Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Meta's blockbuster trial draws parallels to big tobacco

economist.com

Post reply on HN
Print view View original

301–310 of 323 posts

Re: Meta's blockbuster trial draws parallels to big tobacco

#301

Post by mantas »

13hours wrote: post #112
mantas wrote: post #88

Earlier quoted context omitted.

Let’s start with banning algorithmic personalized feeds. That seems to be the common determinator of the bad apples.

I find it very convenient for music discovery for example. Maybe start by making it opt in and give parents control over their children's feeds?

Even for music, it sucks compared to the old human-powered Pandora or legendary torrent tracker forums.

Although it may be possible to replicate either with AI. But it still should be conversation style rather than infinite feed thing without any input.

Re: Meta's blockbuster trial draws parallels to big tobacco

#302

Post by alexthedigger »

shimman wrote: post #256
alexthedigger wrote: post #248

Just another shakedown

Good there should be way more. Meta is an enemy of humanity and they must be destroyed.

Yeah but follow the money, those affected never get any of it. All goes into the government coffers.

Re: Meta's blockbuster trial draws parallels to big tobacco

#303

Post by rockskon »

Dylan16807 wrote: post #157
rockskon wrote: post #86

Earlier quoted context omitted.

"You can stop smoking at any time" "You can stop gambling at any time" As much as I may agree with you, you need a different argument.

I don't see why. The problem with your examples is that it's a person lying to themselves. GP is acting as an objective third party and telling the truth about their hypothetical. There are lots of things you really can stop at any time. I could stop eating peanut butter at any time, if I had a compelling reason.

Because if you want to convince someone away from believing it's an addiction so severe it needs an act of government to address, presenting an argument that all-but-panders to their "people spend so much time on it that it must be an addiction!" priors is counter productive.

Re: Meta's blockbuster trial draws parallels to big tobacco

#304

Post by solid_fuel »

charcircuit wrote: post #127
solid_fuel wrote: post #79

Earlier quoted context omitted.

voluntary feedback - things like "I wish the customer service was faster" - is not the same thing as measuring engagement time down to the millisecond and using that data to feed a presentation loop.

Engagement is measuring voluntary feedback. If you aren't engaged you typically either move to the next piece of content or close the app. These are the voluntary mechanisms you as a customer report how good the recommendation is.

No, engagement is not voluntary feedback, because there is no way to use the product without it being measured.

Re: Meta's blockbuster trial draws parallels to big tobacco

#305

Post by wpietri »

danielmarkbruce wrote: post #292
wpietri wrote: post #281

Earlier quoted context omitted.

That definition sounds plausible but breaks down in practice. Every Chicagoan has snowfall-related risk. And conversely, a lot of those guys legally traded where they were not exposed to risk. They were just gambling.

It doesn't break down. If you are creating risk for it's own sake, it's gambling. If you are doing something to hedge risk, it's not gambling. If your position is unrelated to a risk or wildly above the magnitude, you are creating risk for it's own sake. Gambling. The legality of it has zero to do with anything. You can gamble legally and illegally. You can gamble in markets which are made for hedging.

I do think it breaks down in practice for the goal of deciding which given trade/bet is gambling.

I worked for market makers. Their theoretical job is to provide liquidity. They are required to always give a price. So if the theoretical cereal maker wanted to buy a wheat future to reduce risk, the market maker has to offer to sell them one. Then the theoretical farmer comes along later and the market maker buys a matching future and they're back to even.

That by your definition is not gambling. But the fraction of trades that can be justified as actually helping hedge a real risk is very small. Some people trading are pure speculators, pure gamblers. But some of the non-gambler participants will trade in ways that are effectively gambling as well.

I think it's effectively impossible to always tell on a trade by trade basis which ones are gambling. We could say it's all related to hedging, as the law does, but I think that's in practice ridiculous. It's a business filled with degenerate gamblers.

So I return to my point that even though there's no bright line, it's still worth muddling though and finding some practical guidelines.

