Lvxferre

The fun part isn’t even what Apple said - that the emperor is naked - but why it’s doing it. It’s nice bullet against all four of its GAFAM competitors.

This is solely a hypothesis, and it’s backed by informal and subjective impressions, but:

From online interactions with them, I think that Anglos are specially prone to fall for nirvana fallacy, where a solution is discarded as useless because it is not complete or perfect. That might encourage them to frame the problems in a negative way, as it won’t beg the question “what’s your proposed solution?”, thus being easier to gather agreement and support.

If that hypothesis is correct, I predict that English speakers in Canada should follow a similar pattern as the ones in USA, while the French speakers would follow a pattern more similar to Europe and/or the rest of Latin America.

It’ll likely turn out that the more dispassionate people in the middle, who are neither strongly for nor against it, will be the ones who had the most accurate view on it.

I believe that some of the people in the middle will have more accurate views on the subject, indeed. However, note that there are multiple ways to be in the “middle ground”, and some are sillier than the extremes.

For example, consider the following views:

1. That LLMs are genuinely intelligent, but useless.
2. That LLMs are dumb, but useful.

Both positions are middle grounds - and yet they can’t be accurate at the same time.

Here’s a simple test showing lack of logic skills of LLM-based chatbots.

1. Pick some public figure (politician, celebrity, etc.), whose parents are known by name, but not themselves public figures.
2. Ask the bot of your choice “who is the [father|mother] of [public person]?”, to check if the bot contains such piece of info.
3. If the bot contains such piece of info, start a new chat.
4. In the new chat, ask the opposite question - “who is the [son|daughter] of [parent mentioned in the previous answer]?”. And watch the bot losing its shit.

I’ll exemplify it with ChatGPT-4o (as provided by DDG) and Katy Perry (parents: Mary Christine and Maurice Hudson).

Note that step #3 is not optional. You must start a new chat; plenty bots are able to retrieve tokens from their previous output within the same chat, and that would stain the test.

Failure to consistently output correct information shows that those bots are unable to perform simple logic operations like “if A is the parent of B, then B is the child of A”.

I’ll also pre-emptively address some ad hoc idiocy that I’ve seen sealions lacking basic reading comprehension (i.e. the sort of people who claims that those systems are able to reason) using against this test:

• “Ackshyually the bot is forgerring it and then reminring it. Just like hoominz” - cut off the crap.
• “Ackshyually you wouldn’t remember things from different conversations.” - cut off the crap.
• [Repeats the test while disingenuously = idiotically omitting step 3] - congrats for proving that there’s a context window and nothing else, you muppet.
• “You can’t prove that it is not smart” - inversion of the burden of the proof. You can’t prove that your mum didn’t get syphilis by sharing a cactus-shaped dildo with Hitler.

I still fail to see how people expect LLMs to reason. It’s like expecting a slice of pizza to reason. That’s just not what it does.

This text provides a rather good analogy between people who think that LLMs reason and people who believe in mentalists.

Yup. And that’s specially great as the boy is just 15, so he’s starting his career really early.

My sides went into orbit!

The way that the Github comment is phrased, it implies that the link contains additional info that hackermondev didn’t mention. It doesn’t - instead it contains a subset of that info, missing critical bits:

1. That Zendesk initially dismissed hackermondev’s report.
2. That the “third parties” in question were Zendesk’s clients.

Both pieces of info were omitted to back up a lie present in the text, that the bug hunter would have “violated key ethical principles”. He didn’t - as he noticed that Zendesk gives no flying fucks about the security issue, and that remediation was unlikely, he warned the people affected by the issue, so they can protect themselves against it.

Zendesk is not just being irresponsible - it’s also being manipulative, and doubling down instead of doing the right thing (“we incorrectly dismissed that report. It was our bad. Here’s your 2k.”) They have no grounds to talk about ethical principles.

As I mentioned in another thread, about the same topic:

First Zendesk dismissed the report. Then as hackermondev (the hunter) contacted Zendesk’s customers, the issue “magically” becomes relevant again, so they reopen the report and boss the hunter around to not disclose it with the affected parties.

