AI Bubble Is Bursting! But Software Developers Have a BIGGER Problem! — Transcript

Full transcript

1. 0:02Hi all, welcome back to my channel. So,
2. 0:04from last one or two days, I think you
3. 0:06might have seen the news coming up
4. 0:08because of the AI, what how much money
5. 0:10they have lost because of the AI, right?
6. 0:13So, AI bubble is slowly
7. 0:15busting, right? At least in China, it
8. 0:17has already
9. 0:18got the proof that in their stock 15
10. 0:22decks minus 40% drop they have seen. 480
11. 0:25billion have already wiped up, right?
12. 0:27So, what it's to us?
13. 0:29If we talk about AI bubble, you know,
14. 0:31yeah, let's say last 6 months it was a
15. 0:34very valid news for us that AI bubble is
16. 0:36burst or not, it will be in our favor or
17. 0:38it's, right? But now it's not about us.
18. 0:41It
19. 0:42doesn't matter if AI is AI bubble is
20. 0:44going to burst or not. For us,
21. 0:47specifically for software developer,
22. 0:48either you are a .NET, Java, on UI side
23. 0:51Angular, React, View, Node, anything.
24. 0:54We are bypass that.
25. 0:56You know, so many things are there key
26. 0:58it doesn't matter like AI burst it is
27. 1:00only for the investor. This news is not
28. 1:02for us. Don't get, you know, don't feel
29. 1:05good key yeah if AI is bursting our days
30. 1:07will come, it's not going to happen like
31. 1:09that. Why? Because this is only for the
32. 1:11financial people, those companies those
33. 1:13who have invested so much of into AI.
34. 1:15Now they are asking for a proof. Yeah,
35. 1:17you said that AI will do this or do
36. 1:19that. Now where is the proof? That is
37. 1:21because now proof is not there. We also
38. 1:22know apart from the software
39. 1:23development, no major industry is using
40. 1:26AI as much we are using.
41. 1:28In whether you talk about mechanical,
42. 1:29electrical, medical, medical have, you
43. 1:32know, even robot is out of robot is
44. 1:34somewhat they are using. You know,
45. 1:36China's I don't recall the exact name.
46. 1:38So, the company who was majorly into it
47. 1:41their share are also falling down,
48. 1:42right? So, what is up to us?
49. 1:46Now, the Microsoft severance package
50. 1:48also you might have
51. 1:49seen that Microsoft is giving us giving
52. 1:52their employees that like this amount
53. 1:55and you can leave the company unless
54. 1:56voluntary retirement they are offering
55. 1:58us, Right? Now,
56. 2:01over the last what do we say year?
57. 2:04Around whatever the major IT companies
58. 2:06we have either Infosys, TCS, you know,
59. 2:08Wipro, Tech M. So, around 24,000 people
60. 2:12have
61. 2:13been laid off because of whatever the
62. 2:15restructuring they are doing because
63. 2:17they want to invest the money into AI.
64. 2:19At least AI Indian MNCs companies are
65. 2:21not investing that much into AI. Right?
66. 2:24But majorly, the overall if we see
67. 2:28where it was. Yeah.
68. 2:30Big Tech, right?
69. 2:31The Microsoft, Amazon, Alphabet. So,
70. 2:33these are investing so much of money
71. 2:35into AI. So, they are releasing they are
72. 2:38letting go all its employees.
73. 2:40Because everyone wants to take invest in
74. 2:43the AI now.
75. 2:44But those who are actually using AI,
76. 2:47you know, everyday when you are working
77. 2:49on a particular task, when you are
78. 2:50delivering a task, how much code you are
79. 2:52writing right now?
