I am a game developer and a web developer and I use AI sometimes just to make it write template code for me so that I can make the boilerplate faster. For the rest of the code, AI is soooo dumb it’s basically impossible to make something that works!
The context windows are only so large. Once you give it too much to juggle, it starts doing crazy shit.
Boilerplates are fine, they can even usually stub out endpoints.
Also the cheap model access is often a lot less useful than the enterprise stuff. I have access to three different services through work and even inside GPT land there are vast differences in capability.
Claude Code has this REALLY useful implementation of agents. You can create agents with their own system prompts. Then the main context window becomes an orchestrator; you tell it what you’re looking for and tell it to use the agents to do the work. The main window becomes a project manager with a mostly empty context window, it farms out the requests to the agents which each have their own context window. Each new task is individual, The orchestrator makes sure the agents get the job done, none of the workloads get so large that stuff goes insane.
It’s still not like you can say, go make me this game then argue with it for a couple of hours and end up with good things. But if you keep the windows small, it can crap-out a decent function/module if you clarify you want to focus on security, best practice, and code reusability. They’re also not bad at writing unit tests.
Yes I feel like many people misunderstand AI capabilities
They think it somehow comes up with the best solution, when really it’s more like lightning and takes the path of least resistance. It finds whatever works the fastest, if it even can without making it up and then lying that it works
It by no means creates elegant and efficient solutions to anything
AI is just a tool. You still need to know what you are doing to be able to tell if it’s solution is worth anything and then you still will need to be able to adjust and tweak it
It’s most useful for being able to maybe give you an idea on how to do something by coming up with a method/solution you may not have known about or wouldn’t have considered. Testing your own stuff as well is useful or having it make slight adjustments.
Kind of agree with the rest of your points. Remember though, that the suggestions it gives you, for things you’re not familiar with may very well be terrible ones that are frowned upon. So it’s always best to triple check what it outputs, and only use it for broad suggestions.
I am a game developer and a web developer and I use AI sometimes just to make it write template code for me so that I can make the boilerplate faster. For the rest of the code, AI is soooo dumb it’s basically impossible to make something that works!
The context windows are only so large. Once you give it too much to juggle, it starts doing crazy shit.
Boilerplates are fine, they can even usually stub out endpoints.
Also the cheap model access is often a lot less useful than the enterprise stuff. I have access to three different services through work and even inside GPT land there are vast differences in capability.
Claude Code has this REALLY useful implementation of agents. You can create agents with their own system prompts. Then the main context window becomes an orchestrator; you tell it what you’re looking for and tell it to use the agents to do the work. The main window becomes a project manager with a mostly empty context window, it farms out the requests to the agents which each have their own context window. Each new task is individual, The orchestrator makes sure the agents get the job done, none of the workloads get so large that stuff goes insane.
It’s still not like you can say, go make me this game then argue with it for a couple of hours and end up with good things. But if you keep the windows small, it can crap-out a decent function/module if you clarify you want to focus on security, best practice, and code reusability. They’re also not bad at writing unit tests.
Something like speckit is necessary to make big, sweeping changes that continue past the context window
Interesting project, thanks for sharing
Yes I feel like many people misunderstand AI capabilities
They think it somehow comes up with the best solution, when really it’s more like lightning and takes the path of least resistance. It finds whatever works the fastest, if it even can without making it up and then lying that it works
It by no means creates elegant and efficient solutions to anything
AI is just a tool. You still need to know what you are doing to be able to tell if it’s solution is worth anything and then you still will need to be able to adjust and tweak it
It’s most useful for being able to maybe give you an idea on how to do something by coming up with a method/solution you may not have known about or wouldn’t have considered. Testing your own stuff as well is useful or having it make slight adjustments.
Works in this case doesn’t mean the output works but that it passes the input parameter rules.
For a very lax definition of “works”…
Kind of agree with the rest of your points. Remember though, that the suggestions it gives you, for things you’re not familiar with may very well be terrible ones that are frowned upon. So it’s always best to triple check what it outputs, and only use it for broad suggestions.