[EasyMark]

I never thought twitter was a failure until Musk introduced $8 “pump you to the top” blue check accounts, then that’s when the bots -really- moved in because they knew they could pump propaganda for a small fee and get far more traction than they every could by setting up bot armies that were easy to detct. Bsky will be fine as long as they try to control the bots.

[karaterobot]

I guess I didn't define what I think the problems with Twitter are, and now I don't have time for a long message. But, in short, I think that format encourages echo chambers, shortens attention spans, and rewards bad behaviors like bullying, cheap shots, ungenerous reading, etc. None of those are directly related to bots (though maybe bots are a symptom of the same issues). Nor did they emerge with Musk. That Elon Musk wanted to control Twitter might even be diagnostic of a certain kind of basic problem with it—or a basic power—but I'd have to think about that a little more.

[swed420]

> But, in short, I think that format encourages echo chambers, shortens attention spans, and rewards bad behaviors like bullying, cheap shots, ungenerous reading, etc. None of those are directly related to bots (though maybe bots are a symptom of the same issues). Nor did they emerge with Musk.

Having avoided twitter throughout its entire existence for these reasons, I absolutely agree.

Political commentator Matt Christman has a great vlog episode about the inherent problems with platforms such as Twitter, including the societal cost of being an impediment to real progress.

MIND FORCE: FIBONACCI SNIPER | CushVlog 4.29.30:

https://www.youtube.com/watch?v=OsuTh1aK7Tc

https://cushvlog-catalog.vercel.app/transcript/OsuTh1aK7Tc (transcript)

[maxerickson]

Twitter was trying to boost engagement before Musk bought it, and several of the changes they have made since then supercharged it.

Bluesky isn't yet aggressively working to boost engagement. There's the discover tab, and feeds, but those are more opt in types of things. Blocks also work differently, perhaps increasing the echo chamber effect, but also tamping down on squabbling.

[android521]

what you need is blogs or essays. But the audience has become too small so there are fewer writers

[AnthonyMouse]

It's not the audience, it's the monetization. Why are so many things that ought to be a blog post now a YouTube video? Because YouTube has video ads, which pay better.

And it's not just from the perspective of the creator. Lots of people would write blog posts for free, if people would read them. But then Google dumps them out of the search results in favor of YouTube videos that take ten times as long to tell you the same info, and then the bloggers stop writing because nobody is reading.

What we need is a search engine that goes back to giving people what they're looking for instead of showing what pays the search engine company more.

[sofixa]

> And it's not just from the perspective of the creator. Lots of people would write blog posts for free, if people would read them. But then Google dumps them out of the search results in favor of YouTube videos that take ten times as long to tell you the same info, and then the bloggers stop writing because nobody is reading.

I disagree with this take. I hate learning by video unless it's for physical stuff where seeing it is useful (e.g. how to wire something), and thus I always prefer text content when searching for information (due to the higher flexibility in moving at my own pace instead of the narrator's).

I've never had issues with Google pushing videos too much in search results. You have an initial "videos" results block I just scroll past by, and that's that.

The much bigger issue is the narrowing attention span of people. Thanks to social media, like Twitter, but also Stories, Shorts, Snaps, Reels, TikTok, etc. people's attention span is going down. Many people would prefer watching 10 tiktoks instead of spending 10 minutes reading a blog post.

[AnthonyMouse]

> I've never had issues with Google pushing videos too much in search results. You have an initial "videos" results block I just scroll past by, and that's that.

The issue is that some large portion of people don't do that, they watch the video because it's listed first. Which drops the readership of the later-featured blogs below the threshold where the author keeps making them.

[verzali]

Substack, maybe?

[swed420]

Although platforms like Substack alleviate some of the issues mentioned above, there still remains the most important issue of all:

FOR-PROFIT CONTENT CREATION

When I reflect on how awesome the web used to be, the one common theme was the creators were doing it for free, because their passion for whatever they were creating was simply that strong and authentic.

Fast forward to today's hellscape, and we can see the main reason there's a firehose of mediocre/low quality content is because everybody is desperate to make a buck.

[jauntywundrkind]

Twitter had such a strong home team of people trying to monitor the network health.

And they seemed committed to giving researchers and academics lots of access to understand and share how the network was working, look for disinformation & "foreign influence" campaigns.

Now there's no one at the helm and all the researchers and civic interests have been given the boot. They are being threatened by ExTwitter to the tune of 1.5c per tweet accessed if I remember correctly. These people were helping you but there's this sense that finding out what's happening is risky for business, that it could hurt, when in fact the network is just being overrun, turned into the maelstrom, and is deeply into rot and decay stages.

[AnthonyMouse]

You can blame the AI stuff for that. There were so many people scraping the site to do AI training that providing a free firehose API was a request for a denial of service attack.

[mike_hearn]

Not really. I visited Twitter 1.0 HQ some years ago to give them a talk on spam fighting. Actually I went for lunch, and then discovered that a QA session had been volunteered on my behalf by the friend I was visiting.

Twitter's bot fighting teams back then had basically ceased to exist. They had existed once, but at some point the management decided that the biggest abuse problem on the platform wasn't bots or even misinformation but just people being mean to each other. So everyone working in that department had switched to what was basically semi-automated tone policing. By the time I talked to them they were years behind the state of the art in detecting automation.

As for academics and researchers, Twitter 1.0 didn't have the sort of rosy relationship with them you paint here. Researchers had access to the global feed if they paid for it like many other parties did, and Musk started charging much more for that access, but the feed wasn't really useful for finding misbehavior. In fact Twitter 1.0 was pissed off with academia long before Musk arrived on the scene. Academic research into social networks, disinformation and hypothetical foreign influence campaigns is nearly 100% pseudo-scientific garbage pushed for ideological reasons. In 2020 Twitter execs went public with their frustration [1] in a blog post where they pointed out that academics were using an "an extremely limited approach" that was full of useless signals, like:

• Usernames with long strings of digits (the default username generation pattern)

• Lacking info in the bio (even Twitter employees often don't fill it out)

• Posting many times per day (lots of real users do that)

Twitter 1.0: "the threat has evolved and the narrative on what’s actually going on is increasingly behind the curve". Given that there are tens of thousands of such papers coming out of universities, all of them justified by claiming they're ahead of the curve, being constantly behind it is a damning assessment. And that was from people ideologically aligned with academia!

Booting these people off the system was certainly a good move both for Twitter/X's business and frankly, for academia itself. They certainly weren't trying to help.

[1] https://blog.x.com/en_us/topics/company/2020/bot-or-not