← sources

My bumpy week of adjusting to Claude after having my ChatGPT account deleted last week

source: reddit (reddit:ClaudeAI)published 2026-11-01 15:25ingested 2026-11-02 02:37

retained text

I’m an academic who lost access to ChatGPT last week due to a terms violation. Because it seems like I'll never get my account back, I’ve shifted almost all of my daily activity to Claude, which I’ve been using on-and-off for the past couple of months along with ChatGPT. Claude Code has especially been helpful, but maybe because I'm not a coder, I’ve relied on ChatGPT for almost everything else (checking and replying to emails, drafting and revising writing, research, life-related searches, comparing consumer items, legal matters, etc). 

The transition overall has been quite difficult, especially in these aspects, and I was wondering if you had any advice on how I can optimise it.

1. ChatGPT was the king for revising writing. I could paste in a draft of an email or an article I’m writing with a vague command like ‘improve’ or ‘revise’ and it seemed to know almost exactly what I’m looking for from it without changing too much. With Claude, I have to spell this out every time, and it still sometimes drifts into ‘improving’ things I didn’t ask about. I’ve saved the rules in my preferences, and it helps, but it isn’t the same as being understood like ChatGPT understood my intentions.
2. Claude seems to do well with questions (medical, legal) that have an objective answer, but struggle with those that require interpretive, interesting discussion-based answers (politics, film analysis, philosophy). ChatGPT would commit to an interpreation or an argument, and tell me what would change its mind. Claude tends to give a careful summary of the considerations and stop. It’s good for facts, but for anything with no definitive answer, Claude’s answers are very underwhelming and frankly quite boring to read, even after I tell it what I’m looking for.
3. The most difficult part has been losing the full integration of email, calendar, files and past conversations, all feeding one assistant, which meant I could ask a vague question and it knew what I meant. I’m rebuilding that here with connectors and Projects, but it’s piecemeal.

There’s a 4th point, but I’m not going to delve too much into it (because it seems to be quite aggressively discussed here), because it has to do with usage limits. I reached my weekly limit VERY FAST last week when I was using Claude like I did with ChatGPT, basically dishing very poorly phrased questions back and forth on Sonnet medium. I could barely use Opus. 

It’s very frustrating, but on the bright side, I realised that I was overusing AI to the point that I was outsourcing almost all of my thoughts. I started to use Claude more like a tool to achieve specific ends rather than a constant brainstorm conversation partner.

document fetched 2026-11-02 00:22 · canonical https://reddit.com/r/ClaudeAI/comments/1xl5t9b/my_bumpy_week_of_adjusting_to_claude_after_having