I cannot figure out what people are doing to spend all this money.
I have used a $60 per month Cursor plan on auto, and have never come close to using up my included usage, and I probably have it planning and coding and working for me all through the evenings 4 nights a week.
What on earth are people doing differently that it's costing them so much?
Maybe enabling on-demand usage or other paid models, or on higher modes? What are you doing that requires this? The output from Auto for me is crazy good for the tasks I'm working on, and have yet to run into an issue where it couldn't perform at a high enough level.
We have been interviewing people at work to join our team and they tell us they use $2K per month in tokens with their current employers.... I can't even fathom what's going on here where that would be happening.
> and they tell us they use $2K per month in tokens with their current employers...
perhaps they are simply trying to impress you with their mad prompting skills and like, what self-respecting engineer would be caught dead using less then $2k/month?
giving the context of your interaction with those people, it probably is the simplest answer to your rather baffling question. for the life of me the idea of using $2k/month doesn't even seem possible unless your telling it to waste credits.
Sounds about right. Those people seem to want to create the impression that they are "AI Power Users". That gives them more power inside the organization. People come to them to ask for advice. Also if their output is not good they can claim that is because the AI budget didn't allow them to do more.
If you can give your agent broad access and an effective feedback loop, you just need to steer it and do a final check on outputs.
As an example I might have an agent with access to a browser, logs, metrics, GitHub& CI logs etc. and ask it to implement a new feature.
In Slack I have a few bug reports so I spin up a few more agents. A PM needs a UI tweak so I spin up an agent. You can imagine that a lot of work a dev does isn’t necessarily that complicated and I just need to be there to review the final PR and leave comments as if it were a colleagues (and then my agent goes back, fixes the comments, requests a new review…)
While that’s happening I might be using my actual attention for a meaty feature, design doc, data analysis, etc.
I spend $300/mo for personal use, and a couple thousand at work. Agents can be really transformative and well worth the cost.
Would my company rather pay a few thousand per month, or a several hundred thousand per year for an extra fully loaded engineer? At this point it is _at least_ a 2x multiplier for myself
Totally agree, but then a lot of the same people will be talking about all of the custom instructions/rules/skills/features etc they have set up, so that's eating up a lot of the context window before you even start
When I do use AI, it's just the pure tool itself, and the context is the exact code I'm working with (because I'm trying to see if it can help me solve a specific problem), and I understand the rest of the codebase well enough to know if it's giving me good answers or bad ones
To be honest I’m not doing too much. I’m just on one of the $200 plans, but always hit limits. I only use the best models and mostly use it for various software projects I always wanted to build, but didn’t have time for. I just closely monitor the usage caps and have something running on a Ralph loop most of the time, unless I get near the cap. The post here is more about how I’d start a self-funded software company, if I wasn’t already working full time.
A few things imo, 1) not prompting precisely enough (narrowing scope) means your agent will scan your entire code-base and sometimes get stuck looking at things repeatedly. 2) not checking the output is usually fine but sometimes it produces junk because it doesn't understand, and you cannot prompt your way out of it without reading the code and figuring out the problem. If you leave it on auto it will burn tokens.
Plenty of low level things can trip agents up, too. I just had one inexplicably refuse to read an error about a function needing a bool return value - trying about 10 variations of the same thing before I interrupted it. Skills probably cause issues too, it loves to for example read the source code of libraries I'm using if I give it permission. That's a rabbit hole.
>So if I have a game for a year I paid $70 for, that’s fair, if it goes away, I hope I had a few hours of fun with it.
This example is humorously short and this is why there is backlash to game companies shutting down games. What about the people who bought it towards the end? They just get nothing? All that time and money spent just gets thrown in the trash because they don't want a cloud bill? They either need to opensource the games and servers or keep supporting them for a decade or longer.
So who gets to be the arbiter of how much time $70 is worth?
You?
The companies making the games?
Why should they get to destroy games—gone, forever, with no chance of retrieval or resurrection—that hundreds of people put their time and love into, and millions of people want to play, just because they think it'll make this quarter's stock price numbers look better?
Copyright was created to protect the rights of the creator for a limited time to promote the useful arts. Creations are supposed to become part of the public domain once the creator is no longer getting use out of them. Game companies want to break that bargain, scorched-earth style, and ensure that no one can ever use the things that they create to make anything new.
because you made an effort born of true craftsmanship, because you found the properties of the game that appealed to users and preserved them, instead of taking them away, or locking them into a premium paid tier version.
Not everything is economic value. For gamers, an online game can be a community hub, part of their identity, a hobby. It’s not about whether they got their money’s worth, it’s about destroying a virtual “place” they’re emotionally and socially invested, and the specific skill they posses when they’re there.
I think this is the root of it and what the article describes in the first half. I suspect owning a copy of a game will soon be completely eliminated and replaced w the subscription model. Then when subscription dollars stop flowing, the company naturally winds down the service.
Every developer on videogames has some kind of offline mode already implemented, because its necessary to be able to playtest the game builds on the developer machines. Any argument against SKG is lobbyst nonsense. With the very specific exception of stuff like MMOs. We are seeing cases of pirates being able to play those "turned off games" through cracks and private servers, so there is absolutely 0 reason why the publisher cant already do it.
> Every developer on videogames has some kind of offline mode already implemented, because its necessary to be able to playtest the game builds on the developer machines.