80. 2:54Those who have got the paid access of
81. 2:55Copilot or even
82. 2:57Lovable or any AI agent,
83. 3:00Cloud also. So, we are not writing the
84. 3:02code by ourselves anymore. We are just
85. 3:05writing the prompt. Yeah, this is these
86. 3:06are the scenarios, these are the API,
87. 3:07these are the payload, this is the Figma
88. 3:10this is the Figma MCP server. Uh
89. 3:13it only does everything. We are not
90. 3:16writing code manually at all. Yeah, some
91. 3:18some cases are there where you need to
92. 3:20write the specific business that also we
93. 3:22can
94. 3:23ask it to write, but some people are
95. 3:24still writing it. If they're not asking
96. 3:26everything to get
97. 3:28Some are doing a smart development that
98. 3:30instead of asking it a one go, they're
99. 3:32asking it step-by-step. Yeah, first do
100. 3:34this, then do this, verify. So, now the
101. 3:36role of our software engineer is just to
102. 3:39review the code what AI is writing.
103. 3:41So, really does it matter if AI bubble
104. 3:43is going to burst or not?
105. 3:45It has nothing to do with us with us
106. 3:47now.
107. 3:48Because everyone knows even if you don't
108. 3:50have a paid access of any AI agent, you
109. 3:52know, the Microsoft that 360 Copilot
110. 3:54also can does the job for you.
111. 3:56Any let's say you don't want to use
112. 3:58GitHub Copilot or Cloud or any AI agent.
113. 4:02You host your LLM model locally. You
114. 4:03create a harness and you can use that.
115. 4:05You don't need to depend on that also.
116. 4:07Any basic model, either you are a .NET
117. 4:09developer, Java, Angular, React, any
118. 4:12basic model will do the coding.
119. 4:15We don't need Cloud or GitHub Copilot.
120. 4:17Right? So, we should not get worried
121. 4:20about that AI bubble. It's not related
122. 4:23to us. It is only for that financial
123. 4:24people. What companies are investing in,
124. 4:27that is only for them.
125. 4:28Right?
126. 4:30Okay? So, as I said the last 1 year,
127. 4:32this is the count what all major IT
128. 4:35companies have released the people. Now,
129. 4:37what they are doing then? So, as I have
130. 4:39in previous video I have done that is
131. 4:40slower
132. 4:42Uh what where is that slide? Less
133. 4:44people.
134. 4:45You know, previously we used to have so
135. 4:47big head count in every project, but now
136. 4:49that is limited to very small people.
137. 4:51Small team same output we are getting
138. 4:53within the less time.
139. 4:55You know, previously when we used to get
140. 4:56a feature, we used to ask for yeah, I
141. 4:59need 2 days, 3 days. You know, story
142. 5:01estimation we used to do. But, now that
143. 5:03is slowly getting reduced to 1 day only.
144. 5:06Because AI is writing the code. You are
145. 5:08just reviewing and you are testing what
146. 5:09AI has written. Is it according to your
147. 5:11business use case and the standard what
148. 5:13you are needing? So, small teams
149. 5:17agents and the small team is delivering
150. 5:19the project now.
151. 5:21Okay?
152. 5:22That's why this number is there key that
153. 5:24over a year this many this are these are
154. 5:26the people that got laid off. That's why
155. 5:28in the market also so many people are
156. 5:29there with immediate joiner. Because
157. 5:31because of the layoffs happening
158. 5:33so many so many people are without job
159. 5:36right now.
160. 5:37Because this is very huge number for 1
161. 5:39year. You know, previous we have not
162. 5:41seen this downfall in IT Indian IT
163. 5:43specifically from so many years, but
164. 5:46from last year it is changing
165. 5:47drastically.
166. 5:49Okay? Now,
167. 5:51as I said key it doesn't matter AI burst
168. 5:54or not, we should not be worried about
169. 5:56that. That is not for us. What we should
170. 5:59do?
171. 6:00Right? If you are a back-end developer,
172. 6:02learn UI. If you are a UI developer,
173. 6:03learn back-end. Be full-stack. That is
174. 6:05the first priority. Then learn system
175. 6:07design. How do How the whole application
176. 6:09works?
177. 6:10If someone asks you, let's say we have a
178. 6:12client who has this this requirement.
179. 6:14Now we have to deliver a solution to
180. 6:16him. So what should be your thought
181. 6:17process?
182. 6:18Right? So system design, API
183. 6:20development, databases, cloud, where you
184. 6:23deploy it, how your deployment should
185. 6:24happen, Jenkins, automatic build
186. 6:26deployment, testing, how you will test
187. 6:28it, AI-assisted development, how you
188. 6:30will use AI to maximize your work.
189. 6:33This is what what we should be focusing
190. 6:35on rather than AI bubble, right? That is
191. 6:38not our use case now. Everyone knows AI
192. 6:42AI is there, which is going to reduce
193. 6:44the number of people needing. It will
194. 6:47increase your productivity, right?
195. 6:50So what we can do? Again, once you are a
196. 6:52full-stack developer, what next thing?
197. 6:53Business requirement, understanding the
198. 6:55business, what solution is expecting.
199. 6:58Right? Solve the complex problem. If
200. 7:00validate whatever As I said, like you
201. 7:02know, no one is writing the code.
202. 7:04Everyone is writing just the prompt, and
203. 7:05it is asking. We are the role of just
204. 7:08validating the
205. 7:09We are the We have to play the role just
206. 7:12to validate the code what AI is
207. 7:13generating.
208. 7:14Right? So this is what is currently
209. 7:17expected from us.
210. 7:19Okay? So some of the things what you can
211. 7:21say.
212. 7:22Doesn't matter AI bubble burst or not.
213. 7:24AI raises the productivity, and AI cut
214. 7:27cut down the cost, you know, the head
215. 7:28count. Previously, we used to have big
216. 7:30team. Now it is very less.
217. 7:32Okay? Fewer developer will lead the
218. 7:34team. We don't need so much big team to
219. 7:37complete the project. Very small team
220. 7:39can do the job.
221. 7:41System design, understanding the
222. 7:43business, validating AI output. This is
223. 7:44what is expected from us now.
224. 7:46Okay?
225. 7:47So future belongs to the developer who
226. 7:50use who solve the business.
227. 7:52Deliver the complete code. No one is
228. 7:54expecting you to write the code. Yeah,
229. 7:56when it comes to interview, that thing
230. 7:58is still there. We will get by
231. 8:01bypassing that also. Right? But future
232. 8:03belongs to the developer who are
233. 8:05delivering the complete solution, not
234. 8:07just writing the code.
235. 8:09So, that should be our final goal that
236. 8:11yes, I can deliver the complete
237. 8:13solution. I can deliver the complete
238. 8:14business.
239. 8:15Right? Whether you talk about UI,
240. 8:17backend, deployment, I can take care of
241. 8:19that. That is the mindset, you know,
242. 8:21forward deploy engineer we are saying.
243. 8:23That is nothing but this.
244. 8:25So, don't worry about that AI bubble.
245. 8:28Focus on the development. Learn as much
246. 8:31as you can. If you are UI person, try to
247. 8:33learn a backend. If you are a backend,
248. 8:35try to learn the UI. If you are already
249. 8:37a backend, learn the cloud. If you are
250. 8:40If you already know the cloud, the AI
251. 8:41assistant, you know, rag application,
252. 8:43how do you use AI into your existing
253. 8:45deploy existing applications? Then, how
254. 8:47do you use AI to complete the task more
255. 8:50properly? You know, either you talk
256. 8:52about GitHub Copilot, Cloud, Wind sub,
257. 8:55lovable, cursor, spekit. So many things
258. 8:58are there. Wind sub.
259. 8:59Be aware how which model how which AI
260. 9:02agents are doing which the job better.
261. 9:05Right? How to get the work done with
262. 9:08limited number of token? How to add your
263. 9:09skill? How to create your own AI agent?
264. 9:12These are the things what you should
265. 9:13explore now. So, when the next interview
266. 9:15come, you will have so many things to
267. 9:17tell. Yeah, this much things I know
268. 9:19around the AI.
269. 9:21Right?
270. 9:22So, I hope this video was helpful to
271. 9:24you.
272. 9:25Uh
273. 9:26specific to this channel again like I'm
274. 9:28going to come up with the more AI
275. 9:29specific videos, you know, how do we
276. 9:31host LLM model locally and not. Right?
277. 9:34So, so many things I'm also exploring.
278. 9:37Once my exploration is done, then I will
279. 9:39be coming up with the videos also.
280. 9:41Thank you. Thanks for watching.