Re: Meta's blockbuster trial draws parallels to big tobacco

#306

Post by danielmarkbruce »

wpietri wrote: post #305
danielmarkbruce wrote: post #292

Earlier quoted context omitted.

It doesn't break down. If you are creating risk for it's own sake, it's gambling. If you are doing something to hedge risk, it's not gambling. If your position is unrelated to a risk or wildly above the magnitude, you are creating risk for it's own sake. Gambling. The legality of it has zero to do with anything. You can gamble legally and illegally. You can gamble in markets which are made for hedging.

I do think it breaks down in practice for the goal of deciding which given trade/bet is gambling. I worked for market makers. Their theoretical job is to provide liquidity. They are required to always give a price. So if the theoretical cereal maker wanted to buy a wheat future to reduce risk, the market maker has to offer to sell them one. Then the theoretical farmer comes along later and the market maker buys a mat…

One side of a trade can be a gamble and the other a hedge. A single transaction doesn't need to be classified as a gamble or hedge - it can be both.

As for the market maker themselves - they are definitely not gambling. They are trying to make money while reducing risk to the extent possible. You don't walk into the risk management committee and get an excited response because you increased risk all else equal, right?

Re: Meta's blockbuster trial draws parallels to big tobacco

#307

Post by charcircuit »

solid_fuel wrote: post #304
charcircuit wrote: post #127

Earlier quoted context omitted.

Engagement is measuring voluntary feedback. If you aren't engaged you typically either move to the next piece of content or close the app. These are the voluntary mechanisms you as a customer report how good the recommendation is.

No, engagement is not voluntary feedback, because there is no way to use the product without it being measured.

The user chooses themselves what part of a product they engage with. It's like reporting a post would be considered voluntary feedback on the anti spam systems even if it's the case that every report is measured and fed back into the model for predicting reports.

Re: Meta's blockbuster trial draws parallels to big tobacco

#308

Post by itake »

froidpink wrote: post #217
itake wrote: post #72

Earlier quoted context omitted.

How? The 2 feeds on my IG account: "posts" and "reels" show content of people that I do not follow.

if you click on the ig logo you can switch to a feed that only shows accounts you follow

oo thanks for the tip. sadly, this view still shows posts from advertisers that I don't follow. but that is the nature of the beast I guess.

Re: Meta's blockbuster trial draws parallels to big tobacco

#309

Post by solid_fuel »

charcircuit wrote: post #307
solid_fuel wrote: post #304

Earlier quoted context omitted.

No, engagement is not voluntary feedback, because there is no way to use the product without it being measured.

The user chooses themselves what part of a product they engage with. It's like reporting a post would be considered voluntary feedback on the anti spam systems even if it's the case that every report is measured and fed back into the model for predicting reports.

> The user chooses themselves what part of a product they engage with. It's like reporting a post would be considered voluntary feedback on the anti spam systems even if it's the case that every report is measured and fed back into the model for predicting reports.

If the user cannot use e.g. instagram without generating engagement metrics, then it is not a voluntary metric.

User reports - which are generated via a deliberate action from the user - are voluntary.

Re: Meta's blockbuster trial draws parallels to big tobacco

#310

Post by charcircuit »

solid_fuel wrote: post #309
charcircuit wrote: post #307

Earlier quoted context omitted.

The user chooses themselves what part of a product they engage with. It's like reporting a post would be considered voluntary feedback on the anti spam systems even if it's the case that every report is measured and fed back into the model for predicting reports.

> The user chooses themselves what part of a product they engage with. It's like reporting a post would be considered voluntary feedback on the anti spam systems even if it's the case that every report is measured and fed back into the model for predicting reports. If the user cannot use e.g. instagram without generating engagement metrics, then it is not a voluntary metric. User reports - which are generated via a d…

Open the app. Liking videos. Swiping off of a video early. These are voluntary too.
Post reply on HN