GPT-4 won't build your application for you, but it can be a very helpful addition to your toolbox.

The RubyLA blog offers insights, tutorials, and community updates for Ruby developers in Los Angeles and beyond. Developers can explore Ruby language features, web development frameworks, and best practices for building Ruby applications. Additionally, the blog covers Ruby meetups, events, and job opportunities in the Los Angeles area, fostering collaboration and networking within the local Ruby community.

RUBYLAND

GPT-4 can be helpful for Ruby developers in creating a Ruby script that downloads a RubyGem, parses the docs using YARD, and stores the data in a SQLite database. However, GPT-4 has limitations and may require additional troubleshooting and refinements. The provided post offers step-by-step instructions for creating the script and addresses common issues encountered.

Pairing with GPT-4

More experience helps GPT-4 give better results

Let’s have GPT-4 clean this up for a demo

GPT-4 likes to fake it until it makes it (or you lose your patience)

Let’s make a version for squirrels 🐿️

I have the same concern always… You can have the best tools but you need to know what you want to achieve and how you want to achieve. Some practices should be taught to be able to extract the best from this tools and in top of that you should be aware that not all the answers an IA are correct, they can be wrong because of the way they were understood or gathered.

I’ve had some really interesting chats with ChatGPT lately. Especially on “solution code.” I was trying to work out a test. It was erratic at times. It would also not retain certain things in a conversation. It also doesn’t help when you forget what you’ve told it in a given convo. Forget to tell it you use IntelliJ or Gradle and you get VS Code and Maven responses.
It was also a bit annoying that it couldn’t keep the code language it was typing consistent with the syntax choice of its code blocks. And it doesn’t render my code blocks as block as all, so that’s weird. But it comprehends what I wrote fairly well most of the time.
For me, the next step is it being able to load all previous conversations for a given user so you don’t have to boilerplate the conversation for context. I already write a ton, if I’m going to be using AI, can we get DRY? 😜
But hey, it’s free right now, right?! 😂

Great post. Been using it as well and agree with what you’ve said … think it’s good if you want to do small specific stuff, like implementing a client etc, especially if it’s all boiler plate … but once you add complexity and abstraction levels to it … it just gets confused. Might be something that can be improved, but so far going from gpt-3 to gpt-4 no real improvement, at least for what I tried.

I think the same as you. I see it on my work day, and beginners should not use it for code competition. But after seeing the next evolution of GitHub Copilot for example, think it can help newbies with code explanations

Did you guys know that there’s a user agent for Chat-GPT?
<a href="https://platform.openai.com/docs/plugins/bot" target="_blank" rel="noopener nofollow">https://platform.openai.com/docs/plugins/bot</a>