site banner

Small-Scale Question Sunday for January 7, 2024

Do you have a dumb question that you're kind of embarrassed to ask in the main thread? Is there something you're just not sure about?

This is your opportunity to ask questions. No question too simple or too silly.

Culture war topics are accepted, and proposals for a better intro post are appreciated.

4
Jump in the discussion.

No email address required.

Has anyone else noticed Twitter being really buggy since Elon Musk took over? It's worse than Facebook Messenger. For example, it has had trouble loading images for the last day. I don't get why people have concluded that you can fire all these people and have everything be fine. The content moderation is better, but Twitter as a piece of software is obviously much worse now.

I'm not saying that means firing those people was a bad idea. Maybe having software work perfectly is not worth the cost, and the stock market's reaction to the mass layoffs at FAANG suggests they weren't pulling their weight. But that brings me to a related question. Why was it so common up until recently for people to say that, despite the very high salaries of software engineers in California, they were actually very underpaid given the amount they made for their employers? This now appears not to be even close to true. Why did people think this? Was it just some dumb profit divided by headcount calculation?

I've encountered bugs on Twitter since Elon took over, including images not loading a couple days ago IIRC, but I haven't noticed Twitter being worse in this regard since Elon took over. Maybe there are some actual stats on this, but my experience of Twitter pre-Elon was that we would get various minor outages and/or weird nagging issues every few months that would usually take a day to get fixed, and I haven't noticed this rate changing post-Elon.

The only significant QOL issue I've noticed is the enforcement of the official Twitter app on Android, which also does an absolutely awful job at maintaining your position on the timeline. The amount of times I've seen a Tweet and wanted to look at it more carefully, only to have it forever lost to the abyss because the Twitter app decided to invisibly refresh my feed and force me to jump to the top is uncountable, and this severely negatively affects my Twitter experience.

This is now the third day of images not loading for me.

Twitter has worked just fine for me since the purchase, barring the issues from dumb decision making like the rate limiting and so on. I can't say the actual experience of using the site is any worse overall, so I draw the exact opposite conclusion, that the aggressive culling of the employees was warranted.

Weird, I've had near constant issues. It's probably the buggiest app I use on a regular basis.

The app is indeed weirdly buggy. But it has always been that way for some reason

I don't use the app, just the mobile site, and it works about the same as ever.

Sometimes I see all the tweets on my feed become blank tweets dated 01/01/1980, but that's the only bug I've seen.

The real issue that needs fixing is onlyfans thot follow spamming bots.

I've seen sponsored ads (with the "ad" tag) for individual OnlyFans thots. I've also seen ads for "Bubbling" AI services like it's 2014 4chan. I love the site, but it's going to die.

Yeah, the other issue is that bots have become a lot worse, though I can't clearly connect that to anything Musk has done. The biggest issues for me are crypto scammers and people who post irrelevant videos as replies. My best guess is that this is actually due to my following Vitalik Buterin.

You can't see threads anymore without being logged in, which I assume is a load shedding measure to keep the servers from falling over.

User hostile design like "you have to log in to see this" is really common in the social media site space, it's not just Twitter. Instagram does it, Pinterest does it, etc. I assume that it has something to do with wanting to show advertisers "look we have this many users" so that they can demand higher prices for ads, but I don't actually know for sure.

I think it’s something more substantive than that. In particular, a logged-out lurker is far less valuable than a user. Users are able to engage, create content, drive interactions, and yield far more data for the website to track. So if a lurker is led to sign up for an account because of a logout wall, even if that lurker is initially only intending on using that account to continue to lurk, it’s very likely that they’ll end up engaging with the site in some meaningful way. Even just liking posts gives the site owner a ton of data that can be used to target ads.

(An anecdotal datapoint: this account of mine on The Motte was originally only intended to allow me to lurk more effectively and view the most recent comments across subthreads, a feature forbidden to non-logged-in lurkers. But sure enough, I ended up posting here (albeit very rarely). That’s the sort of behavior that Twitter is trying to capture, but on a far more massive scale.)

There must have been a reason Twitter didn't do it previously.

