Hacker News new | past | comments | ask | show | jobs | submit login
Musk vows 'thermonuclear lawsuit' as advertisers flee X amid antisemitism storm (independent.co.uk)
12 points by croes 6 months ago | hide | past | favorite | 19 comments



I like how free speech somehow now means silencing your critics.


If only Elon could curtail his personal toxic political views, he would have less issues around X.

There are all sorts of terrible stuff on Meta (esp India, Burma) but people don't get as riled up about Zuck because he comes across as a robotic weird pasty nerd on jetski that does not seem to have any sort of biases as long as you use Meta and seemingly, they try their best i.r.o moderation to keep the nuts in check.


"There are all sorts of terrible stuff on Meta" Ah the old everything is evil so you're already debased argument. Sorry, some things (sympathizing & echoing Nazi tropes) are on a completely different plane and I reject your attempt to drag me down to that level.


Calm down, OP wasn't make a false equivalence to defend the presence of evil content, OP was saying no one complains about META because Zuck isn't a spoiled little white supremacist taking potshots at oppressed groups. OP wasn't trying to diminish METAs evil, just explain why fewer people care. Chill.


The problem isn't that Elon says those things, the problem is he is who he is. Oh, and saying "calm down" is pretty much the pinnacle passive-aggressive move, but you knew that.


He does say those things. I'm not being passive aggressive. I'm being assertive. Learn the difference. Now calm down.


Assertive people don't run around explaining themselves.


Thanks that is my intent.


According to Musk, apparently citing Twitter logs, Media Matters presented an extremely skewed story:

"Of the 5.5 billion ad impressions on X that day, less than 50 total ad impressions were served against all of the organic content featured in the Media Matters article. For one brand showcased in the article, one of its ads ran adjacent to a post 2 times and that ad was seen in that setting by only two users, one of which was the author of the Media Matters article."

https://twitter.com/elonmusk/status/1725771191644758037


That's similar to a problem I have at work, where most things work fine, each problem is rare, but a typical user's chance of running into a few problems during one day is high even if each individual problem is rare.

How long did it take Media Matters to find the ad that ran only twice in that problematic context? I'm guessing they clicked a few times and in a few minutes had found some, uh, problematic context. The ad that they also saw ran in a problematic context in the eyes of most advertisers.

https://nitter.net/Shayan86/status/1725566201760256164#m BTW.


I don't know about any of that but thanks to this lawsuit I'm now aware of what Musk re-tweeted.


Let's try it differently though.

Suppose you want to know how common ads next to problematic tweets is, or some question like that. You can either start by picking an ad and then survey the tweets it accompanies, or you can look for problematic tweets and then look at what ads run next to it.

Elon is implying that the former is the right way. But I'll bet Media Matters did the latter. IMO if Twitter is at all well-designed, the latter is much, much easier, since finding tweets is a core task for the UI.


Does Twitter internally define what is 'problematic' content and not place ads next to it? If it does, that raises a whole bunch of new questions. If not, then how often this happens is going to be a function of how common 'problematic' content is. In other words, if I'm an advertiser and I'm not seeing adequate moderation happening and an ever-increasing volume of shit-to-signal ratio, I'm going to jump ship.


You may reasonably asssume that Twitter does and does not.

Twitter will try to group users into audiences, probably in ways no human really understands even if some humans know enough to read the log files and databases. Do you read Czech tweets? You probably understand that language, then, so you're added to the appropriate bucket of users. Do you retweet things later that go viral? Do you retweet things that are later deleted by moderators?

Another company that did something very like this for recommendation purposes ended up with considerably more than 1000 audiences, and the nazis were concentrated into fewer than 100 of those, even though separating nazis from the rest wasn't a goal. Not perfectly concentrated.


> look for problematic tweets and then look at what ads run next to it.

But this wouldn't answer the question at all, would it? It's like surveying only murderers to determine how many people commit murder - the only answer you can get is "all of them".


It answers other questions, such as "how long does it take you to run into […] if you just click around on the site", which is IMO quite an interesting question, even if Elon Musk would prefer that advertisers don't ask that question.


I really wish that doctors could do some sort of separate-the-conjoined-twins surgery on Elan - to turn the (awesome achievements) SpaceX Elan, (awesome with more caveats) Tesla Elan, and (radioactive dumpster fire) Twitter Elan into 3 separate people.


So we will see if its hot air 17:30 UTC.


Did he actually sue or is it yet another bunch of hot air from Mr. Musk?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: