Lesson 4 folder - ReadMe is not there, I took it from Lesson 5 folder
Lesson 5 folder’s tech stack and roadmap has details such as using Hono.
Nowhere was this mentioned in Lesson 4, and already it diverges completely from the lesson.
Could the teaching staff take a look to see what else is missing? This course has too many holes in it to follow along
same here. Was a bit confused and could not easily follow the lectures as it does not quite match the repo. Found README.md in Lession 05, which seems to be the right one needed in prompt 1 of Lession 04
Hi guys, thanks for your comments.
I will raise an issue so the files can be revised.
If there’s anything we could clarify so you can continue following the course, please create a topic here in the forum.
Hi everyone!
Thank you for taking this short course and let us know your thoughts! I’m sharing this with the team so we can help you accordingly 
Hi everyone!
The github repo was updated for better readiness! 
-
Lesly
Hi @imgabidotcom I have generated mission.md, roadmap.md and tech-stack.md with the prompts indicated in promts.md and the files are not much close to those in the course. So a couple of interesting things here:
- Under the same inputs (prompts) the result differs, which is the probabilistic nature of LLM and Agents. How is this approached in the real world?
- I guess for the sake of the course, the particular result is not relevant, the focus is the methodology I guess.
Hey @mchiuminatto, you’re spot on with both observations 
On point 1 — yes, LLMs are probabilistic, so identical prompts won’t produce identical outputs. In practice, teams handle this by focusing on whether the output meets the defined requirements of the project, not whether it matches a specific expected output. That’s exactly the right mindset here too.
On point 2 — correct, the specific content of your files isn’t what’s being evaluated. The methodology is: how you prompt, iterate, and structure the outputs into a shared source of truth. If the core project requirements are covered, you’re doing it right 
The prompts.md file on each lesson folder is suppose to be the file that contains all the prompts in that lesson right? They are also repeated in “prompts” folder at the root of the repo
The prompts.md in “The second feature phase”, which is supposedly is Lesson 09 (or Video 10 in the updated repo) does not match what the instructor is showing on the video. For example,
Prompt 1:
Our brand colors are orange and black.
is never issued, and first prompt is actually
(see sc-spec-driven-development-files/prompts/Lesson_09_prompts.md at main · https-deeplearning-ai/sc-spec-driven-development-files · GitHub also)
Instead, this prompt is given
Find the next phase on specs/roadmap.md and make a branch, ask me about the feature spec.
Create:
A new directory YYYY-MM-DD-feature-name under specs for this feature work
In there:
plan.md as a series of numbered task groups.
requirements.md for the scope, decisions, context
validation.md for how to know the implementation succeeded and can be merged
Refer to specs/mission.md and specs/tech-stack.md for guidance.
Important: You must use your AskUserQuestion tool, grouped on these 3, before writing to disk.
Which is from Lesson 05. But that’s ok, if we are repeat the prompt for any new features.
But then again, the next prompt is
Update spec to use PicoCSS
which isn’t anyhere.
Anyways, it is hard to follow
Hey @Mashrur_Mia_Sa_ad.
I am sorry to hear that there are still some issues with the files in the repo.
However, if I may suggest you an alternative: Follow the lessons first in a more passive way, and then try to apply the Spec-Drive Development process in a your own project. You can either follow the ‘Tamagochi’ challenge that was posted here in the forum or go for your own project (just keep in mind not to create a very complex requirements list for this first one).
For me, once I did this, the lessons then ‘became clear’. I started the challenge, and then I revisited some of the videos. By doing so, it was easier to understand and replicate the process the instructor was presenting.
On the other hand, if you find it easier to follow text / written documents rather than follow the videos, I put a summary of the course here: https://community.deeplearning.ai/t/spec-driven-development-with-coding-agents-course-summary/891863