Sure. But whatever the reason they didn't do it previously, they are now. It seems more likely to me that they are doing so because they hope to capture whatever value other social media platforms find in forcing logins, rather than they are trying to prop up failing infrastructure.

Not sure that I'd call that a "bug" to be "noticed". Didn't Musk publicly announce it, saying that they were finally fighting back against scraper bots, or something?

Scrapers can't log in? I can still use nitter.

It adds a barrier and allows more targeted rate limiting, I guess? The controversy was a while ago, and I didn't pay much attention at the time.

Edit: I feel like I've also seen Nitter display isolated posts the same way Twitter does.

Was it just some dumb profit divided by headcount calculation?

It's precisely this because many leftists think the worker is entitled to most if not all the profits generated from their labor.

Okay, to be more fair, I think some people actually did some more in-depth analysis or had a more nuanced take. For an extremely naive example, they hire a software engineer at $200,000 to make some optimization on their servers that cuts costs of those servers by 10%, and if the company was spending $20 million on servers that's $2,000,000 saved. That example is probably not close to reality though, and I think people were thinking of software engineers in those FAANG companies, in which the software product is the driver of profits.

But it becomes extremely difficult to properly attribute what percentage of a single person's work is responsible for the success/profits of a product, especially when you have tens or hundreds of people working on it. And it's not just the software, there are also other aspects of business such as sales and marketing. Think of how many Google products fail due to shoddy marketing, or how many Google Engineers leave to start their own products/companies and fail (because while they might be technically talented, they lack people management ability for a CEO role or lack the sales/marketing ability to drive consumers to their product).

If it was so easy to generate millions in profit, why don't they A) Start their own business or B) successfully negotiate a higher salary? You could say this was their attempt to create some kind of public awareness/narrative to give a higher chance of them succeeding in option. I also think part of this mentality is driven by hatred for the rich, which is hilarious because as you pointed out, those software engineers are rich, earning the top 1% of income in the United States. But if there is someone richer, I guess you can go hate the richer guy first. Being a millionaire is okay now as long as there are billionaires to hate on.

There's also another perspective to consider. Most software engineers may be costing FAANG money, not generating profit for FAANG. They simply pay those software engineers a high enough salary that they are happy to work at those companies and don't leave to create competing products or go work for their competitors. That's why it was so easy for these companies to let go of people when profits went down because they weren't that valuable to the company. Think of how many "A Day in the Life of a Software Engineer" videos used to come out a few years ago and they showed very little of how much work they do and more on how many free smoothies they can get at the snack bar. (One could argue this is marketing to make the life at the company look better, or it was only focused on the positive aspects of the company). And some people claimed to work only 10 hours a week while being paid mid-high 6 figure salaries, and even more people that were working 2 remote software engineering jobs at the same time. These people are likely not working on the critical products/features that would be the main revenue drivers for the company.

Oh, this reminds me of a dumb take I saw on the system adminstrator subreddit, where several people argued (and everyone else agreed) that, as a system admin, they were worth whatever the lost revenue would be if they didn't do their jobs, which, as they pointed out, was basically all of it. Thinking about this for two seconds should reveal why this makes no sense.

There are lots of employees doing work upon which the same revenue depends. They obviously can't all be paid out of that same revenue. The value added by each step in a process cannot be equal to the total value created by the entire process, even though stopping any given step would destroy all value.

To expand on your example, the software engineer that saves the company $2,000,000 cannot do that without the system on which they're working being built in the first place. Someone had to make that investment, and they did so with the expectation that there would be optimizations that could be paid for with $200,000. If the resulting profits weren't expected, they would have invested in something else.

If it was so easy to generate millions in profit, why don't they A) Start their own business or B) successfully negotiate a higher salary?

You could argue that their employers have monopolies.

There's also another perspective to consider. Most software engineers may be costing FAANG money, not generating profit for FAANG. They simply pay those software engineers a high enough salary that they are happy to work at those companies and don't leave to create competing products or go work for their competitors. That's why it was so easy for these companies to let go of people when profits went down because they weren't that valuable to the company.

