re: a response to my ai class divide post
Came across a response by Randomly Short today. Appreciate the reply, and I thought I'd go more in-depth about some of it.
The author writes:
"Unfortunately that type of work is not what free plans are designed to do. That type of work would honestly be considered something a professional would be doing. Pay for a month of a pro-level plan and get the work done. It’s not like it’s going to disappear afterwards."
In my post, I acknowledge that paying for services that are better and more resource-intensive is a no-brainer. That's obvious. The problem is also not having a short phase of a use case and being too stingy to pay for a month; this use case is ongoing. I'm not only studying for a month, I am studying all year. My volunteer work also involves court cases.
"For students in the US Google offers Gemini AI Pro for free for a year."
Irrelevant point; most of the world is not in the US, other providers offer free plans as well, and as the author also said, this type of work is not what free plans are designed to do. So what gives?
"She also talks about a class divide. Are working students working with many lengthy court cases really part of a class that is being priced out of all of this? If I were still a student I’d not bat an eye at paying for a year of any of the pro plans from the big three if it was going to save me that much time. As it turns out she just admits she doesn’t want to pay. Well…sorry then you deal with the result of that and it has nothing to do with a class divide."
The author is insinuating that law students are part of the upper class. This tends to be true, however, is not true for everyone. I come from a poor background, most of my family has not attended university, and I do it part time next to a full time job. I can also do this because public university in Germany is a lot less expensive than the US.
It misses the overall point of the post though: It isn't about law students being too poor to pay for AI. It's about how different groups and areas of the world have different economic power and access to these systems, and both seminars at work as well as official ads and viral posts online show LLMs doing things that are not feasible with a free plan, but are presented as if they are default possible for everyone. It is about how people deliberately tend to hide how much they pay for AI to be able to do the things they boast about online.
People not wanting to pay could be because they cannot justify the costs. I know US people are used to being thousands in debt, other places aren't. If I was 30k in debt, what more are expenses of 250 dollars a month added to the credit card? Nothing, I guess. But this isn't how other places in the world work, and people very much need to choose between rent, groceries, and shit like this. There is a tension between rising costs of living, and the fantasy online where people pretend a vibecoded app can get you passive income if only you invest into an AI subscription. This doesn't even take into account that many of these subscriptions don't have regional pricing and are a lot more expensive in poorer countries than elsewhere.
It also shouldn't be news that the more money you have, the more a subscription seems like peanuts for the "value" you get, and the more risks you can take. You can just try it, and if you don't like it, cancel - you won't feel like much money was wasted.
Meanwhile others in less fortunate positions have to really weigh the costs and benefits. They also recognize that once they're in it, their projects and workflows rely on it, so this will be a constant recurring cost they might have to shoulder. Many of these people are not even in a position where they can see saved time as saved money; they're not entrepreneurs who, if they take less time for a task, they can accept one more opportunity that brings in more money. Instead, the actual real money on the bank account counts, not time savings. If they're not making rent that month, it doesn't matter how much time it saved, because they cannot as easily turn time into money.
But this especially proves my point:
"If I were still a student I’d not bat an eye at paying for a year of any of the pro plans from the big three if it was going to save me that much time."
Seems like paying more for better models saves more time, which in turn can generate better results that might be profitable, or create more time to earn money in other opportunities or earn more qualifications, or learn faster, which all adds up into more of an advantage. So what happens then if you cannot afford that? What happens then, if your friend on a subscription gets the same thing done three times as fast than you on a free plan, who keeps having to correct the output and wait between cooldowns? What about the likely future in which the free tier enshittifies and the best paid tiers are almost unrecognizably better than the others? What about the people who are subscribed now, but will be priced out of it soon as the prices keep increasing?
Any tool that does not work equally for everyone due to financial constraints is furthering a class divide. That applies to many, many more things than AI, but it applies here too.
The response reads as if it was coming from someone that cannot even conceptualize that poor people or any sort of inequality exists, and it is baffling.
Reply via email
Published