Elon & Twitter

[MC3OZ]

In the same vein, here's another article which shows, among other things, that
Musk's tweets, even if later erased, aren't helping:

Elon Musk Is Turning Twitter Into a Haven for Nazis

Musk has welcomed neo-Nazis back onto the platform, engaged with them on his timeline, and posted multiple tweets that appeal directly to them.

www.vice.com

Taken together with progressive Prof. Seth Abramson's damning pinned Tweet thread from today
(which intimates that Musk's "bot purge" is only happening to left-wing worthies), it doesn't
look good for Musk.

We are more likely to hear from left wingers being banned, than we are from right wingers being banned,

Left wingers are more likely to assume that the reason that they have been banned is because they are left wing.

People are more likely to aggregate a collection of left wingers being banned, and present that as evidence.

In terms of being detected as a bot we don't know the algorithm, but I would be surprised if it is looking at the identify of the poster.
Bots are being banned under the assumption that they are not real people, that was always going to be a difficult software challenge,

Anyone in the non-bot category should be banned entirely because of the contents of their posts, not who they are.
The evidence here is sketchy, still I'm not seeing a definite smoking gun that people are being banned because of who they are, There probably is a definite organised campaign to report some posters based on their political leanings, and who they are. Maybe some staff are making bad calls.
This could be skewing the statistics, via deliberate gaming of the reporting system, report the issue to Elon, and see what he does about it,

 

[italian bread]

Can you tell me which part of this tweet is in violation of Twitter TOS?

I’m seeing a guy who was apparantly criticised by Jewish organisations for being antisemitic and says he can’t be antisemitic ‘because all black people are Jew’. A stupid statement, but which rule does it break?

”death con 3”, despite Ye not understanding what defcon means, sounds threatening to me. I’m assuming threats violate Twitter rules.

 

[heltok]

In terms of being detected as a bot we don't know the algorithm, but I would be surprised if it is looking at the identify of the poster.
Bots are being banned under the assumption that they are not real people, that was always going to be a difficult software challenge,

The question of who is a bot or not is gonna get really challenging very soon. ChatGPT is getting crazy good. Here are some examples:

Add to this that image generation is getting crazy good. One day after ChatGPT came out, someone came up with the idea to pair it with a diffusion algorithm:

https://twitter.com/x/status/1598020781065527296

Is your mind blown yet? This is just GPT3.5, GPT4 is rumoured to be coming out soon and be mindblowingly much better.

Imo the $8/account will be a much needed filter to reduce the amount of bots at least somewhat. Any free platform will have more bots than humans by a large margin very soon. And the bots will be much better than the average human at whatever humans do at the platform. Actually I think I prefer to read the bots thoughts over humans a lot of the time. Hmm....

 

[JRP3]

Any free platform will have more bots than humans by a large margin very soon.

They will probably make more sense then many of the humans soon, so I'm calling that a win. I suspect our future AI overlords will not take kindly to Elon's war on bots.

 

[STS-134]

Imo the $8/account will be a much needed filter to reduce the amount of bots at least somewhat.

A monthly fee might get rid of most of the bots, but it'll also get rid of most of the humans. People are cheap. There's a reason why there's never been a social media company that had its services as the product instead of its users as the product (to advertisers).

 

[MC3OZ]

The question of who is a bot or not is gonna get really challenging very soon. ChatGPT is getting crazy good. Here are some examples:
View attachment 880317

View attachment 880318

View attachment 880320

Add to this that image generation is getting crazy good. One day after ChatGPT came out, someone came up with the idea to pair it with a diffusion algorithm:

https://twitter.com/x/status/1598020781065527296

Is your mind blown yet? This is just GPT3.5, GPT4 is rumoured to be coming out soon and be mindblowingly much better.

Imo the $8/account will be a much needed filter to reduce the amount of bots at least somewhat. Any free platform will have more bots than humans by a large margin very soon. And the bots will be much better than the average human at whatever humans do at the platform. Actually I think I prefer to read the bots thoughts over humans a lot of the time. Hmm....

Yes, that is why they want to $8 per month, not only is it a good income stream, it at least makes having an army of bots expensive.

It think what they have now for bot detection is crude and may be looking at frequency of posts and the topis the account posts on.

For example 3 weeks of inactivity follow up by 20 posts on a particular topic in 1 hour. The problem is the activity of a real human could fit that profile or any other arbitrary profile they pick.

Similarly if they based the check on particular words or phrases in posts, humans could fit that profile.

If those making the bots can find out what the check is, it is easy to avoid.

 

[MC3OZ]

A monthly fee might get rid of most of the bots, but it'll also get rid of most of the humans. People are cheap. There's a reason why there's never been a social media company that had its services as the product instead of its users as the product (to advertisers).

This is why the banking part of the platform is critical.

If say $12 per month includes a Twitter blue check, a low interest credit card, and a banking transaction account, the Twitter users might save money compared to their normal banking system,

If the credit card has lower fees for merchants then it may be readily accepted or even preferred by merchants, that might encourage modest discounts on some products and services,.

When they don't have the banking part, $8 per month is a much harder sell.

 

[ZenRockGarden]

But, Elon got concessions out of them that no-one else in his position could have gotten. That's what I mean by "win" in this case.