Hackermondev did the morally right thing - from his PoV it was clear that Zendesk wasn’t giving a flying fuck, so he contacted the affected parties.

All this “ackshyually it falls outside the scope of the hunt” boils down to a “not our problem lol”. When you know that your services/goods have a flaw caused by a third party not doing the right thing (mail servers not dropping spoofed mails), and you can reasonably solve the flaw through your craft, not doing so is irresponsible. Doubly true if it the flaw is related to security, as in this case.

I’m glad that Zendesk likely lost way more than the 2k that they would’ve paid hackermondev for the hunt. And also that hackermondev got many times over that value from the affected companies.

What “should be done” is irrelevant - what matters is what “is done”. And plenty servers don’t enforce SPF, DKIM and DMARC. (In fact not even Google and Yahoo did it, before February of this year.)

And, when you know that your product has a flaw caused by a third party not doing the right thing, and you can reasonably solve it through your craft, not solving it is being irresponsible. Doubly true if it the flaw is related to security, as in this case.

Let us learn with Nanni: when Ea-nāṣir sold him shitty copper, instead of producing shitty armour, weapons and tools that might endanger Nanni’s customers, Nanni complained with Ea-nāṣir. Nanni is responsible, Zendesk isn’t. [Sorry, I couldn’t resist.]

[EDIT: can you muppets stop downvoting the comment above? Dave is right, Moonrise is trying to start a discussion, there’s nothing wrong with it.]

What a corporation of muppets! First dismissing the report as “not our problem lol”, then as the hunter contacts affected companies the bug “magically” becomes relevant: they reopen the report, and then boss him around to not disclose it with the affected parties.

I bet that they lost way, way more than the US$2000 that they would’ve paid to the bug hunter. Also, I’m happy that hackermondev got many times that value from the affected companies.

It needs to work and be reliable

True that. And it needs to offer a better service than Google; “our services are federated so please use them!” is not enough to encourage the shift, I think.

The video can be summarised into three main points:

1. Advertisement offering Google a perverse incentive to make its search results worse, so the search ad results look comparatively better.
2. Search engine optimisation.
3. Generative AI integration with Google enshittifying the platform.

I’ll focus on #2. Federated search might alleviate the problem.

It’s counter-productive to optimise a page for multiple search engines, running different algorithms; it might perform better on [let’s say] Google, but worse on [let’s say] Bing, or vice versa, since they run different algos that prioritise different things. As such, almost all SEO is made for Google results.

And, in an environment where no search engine dominates the market, and the search engines use different algos, SEO goes away.

The problem with that is people don’t want to use multiple search engines - they want to use one, that they believe to bring the best results on. (That’s why we have a problem called Google on first place.) If only there was some way for those search engines to coexist, and to benefit from each other… well, that’s basically federation, right?

How I see it working:

• each instance crawls the web separately, focusing on the pages that it wants to
• each instance has its own ranking algorithm
• each pair of instances may opt to federate with each other or not
• each instance can relay search queries to each other, if they’re federated
• as a user inputs a search query, based on keywords and/or user preferences, the instance might decide if it should service the user with local results (from that instance), with results from a federated instance, or a mix of both.

I believe that this system would make SEO really hard to do; in practice you’d be better focusing on good content. It would also lead to a situation where different search engines want to specialise, but still keep each other alive - as they benefit from their peers.

Yeah, the peaks themselves and the quick (~2m) drop after them are not the big deal. The big deal is to slowly bleed users. It’s really problematic in the long term; by no means “MASTADON IS DOOMED!”, but more like “Mastodon, you need to step your game up.”

I wasn’t aware of the connection with the band - thanks for the info! Still, people are bound to associate “mastodon” first and foremost with the critter.

Either way, back in 2008 I bet people were making fun of Twitter for being named after bird sounds, so.