Something would have to have changed such that the risk of their starting competitors had gone down.

And some people claimed to work only 10 hours a week while being paid mid-high 6 figure salaries, and even more people that were working 2 remote software engineering jobs at the same time.

It's anecdotal, but I do remember reading a comment somewhere that, at the height of the the lockdowns, expectations had gone out the window and the culture had become one of just making sure everyone's mental health was good. No one was put under any pressure at all to get anything done. It's hard to believe that was common though.

In economics the problem of how to value various actors in a cooperative game is solved by the Shapley Value. In order to calculate it, one needs the profit the company generated if it employed a only certain subset of employees, for every subset. So if it employs 1 sysadmin, 1 rockstar programmer, 2 average programmers, and 2 secretaries, one needs to consider 2^(1+1+2+2)=64 possibilities.

What if it's zero for everything but the set of all employees?

Clearly the profits, if one doesn't take into account risk or replacibility, should be divided equally. It is a simplified model.

This is true of a lot of people though. Anyone working with large amounts of data and money have this large effect on profits through their labour.

Does an IRS worker that prevents ten million in tax fraud every year deserve 200k a year? I feel like should have a lot to do with how replaceable the labour is and some people in SWE (especially at FAANGS) have gotten fairly overconfident in exactly how irreplaceable or necessary they are on an individual level.

One can of course turn this around and say that some people in other sectors are very underpaid for the very valuable and hard to replace labour they do.

Having an effect on profits is not the same as solely generating that profit. Let's say you have a small business that generates widgets generating revenue of $300,000 per year, and it has two employees who operate a machine to make the widgets. If either employee or the machine stops working, no widgets get made and the profits go to zero. So each employee and the machine can claim that 100% of the profits depends on its role. But that can't mean that each employee is worth $300,000 a year. If each employee were paid $300,000 a year and the machine were rented at $300,000, the company would be losing $600,000 a year. If the machine costs say $100,000 a year to rent, then each employee can only be paid $100,000 a year. So an employee who can correctly say that $300,000 of revenue depends on their work is only worth $100,000. More complicated processes with more employees would result in starker differences between these two figures.

I'm sure there's some capacitor somewhere in some multi-million dollar machine that would stop it from working if it failed, but that doesn't mean the capacitor could possibly be worth more than a few dollars, because there are many other components like it and if it did cost more than that, the machine could never have been built in the first place.

That is what I'm saying, it depends on the value of the process and the replaceability of a particular part/person (and to a lesser extent the value add of the person). The tech worker depends on their position in a process/system just as much as the IRS analyst to generate value but one is harder to replace than the other.

Only... Sometimes it other things disrupt this. Collective bargaining (and control over labour supply) and employers with employees in vulnerable positions are common examples, or people that are just too complacent/risk averse to effectively negotiate. People collectively overvaluing something is common as well.

I believe that people in tech (both employees and their employers) have been overestimating their value and replaceability/redundancy. There are still plenty of people in tech that actually are valuable and hard to replace and many of the people that are replaced are needed in other parts of the economy, but they might not command as high wages.

In the specific case of software engineer layoffs, replaceability isn't the issue. They weren't replaced. Their employers got rid of their positions and expected profits went up, which meant that whatever they were doing wasn't producing enough to pay their salaries. So, their mistake was not just to think they were irreplaceable, nor was it to neglect that other employees and capital were needed for them to their jobs.

They must not have understood that what they were doing was not nearly as valuable as they thought.

I agree.

I believe that people in tech (both employees and their employers) have been overestimating their value and replaceability/redundancy.

Yup, anyone with a basic knowledge of economics would understand why the average salaries are the average salary for a particular profession.

I wonder how much of it is a complete lack of knowledge or misunderstanding of the labor market versus fully well understanding how they work but then choosing to argue against the premises/reality that created those systems in the first place and wanting to replace it with something else.

I don't get why people have concluded that you can fire all these people and have everything be fine.

Precisely because I haven't seen those bugs. I have seen people say that they experience a bunch of bugs and I suppose I believe them, but the app seems to just work fine.