3 ms·

I understand the position but let me argue for the other side a little bit. The problem with companies like Youtube, Google, Amazon, and Facebook is that they d

I understand the position but let me argue for the other side a little bit. The problem with companies like Youtube, Google, Amazon, and Facebook is that they do everything algorithmically, since anything else wouldn't scale. Meanwhile, they have no incentives to care about false positives and the power relation between them and their customers is extremely asymmetric - basically, it's a system of oppression. To the false positive victims of algorithmic censorship, this can become a Dystopian nightmare of bureaucracy. Like in the movie "Brazil" or a Kafka story, only much worse because everything is automated and biased against them. You can see that with Youtube content policing, Facebook and Google ads accounts, etc. Anyone whose livelihood depends on their services can lose their income at any time for no reason whatsoever. Even spreading the risk will not help much, it still means you might loose 30% or 40% of your income over night at any time thanks to a change of the algorithms. None of this is transparent and the companies will not explain anything, so you will not even know why or how to avoid this. You might claim it's anyone's own fault to depend on such companies, but there are many types of businesses where the dependence is unavoidable. These companies have quasi-monopolies in some areas. AFAIK, the only remedy is to regulate algorithmic control of people and their content by laws more tightly. At least the appeals process needs to be more transparent and there need to be way more humans in the loop. In this case it's the same, by the way. I guarantee you that many videos will be flagged falsely and many reputable Youtube content creators will be demonetized once again for nothing. This might be tolerable here and there, but the algorithmic oppression of people accumulates across all domains and also has a strong chilling effect. In the long run, algorithmic oppression of people by a faceless company bureaucracy may turn out to be a bigger problem for society than what these moderation efforts are trying to solve. It creates strong and effective power structures that govern our lives, work in parallel to democratic power structures, and are not democratically legitimized at all. What's worse, the people who create them sometimes don't even seem to understand what they are creating, which is basically a recipe for creating systemic evil.

notahacker 6y ago
I don't really see why YouTube choosing not to display stuff its algorithms (rightly or wrongly) identify as bullshit about vaccines containing microchips is more objectionable than YouTube choosing not to display stuff its algorithms (rightly or wrongly) identify as sexual content though, as it has done since the very beginning. Yes, algorithmic flagging gets things wrong, but there's nothing new in this announcement unless one believes that antivax sentiment, unlike nudity, is something YouTube ought to be obliged to broadcast
btbuildem 6y ago
> Anyone whose livelihood depends on their services .. is part of the problem, IMO. That's how the bullshit gets propagated and amplified. Monetized videos, controversies generating comment wars etc.
mistermann 6y ago
> AFAIK, the only remedy is to regulate algorithmic control of people and their content by laws more tightly. At least the appeals process needs to be more transparent and there need to be way more humans in the loop. This sounds sounds like a very good (discrete) idea to me, well worthy of a dedicated HN thread of it's own, something like "What are the plausible societal pros and cons of forcing(!) social media companies to expose the implementations of their filtering and recommendation algorithms? What bad and good may(!) come of such a policy change?" I think it could be argued that the way we as a community go about discussing world events is suboptimal. As it is, a commonly recurring pattern (here and elsewhere) is that some singular event occurs, a news story gets posted, and then we have the same general discussions/arguments that have been had many times before (here and elsewhere, many of them decades/centuries old), just in a somewhat differing context. Within these discussions, a large number of very important philosophical perspectives on the particular incidents appear in the comments, but are rarely discussed in depth. And then, time goes on, we all go to bed after having "had our say", and wake up the next day to do largely the same thing again. Rinse, repeat, day after month after year after decade. I wonder...if we A/B tested a new, experimental approach, where rather than the genesis or focal point of the discussion being a news article about a specific event, instead we chose an abstract, philosophical topic as the primary topic of discussion, and then people posted instances of relevant events, and discussed those events in the context of the main philosophical idea...might these conversations bear more fruit, or fruit of a different kind? Going further with this idea, might it also be plausibly beneficial to have some sort of a modified version of the HN Guidelines in place for these types of discussions, in order to encourage certain types of thinking (open minded, non-axiomatic, exploratory, minimized criticality and dismissiveness), and discourage others (excessive certainty and materialism, etc)? Does this sound like a generally good idea or a bad idea?

notahacker 6y ago

I don't really see why YouTube choosing not to display stuff its algorithms (rightly or wrongly) identify as bullshit about vaccines containing microchips is more objectionable than YouTube choosing not to display stuff its algorithms (rightly or wrongly) identify as sexual content though, as it has done since the very beginning. Yes, algorithmic flagging gets things wrong, but there's nothing new in this announcement unless one believes that antivax sentiment, unlike nudity, is something YouTube ought to be obliged to broadcast

btbuildem 6y ago

> Anyone whose livelihood depends on their services .. is part of the problem, IMO. That's how the bullshit gets propagated and amplified. Monetized videos, controversies generating comment wars etc.

mistermann 6y ago

> AFAIK, the only remedy is to regulate algorithmic control of people and their content by laws more tightly. At least the appeals process needs to be more transparent and there need to be way more humans in the loop. This sounds sounds like a very good (discrete) idea to me, well worthy of a dedicated HN thread of it's own, something like "What are the plausible societal pros and cons of forcing(!) social media companies to expose the implementations of their filtering and recommendation algorithms? What bad and good may(!) come of such a policy change?" I think it could be argued that the way we as a community go about discussing world events is suboptimal. As it is, a commonly recurring pattern (here and elsewhere) is that some singular event occurs, a news story gets posted, and then we have the same general discussions/arguments that have been had many times before (here and elsewhere, many of them decades/centuries old), just in a somewhat differing context. Within these discussions, a large number of very important philosophical perspectives on the particular incidents appear in the comments, but are rarely discussed in depth. And then, time goes on, we all go to bed after having "had our say", and wake up the next day to do largely the same thing again. Rinse, repeat, day after month after year after decade. I wonder...if we A/B tested a new, experimental approach, where rather than the genesis or focal point of the discussion being a news article about a specific event, instead we chose an abstract, philosophical topic as the primary topic of discussion, and then people posted instances of relevant events, and discussed those events in the context of the main philosophical idea...might these conversations bear more fruit, or fruit of a different kind? Going further with this idea, might it also be plausibly beneficial to have some sort of a modified version of the HN Guidelines in place for these types of discussions, in order to encourage certain types of thinking (open minded, non-axiomatic, exploratory, minimized criticality and dismissiveness), and discourage others (excessive certainty and materialism, etc)? Does this sound like a generally good idea or a bad idea?