Not guaranteed. Many just run a local server, either in-process or externally. Minecraft's singleplayer mode actually runs a server in-process internally. This simplifies development because singleplayer is conceptually the same as playing alone on a server.
This gets more complicated when there are infrastructure servers in the mix for things like player state, matchmaking, etc. You would bypass that in development but they are required for normal play while being external to the game server.
The problem with “code quality” and LLM’s taking over your first 3 “pillars” is basically that LLM’s don’t care.
I recently had Cursor evaluate a huge code base that we took over. All public stuff, nothing scary security wise, but it was so convoluted that it was taking me forever to find the bugs. It was written by a person, I should add.
I did this in cursor and after one prompt using Plan, it found all the bugs, created a plan to fix them, it looked good, and I had the agent create the fix.
It took 30 minutes.
The client had this project in the hands of another company without ai tools and they couldn’t fix the bugs she told them about.
So my point is, if we are holding on to our jobs for dear life on the basis that “code quality” matters, you might as well kick down the 4th pillar. Like I said, the LLM does not care.
I remember firing up gpt and asking for some code. It was simple JavaScript to solve a problem and I knew immediately, this was the beginning of the end for software development as we knew it.
I built an app I struggled to create for years in a weekend.
Started playing with the API, etc.
I was very much in my oh-no period that very first few months.
Comments explaining what the code does, which is what an LLM could answer, are basically useless comments. Comments that describes why the code is how it is, is a bit more valuable, but also something LLMs cannot really reliably infer by just looking at the code.
I don’t know what I do differently, but I can get Cursor to do exactly what I want all the time.
Maybe it’s because it takes more time and effort, and I don’t connect to GitHub or actual databases, nor do I allow it to run terminal commands 99% of the time.
I have instructions for it to write up readme files of everything I need to know about what it has done. I’ve provided instructions and created an allow list of commands so it creates local backups of files before it touches them, and I always proceed through a plan process for any task that is slightly more complicated, followed by plan cleanup, and execution. I’m super specific about my tech stack and coding expectations too. Tests can be hard to prompt, I’ll sometimes just write those up by hand.
Also, I’ve never had to pay over my $60 a month pro plan price tag. I can’t figure out how others are even doing this.
At any rate, I think the problem appears to be the blind commands of “make this thing, make it good, no bugs” and “this broke. Fix!” I kid you not, I see this all the time with devs. Not at all saying this is what you do, just saying it’s out there.
And “high quality code” doesn’t actually mean anything. You have to define what that means to you. Good code to me may be slop to you, but who knows unless it is defined.
When it comes to games I absolutely don’t care what they used AI for because the point of games is to be fun.
If it’s fun and you used AI, that’s fine with me. The game served its purpose.
The line for me is copyright on images. If you use ai to generate images to copy a popular game art style, I think that’s over the line. Create your own art or pay the artist.
Code however, I see it as a tool. You wouldn’t scold me for hiring a cheap programmer to get the work done. So to me, AI for coding isn’t any different than hiring a programmer to do the work for you. No problem there.
That being said, I do game dev, and using AI to help figure out an algorithm or do the work of creating my inputs code, etc is a big time saver. However, at the moment, it really struggles with anything else because it has no vision and explaining to it how to put code together for a weird game mechanic or level generation reminds me of that game where you explain how to make a peanut butter and jelly sandwich in the 3rd grade, and you tell your teacher to put the peanut butter on the bread and she scoops it out with her hand…
> You wouldn’t scold me for hiring a cheap programmer to get the work done.
It’s literally the same. There is no difference, either you acknowledge AI is potentially a useful tool to lower costs of development (especially important for indie devs) or it’s exploitative and puts both artists and programmers out of a job.
There’s plenty of things in the art workflow that can be automated same as code, pay an artist to do key frames/storyboarding and use the AI to animate between them? Is this exploitative?
EDIT: I’m reminded of this thread from 2019 about a successful game dev that admits their games look like shit due to cheaping out on art: https://news.ycombinator.com/item?id=20804998
It's a real shame its raster functionality wasn't integrated into Illustrator. Adobe really butchered the whole Macromedia portfolio, didn't they?
(For those unfamiliar, Illustrator is a pure vector graphics editor; once you rasterize its shapes, they become uneditable fixed bitmaps. Fireworks was a vector graphics editor that rendered at a constant DPI, so it basically let you edit raster bitmaps like they were vectors. It was invaluable for pixel-perfect graphic design. Nothing since lets you do that, though with high-DPI screens and resolution-independent UIs being the norm these days, this functionality is less relevant than it used to be.)
At my last job m our designer was a Fireworks holdout. It was very pleasant. As someone who has to implement UIs, I greatly preferred it to Figma, though with today's flat boring designs there's a lot less slicing.
I have used a $60 per month Cursor plan on auto, and have never come close to using up my included usage, and I probably have it planning and coding and working for me all through the evenings 4 nights a week.
What on earth are people doing differently that it's costing them so much?
Maybe enabling on-demand usage or other paid models, or on higher modes? What are you doing that requires this? The output from Auto for me is crazy good for the tasks I'm working on, and have yet to run into an issue where it couldn't perform at a high enough level.
We have been interviewing people at work to join our team and they tell us they use $2K per month in tokens with their current employers.... I can't even fathom what's going on here where that would be happening.
reply