Spec-Driven Development: AI Assisted Coding Explained — Transcript

Full transcript

1. 0:00Right now, the way apps are getting built is completely changing because before, writing and
2. 0:04reviewing code was the hardest part. But now it's knowing how to effectively convey what you want
3. 0:10to build with an LLM. And then, my friends, is what's known as spec-driven development.
4. 0:16Thank you very much. Nah, I'm just kidding. Because the thing is, spectrum and
5. 0:23development has become one of the most important skills to learn if you're looking to be an AI
6. 0:27engineer or to use AI to help build your applications. But let me explain why it's
7. 0:33different from a common technique that you might also know, which is vibe coding, right? And this
8. 0:39is typically what people think of when they think of AI-assisted coding or coding agents. So I want
9. 0:45to give you a quick example of what vibe coding typically looks like, and we'll kind of compare
10. 0:50how this is different than spec coding here in a second. But as a user, as a developer, as a builder,
11. 0:57what you're going to start from is typically your AI coding agent, whether it's in a browser or it's
12. 1:01on your machine itself. You're going to start from that initial prompt. So you're going to write that
13. 1:08initial prompt to the LLM and say, hey, I want a specific application that does this functionality
14. 1:14in this specific language like Java or Python. And that prompt is then sent to the model, and that
15. 1:20model is then going to start generating code based on what it thinks that you want from that
16. 1:26project and that it's been trained on as well. So at this point, we've got some boilerplate code, and
17. 1:32this is great for testing. That's why I love vibe coding for, and we're going to go from here. But
18. 1:36the thing is we might not have the exact desired implementation that we want. So we're going to go
19. 1:41ahead and edit that prompt. So we're going to say actually I wanted um, something different
20. 1:48or to use a different type of library or something like that. So we'll go from the prompt
21. 1:53editing over to the AI uh, still continuing to code, and we'll go back and forth until we
22. 1:59reach that desired implementation after a few tries. Right? So we are at the desired state
23. 2:06of implementation. So that's typically what we see when we're doing vibe coding. But the thing is
24. 2:11like how did, for example, the AI model decide to make that specific decision
25. 2:18that it did based on the prompt that we gave? Because we could do a hundred different tries of
26. 2:23this implementation of the app we want to create. We might get a different result every time. And
27. 2:27that frustrates a lot of people. And the thing is,uh, vibe coding kind of skips the traditional
28. 2:34software delivery lifecycle, also known as the SDLC that we're used to in software
29. 2:41engineering. So the software development lifecycle takes a little bit of a different approach,
30. 2:46because we start our project by planning and designing using specific project requirements, or
31. 2:52the PRD as it's typically called. And then we take what we need and require for our project and
32. 2:58begin to implement those features. Once we have those features, we test to see if they work. We do
33. 3:04a little bit of quality assurance, and then it goes into the deployment phase—from dev to
34. 3:09staging to production and finally the maintenance of that project. And so while vibe coding is
35. 3:16fantastic, and I know it feels like magic, spec coding takes a little bit of a different approach
36. 3:21and adds in some of these software development lifecycle components to AI-generated software
37. 3:27development. So let's take a look at what spec coding looks like in a hypothetical scenario. So
38. 3:32spec coding or spec-driven development n-again is going to take into consideration some of those
39. 3:38software development lifecycle aspects, right, but also use the LLM just like vibe coding in order to
40. 3:44as a AI agent write code or run these tests. But it all starts, um, with the prompt, of course, at
41. 3:51first. But the thing is, we're not prompting a specific implementation. We're prompting what we
42. 3:56want our system to do, so the behavior, the constraints that we want. And that specific uh,
43. 4:02specification is then used like a contract to create a requirements. And this requirements
44. 4:09is going to be kind of the main hierarchy of how this project is going to work. So how we want the
45. 4:14model and the agent to write code, to do tests, documentation, verification and much, much more is
46. 4:21all going to be focused from around the requirements itself. And the thing is, if we're