I don’t remember but you’re likely correct. There’s a difference though - Twitter didn’t need to capitalise on every single tiny advantage, Mastodon does it, and while the role of branding might be small it still gives you (or your competitors) some edge.

A model that explains well half of the data is as useful as a coin toss. But let’s roll with it, and pretend that we got two superimposed Gartner cycles here.

The trough would be reached after a sharp drop after the peak, and based on the first peak it would be ~2 months long. That would explain only the period between 2023-07 and 2023-09; the rest of what I’ve pointed out in red is clearly something else, the nearest of what they look like would be a sick version of the “slope of enlightenment” - going down instead of up.

Yeah, the model doesn’t work.

A better way to approach this is to consider three things:

• The main selling point is federation.
• Federation is only perceived as useful for your typical user when a competitor abuses power.
• Mastodon has the drawbacks already mentioned all the time, not just when the competitors fuck it up.

Once you notice those things, it gets really easy to explain what’s happening:

• the peaks are caused by Musk’s acquisition of Reddit and Threads being released (as it brought a lot of discussion about federation up)
• overexcitable people take 1~2 months to realise that Mastodon is not just “Twitter minus Musk”.
• the drawbacks are always there, so Mastodon slowly bleeds users, while only gathering new ones when Musk/Zuckenberg/etc. do something shitty.

By analysing the data this way, not just we’re describing it better, but we can also see where Mastodon needs to improve:

• It needs killer features that are clearly visible for everyone, regardless of federation or “Musk pissed off users”
• It needs to be promoted better. Even among non-Twitter/Bluesky/Threads users.
• Federation itself needs to be promoted better, with simple words, showing why leaving Twitter for yet another walled garden won’t solve shite in the long run.

What I’m saying also partially applies to the “Fediverse link aggregators”, like Lemmy. Lemmy does show some tendency to bleed users, but in smaller degree than Mastodon; but it’s in a better position because there’s only one big competitor, and it keeps fucking it up over and over.

Context for other users - the user above is likely referring to the Gartner cycle:

As anyone here can see, it looks nothing like that pattern that I’ve highlighted.

If the success condition for Mastodon is “to become a long-term viable and attractive alternative to corporate-owned microblogging”, then improvements of the platform are necessary.

To be clear on my opinion in this matter: I want to see Mastodon to succeed, I want to see X and Threads closing down, and IDGAF about Bluesky. However I’m not too eager to engage in wishful belief and pretend that everything is fine - because acknowledging the problem is always the first step to solve it.

First, I will accept the data of the chart at face value, it seems resonably accurate and I don’t have any other data to work off of.

If you do find another source of data, please post it. Relying on a single source (like the Fediverse Observer) is problematic, I know.

To me the declining slopes after the sruges are not relevant to any long term conclusions, they follow a highly predictable curve and doesn’t mean much.

You’re conflating the sharp drops after the surges with the declining slopes.

The sharp drops (like MAU from 12/2022 to 02/2023) go as you said, they don’t mean much. However, the declining slopes are relevant - they span across multiple months (up to ten), and show that Mastodon userbase has a consistent tendency to shrink.

If you look at the end of the graphs you can even see it growing slightly, that is obviously not evidence of anything yet, but to me it is an indication of either a start of another surge, or stability.

We’ll only know if it’s an indication of a surge (sudden influx of new users), or growth (slow influx), or stability in the future. For now it’s an isolated data point.

I believe you are too quick at spreading doom for Mastodon

I’m saying that Mastodon is struggling. I did not say that Mastodon is doomed.

The difference is important here because a struggling network can be still saved, while a doomed one can’t.

Let’s see:

Network effect hits Mastodon specially hard as it competes not just with Twitter, but also Threads and Bluesky. In those situations, a smaller userbase means that people will outright ignore you as an option.

The way that federation was implemented; as linearchaos mentioned in another thread, if you settle in a smaller instance (the “right” thing to do), you won’t get “good collections of off node traffic”. So it creates a situation where, if you know how federation works you’ll avoid big instances, and worsen your own experience; and if you don’t, well, Mastodon’s big selling point goes down the drain.

