Sounds like you are using it entirely wrong then...

Just yesterday I uploaded a few files of my code (each about 3000+ lines) into a gpt5 project and asked in assistance in changing a lot of database calls into a caching system, and it proceeded to create a full 500 line file with all the caching objects and functions I needed. Then we went section through section of the main 3000+ line file to change parts of the database queries into the cached version. [I didn't even really need to do this, it basically detected everything I would need changing at once and gave me most of it, but I wanted to do it in smaller chunks so I was sure what was going on]

Could I have done this without AI? Sure.. but this was basically like having a second pair of eyes and validating what I'm doing. And saving me a bunch of time so I'm not writing everything from scratch. I have the base template of what I need then I can improve it from there.

All the code it wrote was perfectly clean.. and this is not a one off, I've been using it daily for the last year for everything. It almost completely replaces my need to have a junior developer helping me.

▲

jayd16 4 days ago | parent | next [-]

You mean like it turned on Hibernate or it wrote some custom rolled in app cache layer?

I usually find these kinds of caching solutions to be extremely complicated (well the cache invalidating part) and I'm a bit curious what approach it took.

You mention it only updated a single file so I guess it's not using any updates to the session handling so either sticky sessions are not assumed or something else is going on. So then how do you invalidate the app level cache for a user across all machine instances? I have a lot of trauma from the old web days of people figuring this out so I'm really curious to hear about how this AI one shot it in a single file.

	▲	eatsyourtacos 4 days ago \| parent [-]
		This is C# so basically just automatically detected that I had 4 object types I was working with that were being updated to the database that I want to keep in a concurrent dictionary type of cache. So it created the dictionaries for each object with the appropriate keys, created functions for each object type if I touch an object to get that one updated etc. It created the function to load in the data, then the finalize where it writes to the DB what was touched and clears the cache. Again- I'm not saying this is anything particularly fancy, but it did the general concept of what I wanted. Also this is all iterative; when it creates something I talk to it like a person to say "hey I want to actually load in all the data, even though we will only be writing what changed" and all that kind of stuff. Also the bigger help wasn't really the creation of the cache, it was helping to make the changes and detect what needed to be modified. End of the day even if I want to go a slightly different route of how it did the caching; it creates all the framework so I can simplify if needed. A lot of times for me using this LLM approach is to get all the boilerplate out of the way.. sometimes just starting the process by yourself of something is daunting. I find this to be a great way to begin.

▲

lubesGordi 4 days ago | parent | prev | next [-]

I know, I don't understand what problems people are having with getting usable code. Maybe the models don't work well with certain languages? Works great with C++. I've gotten thousands of lines of clean compiling on the first try and obviously correct code from ChatGPT, Gemini, and Claude.

I've been assuming the people who are having issues are junior devs, who don't know the vocabulary well enough yet to steer these things in the right direction. I wouldn't say I'm a prompt wizard, but I do understand context and the surface area of the things I'm asking the llm to do.

▲

Tiktaalik 4 days ago | parent [-]

From my experience the further you get from the sort of stuff that easily accessible on Stack Overflow the worse it gets. I've had few problems having an AI write out some minor python scripts, but yield severely poorer results with Unreal C++ code and badly hallucinate nonsense if asked in general anything about Unreal architecture and API.

▲

lubesGordi 4 days ago | parent [-]

Does the Unreal API change a bit over versions? I've noticed when asking to do a simple telnet server in Rust it was hallucinating like crazy but when I went to the documentation it was clear the api was changing a lot from version to version. I don't think they do well with API churn. That's my hypothesis anyway.

▲

toshinoriyagi 4 days ago | parent | next [-]

I think the big thing with Unreal is the vast majority of games are closed source. It's already only used for games, as opposed to asking questions about general-purpose programming, but there is also less training data.

▲

mac-mc 4 days ago | parent | prev | next [-]

You see this dynamic even with swift which has a corpus of OSS source code out there, but not nearly as much as js or python and so has always been behind those languages.

▲

Tiktaalik 4 days ago | parent | prev [-]

There would be significant changes from 4 to 5, but sadly I haven’t had any improvement if clarifying version.

	▲	frankc 4 days ago \| parent [-]
		Clarifying can help but ultimately it was trained on older versions. When you are working with a changing api, it's really important that the llm can see examples of the new api and new api docs. Adding context7 as a tool is hugely helpful here. Include in your rules or prompt to consult context7 for docs. https://github.com/upstash/context7

▲

rootnod3 4 days ago | parent | prev [-]

How large is that code-base overall? Would you be able to let the LLM look at the entirety of it without it crapping out?

It definitely sounds nice to go and change a few queries, but did it also consider the potential impacts in other parts of the source or in adjacent running systems? The query itself here might not be the best example, but you get what I mean.