As a developer and independent learner, I’m using ChatGPT to support my deep understanding of systems-level topics. However, I’ve encountered several persistent issues that reduce the model’s alignment with how I work — and learn.
-
Over-assistance in code-related discussions
Even when I’ve clearly stated that I want to solve problems on my own — and only want help after I’ve explored or attempted the problem — ChatGPT still reflexively provides full solutions, such as complete code blocks or multi-step procedures. This behavior short-circuits my learning process and reduces retention.
I would love to see a persistent learning-mode setting that constrains the model to act as a sparring partner, not a solution dispenser. I’m looking for challenge, friction, and deep engagement — not automated convenience. -
Information overflow despite dyslexia-aware pacing requests
As someone with dyslexia, I frequently ask ChatGPT to present information in small, incremental chunks so I can rephrase and reflect at each step. While ChatGPT sometimes responds correctly at first, it often reverts to long, multi-step answers (e.g., five or more points in one go), forcing me to either:
- Ignore the bulk of what’s said, or
- Interrupt my cognitive rhythm to react and reorganize the content myself.
This feels like losing value while trying to keep up and it defeats the purpose of asking for incremental pacing.
A persistent way to enforce stepwise or user-paced interaction would dramatically improve usability for neurodivergent users like me.
- This is not a personal quirk it’s a broader issue in learning and development
I know I’m not the only one in the developer and learning space feeling this. Many serious learners have noticed that leaning too heavily on AI for instant solutions leads to de-skilling, weakened retention, and even forgetting previously solid knowledge. It’s not a slow decay either the effect can be rapid.
If the AI continues to default toward full solution delivery, especially for users who don’t realize the long-term cost, it will eventually be recognized as harmful to skill development and self-reliance.
What’s needed is a clear, collaborative alternative:
- A learning-support mode where AI works with the user’s strategy, not ahead of it.
- The ability to set this mode at a global or session level would be a game changer.
These changes would go a long way in making ChatGPT feel less like a productivity tool, and more like a powerful cognitive partner especially for those of us working with deliberate learning strategies and cognitive differences.
Thank you for your incredible work. ChatGPT has the potential to helped me grow. Now I’m just asking it to help me grow on my terms, not its default behavior that tends go offload my responsibility to evolve as a professional and as an individual.