First of all, as someone who is most likely AuDHD, the text layout is really good for the brain to process
'Setting up your AI playground' wasn't super relatable since I use claude code in cursor but this is not a criticism. It's probably hard not to be specific in terms of tech stack, since its easier to explain.
Here, what would be helpful for me is understanding how practitioners use context engineering elegantly in the form of skills.md, cluade.md, agent.md, rule.md etc to improve their system constantly without bloating the context window. Not sure if this is in the scope of the book
---
Loved the 'Rapid prototyping and iteration with AI' chapter. I use the ASCII wireframes all the time. It reduces the cognitive load so so much.
One of the bottlenecks for me is reading through the AI responses lol. I also later move on to building HTML mock-ups/wireframes in claude desktop app.
I mainly use ASCIIs to generate 10-15 ideas for each HMW I define. Going through these ideas is time-consuming so ascii helps a lot
Here I was wondering if delving into the purpose of prototyping and how to design experiments is useful? I definitely tend to over engineeer the prototypes. This results in wasted time and tokens. If AI has the context of the hypothesis that you want to test that might be helpful i'm not sure.
---
I will also take the liberty to share my working style, which is a lot of back and forth conversations with Claude to produce these
1. Product brief
2. Generating HMWs based on product brief
3. Brainstorming ideas with ASCII wireframes based on HMWs
4a. Generating structure and user flows based on ideas selected
4b. HTML mockups for high-resolution wireframes so i can verify the structure
4c. try to get the right CSS tokens by asking claude to provide me visual direction
5. Systems architecture (out of scope for prototyping ig)
6. List of technical risks and spike tests (out of scope for prototyping ig)
7. Once spike tests are done I proceed to do some systems engineering shiz to map needs to system components to verification tests. If spike tests fail, I just give up which i dont think is right and it wastes a lot of time and energy when the spike test fails. I should choose a less risky path
8. then i generate a bunch of handover docs namely 1. spec, 2.tokens.css 3. Prompt for claude code to go all beep boop beep boop
---
so the how to prototype is not the problem. But the "why to prototype" is what my brain forgets and strive to build a perfect app. So for this to happen I should set some hypothesis testing goals or my brain doesnt know which direction to head.
So why to prototype, how to reduce risks, how not to over prompt and engineer the prototype would be really helpful
PS: please take everything I say with a pinch of salt. I haven't been working for the past year due to personal reasons and am not very experienced either. This is my personal take on whatever i remember from the two chapters that i read. I hope this helps
I love the concept. not just the product manager role but we can become solopreneurs. This book is urgently needed. will read this today. Thanks Aman for sharing this and for your work on this book
Aman, do you want to give mention about Antigravity as well? I love this tool better than anything else, but all tools are getting better every day. Probably we are not needed in the near future. :-(. I reached out to the editor if I can be reviewer.
I see until chapter 4. Hope to see other chapters as they get ready. I did reach out to Angela as I am super excited about this book. Want to seek opportunity to review/contribute in anyway possible.
I'm getting this error code as well, Aman
"This code can no longer be redeemed."
Is there limited seats or something like that?
We just re-upped the code so it should work now if you want to give it a shot.
https://learning.oreilly.com/get-learning/?code=APM26
This is the link directly to the title once you've signed up: https://www.oreilly.com/library/view/ai-product-management/0642572251048/
Thanks again!
I've read the two chapters
First of all, as someone who is most likely AuDHD, the text layout is really good for the brain to process
'Setting up your AI playground' wasn't super relatable since I use claude code in cursor but this is not a criticism. It's probably hard not to be specific in terms of tech stack, since its easier to explain.
Here, what would be helpful for me is understanding how practitioners use context engineering elegantly in the form of skills.md, cluade.md, agent.md, rule.md etc to improve their system constantly without bloating the context window. Not sure if this is in the scope of the book
---
Loved the 'Rapid prototyping and iteration with AI' chapter. I use the ASCII wireframes all the time. It reduces the cognitive load so so much.
One of the bottlenecks for me is reading through the AI responses lol. I also later move on to building HTML mock-ups/wireframes in claude desktop app.
I mainly use ASCIIs to generate 10-15 ideas for each HMW I define. Going through these ideas is time-consuming so ascii helps a lot
Here I was wondering if delving into the purpose of prototyping and how to design experiments is useful? I definitely tend to over engineeer the prototypes. This results in wasted time and tokens. If AI has the context of the hypothesis that you want to test that might be helpful i'm not sure.
---
I will also take the liberty to share my working style, which is a lot of back and forth conversations with Claude to produce these
1. Product brief
2. Generating HMWs based on product brief
3. Brainstorming ideas with ASCII wireframes based on HMWs
4a. Generating structure and user flows based on ideas selected
4b. HTML mockups for high-resolution wireframes so i can verify the structure
4c. try to get the right CSS tokens by asking claude to provide me visual direction
5. Systems architecture (out of scope for prototyping ig)
6. List of technical risks and spike tests (out of scope for prototyping ig)
7. Once spike tests are done I proceed to do some systems engineering shiz to map needs to system components to verification tests. If spike tests fail, I just give up which i dont think is right and it wastes a lot of time and energy when the spike test fails. I should choose a less risky path
8. then i generate a bunch of handover docs namely 1. spec, 2.tokens.css 3. Prompt for claude code to go all beep boop beep boop
---
so the how to prototype is not the problem. But the "why to prototype" is what my brain forgets and strive to build a perfect app. So for this to happen I should set some hypothesis testing goals or my brain doesnt know which direction to head.
So why to prototype, how to reduce risks, how not to over prompt and engineer the prototype would be really helpful
PS: please take everything I say with a pinch of salt. I haven't been working for the past year due to personal reasons and am not very experienced either. This is my personal take on whatever i remember from the two chapters that i read. I hope this helps
I love the concept. not just the product manager role but we can become solopreneurs. This book is urgently needed. will read this today. Thanks Aman for sharing this and for your work on this book
Aman, do you want to give mention about Antigravity as well? I love this tool better than anything else, but all tools are getting better every day. Probably we are not needed in the near future. :-(. I reached out to the editor if I can be reviewer.
More on Antigravity coming soon definitely :)
I see until chapter 4. Hope to see other chapters as they get ready. I did reach out to Angela as I am super excited about this book. Want to seek opportunity to review/contribute in anyway possible.
Are we there yet, Aman? :)
Yep the code isn't being accepted right now again
I'm getting "This code can no longer be redeemed." on my end using the updated code link below. 🙁
Aman, I am getting “This code can no longer be redeemed.” I guess I was too late.
We just re-upped the code so it should work now if you want to give it a shot.
https://learning.oreilly.com/get-learning/?code=APM26
Thanks again!
Yeah, that worked. Thanks!
The link just takes me to the homepage. Not the actual book. Looking forward to this anyway.
You can search for the title "AI product management" in O'Reilly and see if you're able to access it there! Let me know if not
Thank you! That worked. Looking forward to this!