.
Anduril has its roots in reactionary right wing circles of Silicon Valley who are determined to become the US government’s next supplier/contractor and beat out the traditional firms (Boeing, Lockheed Martin, etc). They work hard, hire talent, and choose growth over ethics.
These Silicon Valley MICs do this by aggressively pursuing contracts and developing technology that skirts around ethical concerns, often making it a point to do so.
Anduril helps the US government to further militarize the Mexican-US border and makes it a point not to care about the people they’ve hurt along the way. This is simply not compatible with a community who puts people first, nor a community that values free software and the importance placed on digital consent that comes with it.
tomberek assists Anduril in building technologies that inspire terror and harms people indiscriminately. These people could be Nix community members or friends or family of those affected by Anduril’s stated mission. To really get my point across: how can we call ourselves a diverse and welcoming community when the people who are meant to represent us also help hurt us?
And this isn’t an invitation to do vulgar whataboutism and say we have to punish all members employed in IT because we live in a capitalist system and therefore we harm people always so who cares. This is an attempt to whitewash how Anduril is unique in perpetuating the violence that our community should be against by attempting to derail the conversation by assuming that those who speak up “just want to cause trouble or rabble-rousing in the court of social media”
It is not impossible to work and innovate in computer science without having to sell your labor to MIC or any other numerous firms that participate in harmful practices and behavior.
Tomberek said:
I had considered not updating my public LinkedIn, expecting it might be the target of some drama
So you know that this is controversial, you are also aware of the governance crisis that arose just a year ago from this very issue, and yet you still accepted a position at Anduril. Did you do it out of financial distress? Was your personal livelihood, your family, your work at stake if you did not work at Anduril? Was Nix at stake? Are you going to continue to tell us about the “defense industry” while your biggest client (in fact, your only meaningful client) renames itself to the Department of War (NPR Sept 6th, 2025, https://archive.is/2h0w2)? You didn’t mention any of these questions and instead talked down to us as wreckers and drama artists. The fact is, your refusal to do a public disclosure exposes the fact that you know what you’re doing is wrong and you simply don’t want to deal with it.
If you don’t care about anything I’ve said and continue to dig in your heels, you are not fit to represent the Nix community. I’m not going to sit idly while community leaders who work at firms like Anduril shrug this off as “shit happens” and encourage a culture where the ends justifies the means (not coincidentally, the go-to excuse of these companies when they inevitably hurt people).
Nix adoption doesn’t come before people’s lives and your ethics as an engineer and community leader. I ask you to consider stepping down from the SC position and choose to continue supporting Nix without placing yourself in a position of community leadership. I have no interest or power to police your life or to ritually shame you, but if you really care about this community, you wouldn’t put yourself in one of the seats for it and condescend to it as you’ve done now.
- edits for clarification:
- “High level position” is referring to the fact that Bereknyei is a software engineer working directly to shape the products that Anduril produces, they are not a secretary, a custodian, or a HR representative, they are an engineer.
- “Reactionary right wing circles” refers to the clique of men such as Peter Thiel of Palantir fame who also shouldn’t be accepted in our community for similar reasons stated above.
- It is very much possible to work in the field of computer science without contributing to MIC and anyone who works at MIC does so willingly and without coercion.
37 Likes
Silicon Valley was built first and foremost by and on the MIC. The reason we have integrated circuits–integral to the computer you used to type your screed–is because we needed packages that would survive the environments found in ICBMs. Those systems’ sole real use was deterrence by holding civilian population centers hostage. Similarly, the networking tech we’re all using is from the MIC.
Plenty of ethical engineers work on weapon systems (here’s a simple thought exercise: is a munition that is poorly designed and which blows up instead of disarming on time more or less ethical?). There are more views of ethics than you (and others) are evidently aware or tolerant of, and saying “but but but this person works for a company that is unapologetically pro-American interests and therefore is unethical because of their company” is just incredibly lazy. Being a pacifist is not inherently ethical, and neither is working on systems to secure a given population or state.
(And sure, it can be argued one way or the other about if the politics and policy of some country is ethical–or more usefully, perhaps, agreeable–but those conversations are incredibly off-topic and divisive and better served elsewhere. Reinventing geopolitics naively from first principles is a tedious, arduous, and antagonistic exercise.)
Also, if you fully quote that value, it doesn’t look like Tom’s done anything wrong–and it actually makes a lot of posters here look like they haven’t exactly done anything right:
We are here, first and foremost, as individuals working together. Our priority here is to work on Nix projects for the benefit of all their contributors and users. We value building excellent software with a vibrant and diverse community. Individuals gain trust and status by doing the work. Organisations gain prestige by funding the work of individuals and providing resources to support the project.
16 Likes
In the abstract ‘securing a population’ could be ethical, but once the argument hits the real world it must become contextual. You must consider why These protections need to be created.
The perception of a constant security threat is not born in a void. For decades the U.S and western foreign policy, including their militaries and unconditional support for certain foreign governments, has contributed to a destabilization of various regions. This has created a cycle of blowback that is always used to justify more military spending, and more violence.
The cycle has unbelievable human consequences. Every single day millions of young people wake up and watch the latest horrifying imagery of death and destruction raining down on innocent people, all supported by The United States and their private military companies. We fund and develop these protections, all without considering the root cause.
To anyone reading this: Please research, spread the word, and support the innocent people that this sort of funding affects. There are humanitarian organisations providing direct humanitarian relief to those affected. Last time I linked directly to a campaign it was reported for spam, so you’ll have to seek it out for yourself.
21 Likes
The yearly Steering Committee elections are an effective and fair mechanism to influence the membership of the Steering Committee.
I would like all SC members to stay for their term, to minimise disruption, and to provide continuity for the next election.
5 Likes
Correct; Silicon Valley is an intentional collaboration between universities, military organizations, private corporations, and laboring yuppies. However, Silicon Valley is not the locus of computer engineering. In the USA alone, computer engineering is not parochial to the metros of San Francisco, Los Angeles, Seattle, Dallas, Boston, nor Washington, DC; there’s counterexamples for every choice. For your example, consider the history of Texas Instruments. Globally, the modern computer isn’t possible without technology from the Netherlands, Belgium, France, Canada, Taiwan, and Thailand, in addition to the USA. In general, any large metropolitan area today has some tradition of computer engineering.
This might be the root of the misunderstanding. Ethics is not a binary. Rather, ethics is how we make choices. To say something is unethical is to say that it was chosen without deliberation. Following the pragmatix maxim, consider what effects our choices have; then, those effects are the whole of what the choice entails. (Yoneda’s lemma can formalize this in the first-order case!) So, what are the effects of working on weapon systems? Those effects are the only thing entailed by the choice, and they aren’t exactly people-first effects.
This is a conflation of morals and ethics. Morality is about following rules when making choices. They are a sort of cultural gauge for expected behavior. Building weapon systems does not put people first; it does not entail working on Nix projects for the public benefit; it does not entail excellence in software; it actively reduces the diversity of the community; it does not inspire trust; it clearly does not bring prestige to the employer. In that sense, Tom is not meeting the behavioral expectations of the community.
In contrast, what specifically would you say non-Tom posters have done in this thread to violate those cultural expectations?
11 Likes