Me, a software developer who likes to make a living: *audibly gulps*
Jokes aside, this is immensely insightful and helpful. Do you have any suggestions about how to implement this PRD designer for an in-flight project? I've got a decent amount of backend structure built out like the DB schema. Maybe going through this exercise will help me refine that, though, so who knows.
I'm also using ChatGPT rather than Claude, do you know if the gpt projects can be used in the same way?
Also want to say that even as someone with coding experience, I fell into his trap as well. I just thought "oh I'll use AI to help with my project," without asking HOW AI can help with my project. There has been a lot of trial and error on my part, and a lot of iteration in how I'm using the tool. Still, I'm making more progress than I would have otherwise, but with this insight I can further increase my efficiency.
I understand the fear and uncertainty here. I think anyone who works at a keyboard is right to have concerns about their career shifting dramatically. That being said, I do think the next 2-5 years are the best time for those of us willing to learn the tools and scale our impact.
When working with an in-progress project, I recommend inputting any code or documentation as part of your project materials as you try to build the updated PRD and use that to guide specific sub-tasks you want support on with the ongoing project.
Chat GPT projects should work similarly to Claude. The frontier models (Claude, ChatGPT, Grok, and Gemini) all work quite similarly and its more about personal taste to see which one works best for you.
I’m of a similar mind with regard to the career “threat” of gpt. That’s at least in part why I’m jumping in on learning the tool now. If the tool becomes better at my job, being able to use the tool means I’m free to do even bigger and better things later. I’m young enough to learn the new skills to go on to whatever’s next for my field.
I’ve been working on and off with gpt on a PRD for my project and it’s going pretty smoothly. It’s done a good job of remembering a lot of what I’ve already discussed with it and not getting too confused about it.
When that’s all set, I’m debating whether to give a coding AI a whirl now to see what it gives or to just soldier ahead with my current workflow but now informed by a formal PRD.
Another follow up:
Do you have any recourse recommendations for learning how to improve my promoting skills? I could certainly use some help in this area to be more efficient.
The short answer on how to get better at prompting is to ask AI to help you. Give it your simple prompt and ask it to help you improve the prompt for better output. A slightly longer answer is to always give it a clear goal, role, and context when possible. For a deeper dive I suggest Anthropic's guide: https://docs.anthropic.com/en/docs/build-with-claude/prompt-engineering/overview
That’s a great place to start! I was thinking of creating a gpt project called “Project Instruction Generator” and telling it to be an Ai industry expert and power user who can give me instructions to feed to other gpt projects. That way, when I start any new projects I’ll have a leg up. Right now I’ve just been adding piecemeal instructions as I think of them while my project is ongoing.
The AI feedback loop seems quite potent in this way.
Me, a software developer who likes to make a living: *audibly gulps*
Jokes aside, this is immensely insightful and helpful. Do you have any suggestions about how to implement this PRD designer for an in-flight project? I've got a decent amount of backend structure built out like the DB schema. Maybe going through this exercise will help me refine that, though, so who knows.
I'm also using ChatGPT rather than Claude, do you know if the gpt projects can be used in the same way?
Also want to say that even as someone with coding experience, I fell into his trap as well. I just thought "oh I'll use AI to help with my project," without asking HOW AI can help with my project. There has been a lot of trial and error on my part, and a lot of iteration in how I'm using the tool. Still, I'm making more progress than I would have otherwise, but with this insight I can further increase my efficiency.
I understand the fear and uncertainty here. I think anyone who works at a keyboard is right to have concerns about their career shifting dramatically. That being said, I do think the next 2-5 years are the best time for those of us willing to learn the tools and scale our impact.
When working with an in-progress project, I recommend inputting any code or documentation as part of your project materials as you try to build the updated PRD and use that to guide specific sub-tasks you want support on with the ongoing project.
Chat GPT projects should work similarly to Claude. The frontier models (Claude, ChatGPT, Grok, and Gemini) all work quite similarly and its more about personal taste to see which one works best for you.
I’m of a similar mind with regard to the career “threat” of gpt. That’s at least in part why I’m jumping in on learning the tool now. If the tool becomes better at my job, being able to use the tool means I’m free to do even bigger and better things later. I’m young enough to learn the new skills to go on to whatever’s next for my field.
I’ve been working on and off with gpt on a PRD for my project and it’s going pretty smoothly. It’s done a good job of remembering a lot of what I’ve already discussed with it and not getting too confused about it.
When that’s all set, I’m debating whether to give a coding AI a whirl now to see what it gives or to just soldier ahead with my current workflow but now informed by a formal PRD.
Another follow up:
Do you have any recourse recommendations for learning how to improve my promoting skills? I could certainly use some help in this area to be more efficient.
The short answer on how to get better at prompting is to ask AI to help you. Give it your simple prompt and ask it to help you improve the prompt for better output. A slightly longer answer is to always give it a clear goal, role, and context when possible. For a deeper dive I suggest Anthropic's guide: https://docs.anthropic.com/en/docs/build-with-claude/prompt-engineering/overview
That’s a great place to start! I was thinking of creating a gpt project called “Project Instruction Generator” and telling it to be an Ai industry expert and power user who can give me instructions to feed to other gpt projects. That way, when I start any new projects I’ll have a leg up. Right now I’ve just been adding piecemeal instructions as I think of them while my project is ongoing.
The AI feedback loop seems quite potent in this way.
This is great. I had left my app project aside because I felt a bit overwhelmed but this post's renewed my spirit lol Thanks!
Great to hear! I can't wait to see what you create.