Criticism: ChatGPT dot com is so bad for learning engineering!

[potent idol]

I am a Go subscriber!
Please take a look at the screenshot! The same prompt I shared with other AI tools is quite different. I feel bad now for wasting my time on ChatGPT.

Thanks.

[zenith summit]

Hi @potent idol. I'm kapa.ai, your friendly support bot. I will try to answer your question by searching through the knowledge base. :wave:

---

I’m sorry you had a frustrating experience—that’s not what we want, especially if you’re paying for Go.

From the information I have, I can’t see your screenshot or the exact prompt, so I can’t diagnose that specific conversation. The knowledge sources also don’t compare ChatGPT’s answers to other tools, so I can’t directly confirm why those other tools felt better in this case.

What I can share from the docs:

• ChatGPT (and the underlying models) are non‑deterministic: the same prompt can produce different but still “plausible” answers, even with temperature set to 0, because of how probabilities are computed and represented. [Forum answer]
• The models are trained to be broadly helpful, not specifically as an engineering textbook, and they can be confidently wrong or oversimplify technical topics. OpenAI explicitly warns that ChatGPT “can be helpful—but it’s not always right” and encourages users to verify important technical information. [Truth article]
• OpenAI recommends treating ChatGPT as a first draft, not a final authority, and to use critical thinking and external sources—especially in educational or technical contexts. [Practical tips]

For learning engineering specifically, OpenAI suggests using good prompt practices to get more useful, structured answers:

• Be very clear and specific about the task, context, and level (e.g., “Explain this as if I’m a junior backend engineer familiar with Go and Docker, but new to Kubernetes”). [Good prompts]
• Break complex tasks into smaller steps and iterate: ask for an outline, then dive into each section, rather than one huge “teach me everything” prompt. [Good prompts]
• Use prompt‑engineering techniques (roles, constraints, examples) to steer the model toward the kind of explanation or code you want. [Prompt engineering]

If you’re open to it, you could paste the exact prompt you used (and what you were hoping to learn), and I can help you rephrase it using these best practices so you get more useful, engineering‑grade responses next time.

Note: you can ask me a follow up question by @ mentioning me again :speech_balloon: