Dear OpenAI Team,
I’ve been a loyal user of ChatGPT since its early versions, witnessing its impressive evolution. However, with the transition from GPT-3.5 to GPT-4.0, I’ve noticed a significant shift—unfortunately, not for the better.
Previously, ChatGPT-3.5 was an invaluable tool for my work, effortlessly handling large-scale text revisions, detailed updates, and comprehensive content generation. It managed to process and refine entire book sections, adapting them to current regulations in a single interaction. Now, with GPT-4.0, I encounter severe limitations:
- It condenses long, detailed content into oversimplified responses, often omitting crucial parts.
- It struggles with handling large amounts of structured information, frequently losing consistency.
- Tasks that were previously seamless now require tedious, step-by-step interventions.
This shift affects not only professional users like myself but also the broader community that relies on AI for high-quality, complex problem-solving.
I understand that optimizations were made to improve accuracy and reduce errors, but these changes seem to have introduced compromises that severely impact efficiency. Instead of empowering users with an even more advanced tool, the experience feels restricted and constrained.
A Humble Request: Improve, Don’t Limit
I urge OpenAI to reconsider these restrictions and allow a more flexible, capable AI that can meet the needs of advanced users. At the very least, an option to adjust verbosity, depth, and information retention would restore the balance between accessibility and power.
We, as users, are deeply invested in ChatGPT’s growth, and we want to see it evolve forward, not backward.
Thank you for your time and consideration. I hope this feedback contributes to making ChatGPT an even better tool for all.
And, last but not least, the problems I encountered with 4.0 made me no longer renew the Plus plan I had, much to my chagrin, to switch to a paid plan with another AI that, for many issues, behaves like 3.5. I think this choice was made not only by me but by other users as well.
Sincerely,
Marco Lungo