The Unoptimized Performance of OpenAI ChatGPT's Front-end Webpage

ChatGPT Bugs
1

Hello there handsome reader or lazy OpenAI Developer, I want to talk about the webpage lag in OpenAI’s ChatGPT’s webpage and why you lazy OpenAI Developer should take notes from JanitorAI’s quality front-end programming because this is a problem since 2022. It also helps me stick around in your site a lot longer and maybe tempted to buy one of you things but since your webpage front-end is so laggy, it put a bad taste in my mouth.

When I load a large conversation chat in ChatGPT’s Webpage, the animations begin to lag and then finally it loaded successfully but when I do normal things like typing, it is lagging immensely, even doing normal things like sending a message to the ChatGPT took forever. To make sure it wasn’t me that wasn’t lagging, I had other tabs and then click on other tabs if they are lagging too and they are not lagging. I checked Task Manager to see my RAM Usage and it was completely fine at around 60% RAM Usage and the CPU usages aren’t that high at around 20-30% CPU usage because I am using a Intel Core i5-11300H CPU. To make even more sure that I am not the only one that is tripping or having problems, I used my phone to then load the same thing in Google Chrome and then the same thing happened. The only time where ChatGPT’s webpage is smooth is when I have a new chat or I am using ChatGPT’s App on Android from Google Play Store. Of course I didn’t want to use my phone to chat with ChatGPT as much because I wanted to type a long prompt to ChatGPT.

Why JanitorAI isn’t lagging? I’m not sure probably because you, yes you, the OpenAI Developer that is probably reading this, is lazy at making the front-end webpage optimized. So, I have a 400 Messages conversation in JanitorAI and it was pretty smooth, the animations are smooth, sending the messages is smooth, and the day-to-day operations is doing just fine because the front-end webpage is actually programmed pretty well because it wasn’t programmed by a lazy OpenAI Developer.

Conclusion? If it weren’t for the laggy webpage experience with ChatGPT, could’ve bought one of your subscription because, what is the point of paying for your service when your front-end webpage is acting like that? Thank you for being convenient and cooperative, lazy OpenAI Developer and thank you random very handsome reader for taking the time to read this crappy essay.

2

So true… horrible frontend issues…

3

The frontend is way over-engineered, unoptimized, and/or uses very slow rendering libraries/dependencies. The opposite of simple. I even had a chat once where, I kid you not, chatgpt began typing in the entirely wrong elements while chopping off large sections of the response. It began trying to provide answers in the ‘thinking’ area or something like that, if memory serves.

The whole thing is so thoroughly immersed in complexity there would likely need to be an entire redesign. Also willing to bet Sam Altman is too cheap to actually hire a decent web dev team to make a halfway-decent webpage, and by halfway decent I don’t mean the nearly unusable mobile app interface, where efficiency comes at the cost of functionality. Just something that maintains their functionality without being an overengineered and broken code cludge. Admittedly, this would probably break a lot of web extensions for their site if they did this, but it would also make them easier to maintain in the future, and I would not have to wait 10s of seconds for long conversations to load so to me it’s justified.

4

My conclusion is that the frontend tries to load the entire conversation regardless of how far down you are. The tab is showing that it’s taking up 1.77gb of memory. For a $500bb company to fumble something you can watch a youtube tutorial to fix is surreal.