47. 4:26happy with those requirements, so if we're happy we can approve and we can say yes, I want to turn
48. 4:32this requirements specification into a design document that will then have to-dos for each
49. 4:38specific implementation. Or I can say, hey, I want to edit how I want this project to be implemented,
50. 4:44because at this point nothing has been implemented and AI models are all about proper
51. 4:49instructions. So having a spec like this is much better than having the LLM guess what solution is
52. 4:54going to hopefully best fit the user's request. So we go from the design here, and if we're happy
53. 5:01with that design and how we actually want it to be implemented in code, well, then we can have the
54. 5:07model, if we're happy, go off and implement this. So we can go and use that AI agent, or we can
55. 5:13continue to do more requests on the specific implementation features that we want to get back
56. 5:18to. And that's how spec coding works. But really quickly, I want to explain how it's different than
57. 5:23other development cycles. So in traditional development, so the way that,uh, you know, most folks
58. 5:30kind of started by writing code, it was uh, first code, right? And then afterwards it
59. 5:37was documentation, right? So we would kind of start with our intuition um, and we would kind of go from
60. 5:43there. The thing is, we also started to work with test-driven development. So test-driven
61. 5:48development, as you can probably assume in the name, is where we start from the test and what the
62. 5:54functionality we want for this, you know, application's behavior to be, and then go into
63. 6:00writing the code afterwards. So that was also another popular way to approach
64. 6:06development. And the thing with uh, spec coding, so spec-driven development is it kind of turns it
65. 6:13on its head. So we kind of go from specifications, these specs that we have here, to the design
66. 6:20document requirements and actually implementing that. And then we go to code. And so it's kind of
67. 6:26test-driven development and behavior devel ... driven development on steroids, which is really
68. 6:31cool. So let me give you a quick example of how this works in practice. So, we'll start off with
69. 6:36vibe coding. So with vibe coding, again, we're just talking to the model. We're doing quick edits on
70. 6:41the fly. I think that's that's what it's best for, right? We're going to say hey we need um, let's see,
71. 6:46what do we need. Uh, maybe, um, a slash login page for our users, um, to
72. 6:52authenticate. So,okay, that's a great request to the model. But how does it, how do we know what
73. 6:57it's going to do? It might have 30 different ways to implement this, and we might have to go back
74. 7:02and forth with the model. And that can take sometimes longer than just writing the code
75. 7:06ourselves. So, vibe coding is one way to approach that problem. The spec-driven development uh, approach
76. 7:13is a little bit different, but of course still using the LLM to create this requirement that we
77. 7:19have for this feature, right? So, in this case, with spec-driven development, we're going to have a
78. 7:24new feature. And this feature is going to be user authentication, right? So, this is a
79. 7:31new feature that we want to the LLM to build out. Of course, we haven't started implementing yet.
80. 7:35This is just the kind of planning phase.Uh, and we're going to say hey, this is going to be an endpoint
81. 7:40at slash login to do post requests to. Okay, awesome. And then what
82. 7:47variables are we going to take for the username and password. Well, we can say hey we're going to
83. 7:52accept two different variables, so they're going to be user and then pass. Awesome.
84. 7:59So that is in there, and we know when the implementation starts why it got to this
85. 8:03conclusion. Um, let's say for some reason if it doesn't work, we're going to have a fallback, a
86. 8:09failure code. Um, if missing, I don't know, it's a username. Right? And then we can also generate
87. 8:15test cases from here. So, let's go on to test. And we're going to say hey um, valid
88. 8:22credentials. Um, we'll give a 200 code. And this is really cool because
89. 8:29when we're doing AI-assisted coding, we now have less ambiguity for our coding agents. And we can
90. 8:35use spectrum encoding to kind of flip the traditional development model, right, so that we
91. 8:39have this spec and that becomes the primary artifact that drives all this downstream work
92. 8:44like implementation and test and much more. If you learned something today, I'd love if you could hit
93. 8:49the like button and hack the algorithm. And make sure you subscribe for more content around AI and
94. 8:54application development. Have a good one!