# Context window full error: theory on why haven't they fixed it yet? — retained record

> Retained reddit record from reddit:ChatGPT.

Canonical: https://highsignal.app/data/records/d1914260b2ff3c23

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:ChatGPT)  published 2026-11-07 11:04 ingested 2026-11-08 02:36

[original source ↗](https://www.reddit.com/r/ChatGPT/comments/1xq3tzm/context_window_full_error_theory_on_why_havent/)

## retained text



```
Hello everyone, 

I know many of you are undergoing this glitch. I have a short explanation on why open AI hasn't solved it yet. First of all, let's start with the basics.

**WHAT THE HELL IS HAPPENING?**

Some users are being blocked off their conversation, with this message

"CONTEXT WINDOW FULL 

 *Start a chat with an automatic summary of this conversation, or upgrade to continue here with more context available in every chat."*

 Popping up after only 10-20 messages. Even  plus users are experiencing this.

**MY THEORY**

This happens due to what I believe,  "A broken Token system". For those who don't know what a token is, a token is *basic units of text, audio, or visual data that a machine learning or large language model (LLM) reads and generates.*  ChatGPT processes tokens using the Byte Pair Encoding (BPE) algorithm paired with specialized tokenizers. I believe that this BPE may be broken, as a normal free GPT model has 27k token capacity, but during this glitch it is miscalculated by 5x and shortened into 3000-5000 tokens, while according to the official token count Chat gpt should easily handle 50+ messages.

**WHY ISN'T IT FIXED (THEORY)?**

By acknowledging this, the OPEN AI company will face corporate humiliation, along with that they will literally confess that their platform is suffering from bugs.

By this point hundreds of users are experiencing this, and Open AI hasn't spoken a word about fixing it.

Because I believe that they are trying to experiment with this bug on users so that they can buy an upgraded plan.

Why?

 

Sometimes aggressive writers in their creative flow can buy the plus plan, resulting as a profit for the company.

Sometimes desperate students or addicts may also buy the plan.

**TIME PERIOD**

I think this will probably extend for days or for some weeks. They need time to see the public reaction.

**MY SOLUTION** 

If this my theory is true, we should probably stop going to their app, and stop giving them view time along with that MOST IMPORTANTLY we should NOT buy their free upgrade that pops up with the warning, if we follow this the OPEN AI team would be forced to fix this bug as their users drop.

Thank you everyone, and along with that, Im pretty new to this AI stuff so please let me know my errors!
```



document fetched 2026-11-08 00:19 · canonical https://reddit.com/r/ChatGPT/comments/1xq3tzm/context_window_full_error_theory_on_why_havent