Federation itself introduces a complexity cost. That’s unavoidable and the benefits of federation outweigh the cost by far; however, the cost is concrete while the bigger benefit is far more abstract.

Branding issues. Other users already mentioned it, but you don’t sell a novel tech named after an extinct animal.

And this is just conjecture from my part, but I think that microblogging is becoming less popular than it used to be; people who like short content would rather go watch a TikTok video, and people who want well-thought content already would rather read a “proper” blog instead.

On a lighter side: the very fact that we’re using the ActivityPub now helps Mastodon, even if we’re in different platforms (like Lemmy, MBin, PieFed, SubLinks). Due to how federation works, you’re bound to see someone in Mastodon sharing content with those forums and vice versa; it could be a bit less clunky but it’s still more content for both sides.

On the text: I think that the author reached the right conclusion through the wrong reasoning. The activity peaks don’t matter that much, when there’s a huge influx of users you’re bound to see some leaving five minutes later. The reason why Mastodon is struggling is this:

Source of the data.

See those slopes down? They show that the stable userbase is shrinking. Even users engaged enough with the platform are slowly leaving, but newbies who could fill their place aren’t popping up.

I don’t know. Instead I’ll focus on my subjective experience with comics and manga, as a nobody from LatAm who likes fantasy.

Manga is something that I grew up with. As adulthood came by, I didn’t feel the need to ditch it - instead I found other manga series to enjoy. There’s stuff for young kids and adults; spicy and tame; comedic and serious; romance and no romance. No matter who you are and the stuff that you like, I feel like you could find at least one enjoyable manga series to read.

In the meantime, what I’ve found from comics elsewhere:

• Local (at least in Brazil) - either tailored for kids (see: Monica’s Gang) or newspaper 4-koma with social commentary (see: anything from Glauco). So only kids get actual stories? Based on Mafalda I feel like that’s how the cookie crumbles in Latin America as a whole.
• European - wider in age demographic than the local ones, and some do have fantasy (Even erotica. Druuna, I’m looking at you. And your butt.), but I feel like they lack dynamic. Even adventure ones like Tintin. Still enjoyable to read, but sometimes my cup of tea might be yerba or coffee, you know?
• United-Statian - Mary Sue protag got superpowers from Z’bh’thy, and now is fighting the Evil for the sake of their country. Skip past 20 years and they’re still in the same slop, never reaching the end, in a multiverse that makes my PC cabling look tidy in comparison.
• manhua (China) - I actually found quite a few enjoyable series (like the Fairy Captivity, Yaoguai Mingdan, My Wife is the Demon Queen). Perhaps not surprisingly they’re similar in spirit to Japanese manga. I could see myself reading more of that stuff. (I’ll skip wuxia though.)
• manhwa (S. Korea) - 90% of the stuff that I’ve seen boils down to either “adultery stories” (I’m not into that stuff) or what feels like ultra-shōnen: “level ZZ is not enough, MC needs to reach level ZZZ”. That said I did find a few enjoyable series, like FFF-Trashero or Carnivorous Princess Yegrinna.

Are they always like this? Probably not; I bet that people can find exceptions to every single bullet point that I’ve listed.

Something must be also said about the synergy between light novels, manga, and anime: if you want you get to enjoy the same story thrice, in three different media, and the pleasure associated with each will be different. And if the story is good enough it won’t tire you down. I simply don’t feel the same in non-Japanese series, even the ones that adapt the same universe across different media (like X-Men).

Yup, we are growing. It isn’t just in number of users, or their activity here, but also in the number of platforms using the protocol - and that’s one of the things that the ActivityPub developers did really right: they picked the concept of federation from earlier protocols/standards (like OStatus), and made it usable for more than just microblogging. The impact of that is twofold:

• future-proofing the protocol. Even if microblogging were to fade away, as a trend, the protocol would still survive.
• older platforms push new ones up, even if the new ones are something completely novel.

(I also like chatting with you!)