How to learn coding in the age of Artificial Intelligence?

[badlydrawnrob]

Programming takes so long to learn and write because computers are literal. They require precision; computers (and therefore programs) have no "do what I mean" or "you get the idea" in their nature, whatsoever. (Quora comment)¹

📰 So much for 'Learn to Code'

A designer can be lazy. A programmer creates a reusable piece of code, so they never have to solve that problem again.² So how, with the future looming large and the threat of AI taking over, will programmers co-exist with their copilots?

• How will kids learn to program in future, or will they even need to?
• How can we encourage learning how to think, rather than what to think?³
• Which subjects best suit an individual, and how to encourage their path?⁴
• Is AI a good teacher?⁵ (How will this affect jobs?)

How much programming knowledge to you need to use ChatGPT confidently? Will no-code solutions be the answer in future? How good is the code they generate?

Will programming languages become easier to read and write if the AI is sentient? Or will extensive training be required to use it? Is it sentient enough to not have to be treat like a robot?

• General AI courses like this one
• AI prompt engineering courses, some good guides and cheat sheets!
• Dealing with AI hallucinations and "bad Truth™"

Creating flashcards with AI

🤖 Here's an example of generating flashcards from notes.

I'm not sure how good these flashcards' content will be, but you can manually edit them as you go. One potential process could be:

1. Read the book, or study the online course
2. Make succinct notes as you go
3. Feed those notes to ChatGPT
4. Ask your AI study partner to generate Anki flashcards
5. Manually edit where necessary

This would work with simple question/answer note types, but more involved cards (such as Missing!) could be harder to achieve.

Problems with using AI (ChatGPT)

• AI hallucinations
  • It looks factual, but on inspection is completely made up?!
  • Supplying a good looking but wrong json decoder
• Broken links and poor footnote references

Footnotes

1. A Quora top comment by David Carew ↩

2. This is wishful thinking perhaps. Why does art repeat itself? Why do we create so many versions of furniture, ceramics, or paintings? The same themes in authors work? Why do technology companies create proprietary products, rather than collaborating on standards-based tech? For many problems we don't need to reinvent the wheel, yet we do. ↩

3. I have some problems with the Chinese way of learning, for example. It's great for Mathematics and some other core subjects, but to me doesn't instil a love of learning for learning's sake. By the time they reach their teens, it feels too much like an exam-factory, spitting out kids who think education is a competition without the flexibility and self-reliance they'll need for Western Universities. ↩

4. The maths in Structure and Interpretation of Computer Systems is not for me 😉 ↩

5. AI hallucinations and overly confident responses are a big problem. ↩