Comment by smokel
6 hours ago
There are many places at Meta that seem to be quite interesting for researchers. You get to play with a lot of hardware, with other talented people, and you can open-source some of your work.
It's all a slippery slope anyway. If you were to work for yourself and publish your research, people might do bad things with it anyway. Consider YOLO [1] as an example of where things might have gone wrong. Another fine example is Fritz Haber [2], who intended some of his inventions for good, some for bad, but eventually society found a way to reverse his intentions.
Given that most computer scientists are pretty good at putting things in perspective, they might come to the conclusion that working for Meta isn't so bad in the grander scheme of things. Slaving away in academia and having your work ignored isn't a very tempting alternative.
Instead of considering how we can make smart people stop working for idiots, it might be more fruitful to spread the idea that we should stop worshipping idiots altogether. If there is one thing I miss from the days when religion was still a thing, it is this suggestion [3].
[1] https://www.deeplearning.ai/the-batch/code-no-evil/
[2] https://en.wikipedia.org/wiki/Fritz_Haber
[3] Exodus 20:3-5
It seems like the argument is that doing science/tech-development for an organization which both has and adheres to benevolent intents and goals, or even just going on your own is the same as working for a company that is intending from the onset to use the work malevolently. Because, all tech gets abused eventually.
This is a terrible argument and is defeatist in the same was as 'what does anything matter at all if the sun is going to explode'.
If you choose to do work for bad leaders, you are going bad in the same way that 'just following orders' for bad things is also bad. You are responsible for the outcomes in those cases. If you are ok with the resulting bad outcomes because the science was interesting and the pay is good, that's your decision. But there is no absolution just because you can suppose that someone else would have done it so it might as well have been you.
> what does anything matter at all if the sun is going to explode
It would not surprise me if this is the exact reasoning that underpins decisions made by leaders of these big companies.
It's terribly hard to convince some people that this is not a sound argument.
In fact, I think it's mostly an evolutionary trait that most of us have, but looking at other species, I don't think it's universal to help others.
> If you were to work for yourself and publish your research, people might do bad things with it anyway.
There's a whole world of difference between someone using your work in a way that you find objectionable and volunteering to accept a paycheck doing work for a company that you know will be using the work they're paying for in a way that you find objectionable.
This is why I have to assume that anybody working for a company is fine with what that company does.