What "concessions" did Elon get? The knowledge that his made up claim Apple was banning Twitter was nonsense?

 

[ZenRockGarden]

We'll never know, and that's OK. If it becomes public, then everyone else will press for the same deal. Basically, it's the equivalent of Ford trying to get a discount on their Tesla purchase. We all know how much luck they had.

Got it - it's the imaginary concessions that you've made up.

 

[JRP3]

Got it - it's the imaginary concessions that you've made up.

The best kind.

 

[ZenRockGarden]

A monthly fee might get rid of most of the bots, but it'll also get rid of most of the humans. People are cheap. There's a reason why there's never been a social media company that had its services as the product instead of its users as the product (to advertisers).

That's not at all clear. A lot of bots in the twitter-sphere are foreign government backed mis-information, plus an assortment of fake influencer-engagement schemes, both of which have a lot of money behind them.

Meanwhile real humans are proven to be deterred from paying monthly fees for news and info at about a 90% rate (literally look at the paywall success rates of any major publication, news aggregator, etc)

I really don't see any world where you get a meaningful part of the general population paying Twitter $8/mo for the right to tweet. If their goal is to make a copy of WeChat.... guess what, it's not a paid subscription service, it's free.

 

[ZenRockGarden]

That's not at all clear. A lot of bots in the twitter-sphere are foreign government backed mis-information, plus an assortment of fake influencer-engagement schemes, both of which have a lot of money behind them.

Meanwhile real humans are proven to be deterred from paying monthly fees for news and info at about a 90% rate (literally look at the paywall success rates of any major publication, news aggregator, etc)

I really don't see any world where you get a meaningful part of the general population paying Twitter $8/mo for the right to tweet. If their goal is to make a copy of WeChat.... guess what, it's not a paid subscription service, it's free.

Also remember that Twitter literally sells purpose-made bot interfaces for businesses to use for various kinds of automated content (and the businesses pay for them). This traffic may be fairly legit - republishing material from other feeds, etc - but it IS a bot. No human is typing in the tweets to these API's.

So the "eliminate all the bots" topic is a lot more nuanced than the slogans and rants might make you think.

 

[STS-134]

I really don't see any world where you get a meaningful part of the general population paying Twitter $8/mo for the right to tweet. If their goal is to make a copy of WeChat.... guess what, it's not a paid subscription service, it's free.

WeChat is also an anti free speech nightmare with the CCP pulling the strings.

 

[ElectricIAC]

WeChat is also an anti free speech nightmare with the CCP pulling the strings.

A tool in bad hands makes not a tool bad.

 

[STS-134]

A tool in bad hands makes not a tool bad.

Well their 微信支付 (WeChat Pay) service is annoying as hell. So many merchants over there have started to only take either 微信支付 or 支付宝 (Alipay) and do not take credit cards at all. The problem? Both of these services are like using a debit card: they want me to hand them money (interest free!) so that I can then use it to pay merchants, plus they do not give you any rewards or points for spending money through the service. It's just free money for them (I'm sure they make interest off of the money even if you do not). Plus it doesn't come with any of the consumer protections of using a credit card. Be careful what you wish for. The last thing I want to see is a payment system like that become prevalent in the US. It's annoying enough having to deal with this in the PRC.

 

[sunwarriors]

Cant believe Im saying this, but I am feeling envy he's prioritizing Twitter over Tesla a little.

https://twitter.com/x/status/1598381157124837376

Hah, does he really want me to:

“It's better to keep your mouth shut and appear stupid than open it and remove all doubt”

 

[jabloomf1230]

It's a simple approach, but Twitter has recently automated checking for dead accounts and bots. First, the account either gets a text message or an email that "There is suspicious activity on your Twitter account." The account holder is directed to log in to Twitter. Once logged in, the account holder is asked to enter a six digit number that has been sent to the phone number/email. It's not foolproof, but it might get a few accounts locked.

 

[ZenRockGarden]

WeChat is also an anti free speech nightmare with the CCP pulling the strings.

as opposed to Twitter 2.0 with an eccentric billionaire pulling the strings? Yeah. But we were talking about getting people to PAY for it every month

 

[ZenRockGarden]

It's a simple approach, but Twitter has recently automated checking for dead accounts and bots. First, the account either gets a text message or an email that "There is suspicious activity on your Twitter account." The account holder is directed to log in to Twitter. Once logged in, the account holder is asked to enter a six digit number that has been sent to the phone number/email. It's not foolproof, but it might get a few accounts locked.

This type of automated identity checking tactic is used by essentially every platform out there (Facebook, SnapChat, WhatsApp etc), and there is an entire ecosystem which farms and sells fake identities and tools to manage them specifically to bypass these checks.

The bots tend to be experts at this, and simple spot checks will not weed out much other than the least sophisticated ones, which also tend to be low volume. The big guys? They farm fake identities by the hundreds of thousands. Kill off a few - they'll make more.

 

[spacerfc]

Anyone noticed a lot more ads ("promoted" tweets) when browsing someone's tweets? Doesn't appear in Elon's tweets or lists (which is what I use primarily to get information), but if you go to someone's homepage and look at their tweets, there's one promoted tweet per every few tweets by this person, an interesting change from before.