YouTube2Text

These 19 Jev-Claude use cases are blowing people's minds (with prompts) — Transcript

by Jay E | RoboNuggets · 3,086 words · 416 segments · language en · Watch on YouTube

Full transcript

  1. 0:00Jev is the new AI model that everyone's
  2. 0:02building with right now and for good
  3. 0:04reason because it can be up to 200 times
  4. 0:06faster and 400 times cheaper than other
  5. 0:08models. And so if you know exactly where
  6. 0:10to use it and how to use it, then it can
  7. 0:12definitely make your setup cheaper, make
  8. 0:14your systems faster, and ultimately
  9. 0:16transform your workflows. So today, I'll
  10. 0:18give you 19 use cases you can do with
  11. 0:20Jev along with the prompts to get you
  12. 0:22started with your builds so that by the
  13. 0:24end of this, you'll for sure have an
  14. 0:26idea of how to best use it. We'll start
  15. 0:28with the easy ones you can try today to
  16. 0:30some more advanced ones that can truly
  17. 0:32change how you build. And by the way, if
  18. 0:34you need a quick primer on what Jeb is,
  19. 0:35I'll link a previous lesson I did
  20. 0:37somewhere in this video that goes
  21. 0:38through it in just a few minutes. And if
  22. 0:40you're new, my name is Jay. I spent over
  23. 0:42a decade working with brands you
  24. 0:43probably know, have been in AI since my
  25. 0:45masters in data science. And now I'm
  26. 0:47leading our AI business in one of the
  27. 0:48largest AI communities globally. But for
  28. 0:50now, let's dive into it. Starting with
  29. 0:52the easy ones. Number one is to put Jev
  30. 0:54in Google Sheets. One of the best ways
  31. 0:56to feel how fast Jev is is inside a
  32. 0:59spreadsheet. Jev is a decision model,
  33. 1:01which means it can't write a sentence,
  34. 1:03but it can pick true or false, a choice
  35. 1:05from a menu, or a score, and it does
  36. 1:08that in a fraction of a second for
  37. 1:10almost nothing. So, if you have data in
  38. 1:12a sheet that needs categorizing, like a
  39. 1:14bank export, you can have Claude build
  40. 1:16you a plugin where you type a new column
  41. 1:18name, say category, and Jev fills in
  42. 1:20every row while you watch. It does this
  43. 1:22really quickly and because Jev can only
  44. 1:24pick from your list, it never invents a
  45. 1:26category that you didn't ask for. Here's
  46. 1:28a prompt to get you started. Number two
  47. 1:30is to triage customer inquiries. If you
  48. 1:33get a lot of customer questions, the
  49. 1:35first job is deciding what to do with
  50. 1:37each one. Specifically, which one can an
  51. 1:39automation handle and which one needs a
  52. 1:41real person? So, what you can do is to
  53. 1:42build an automation workflow where Jev
  54. 1:44is the decision layer. Instead of paying
  55. 1:46for a big language model to read every
  56. 1:48single message, Jev tags each inquiry.
  57. 1:51Let's say if it's for billing or if it's
  58. 1:52a bug or if it's a question and it also
  59. 1:55tells you how sure it is. So that
  60. 1:57confidence score that is the useful part
  61. 1:59because if you set a rule for that
  62. 2:01confidence score like for example if Jev
  63. 2:03is less than 60% sure let's say you can
  64. 2:05just have a rule where Jev sends that
  65. 2:07straight to a person then you can just
  66. 2:08adjust that confidence number threshold
  67. 2:10as you learn what works best for your
  68. 2:12business. In one independent test, 217
  69. 2:15customer call transcripts were sorted
  70. 2:17into categories by Jev in 13 seconds for
  71. 2:20only 7 cents, which is really cheap and
  72. 2:23can really speed up your customer
  73. 2:24handling workflows. And by the way, I've
  74. 2:26put every use case in this video into
  75. 2:28this free PDF guide, which you can just
  76. 2:30send it straight to your agent and start
  77. 2:31building. The link for that is in the
  78. 2:33description below if you need it. Number
  79. 2:34three is to analyze your competitor's
  80. 2:36ads. The meta ad library shows every ad
  81. 2:39your competitors are running, but nobody
  82. 2:41reads hundreds of them by hand. So, you
  83. 2:43can have Claude build you a plug-in or a
  84. 2:45Chrome extension that does it. It saves
  85. 2:47each ad for you. And Jev reads a text
  86. 2:50very quickly on every ad and tags it.
  87. 2:52What's the format? What's the call to
  88. 2:53action? What is the funnel stage? Then
  89. 2:55everything can land in a database. So,
  90. 2:57you end up with a swipe file of what's
  91. 2:59working in your market. When I tried it,
  92. 3:01Jev tagged 54 ads from three different
  93. 3:04brands in about a second and a half for
  94. 3:06less than a cent. Number four is on
  95. 3:08clipping. If you make long videos or
  96. 3:10podcasts, finding the moments worth
  97. 3:12turning into shorts can take hours if
  98. 3:14you do it manually. So, what you can do
  99. 3:16is to build a clip finder where when I
  100. 3:17tried it out, it scored 411 possible
  101. 3:20clips from a 70minute interview in 6 and
  102. 3:231/2 seconds. Then, it also picked the
  103. 3:26top 10 that is worth posting. So you can
  104. 3:28just record once and Dev would then hand
  105. 3:30you a short list of the best moments
  106. 3:32within that long- form video. Number
  107. 3:34five is to score and profile your
  108. 3:36customer base. If you run a software as
  109. 3:38a service platform, you already have a
  110. 3:40database of your customers, how long
  111. 3:41they've been with you, whether they're
  112. 3:43on an animal plan, whether they've
  113. 3:45started to cancel, Jev can read all of
  114. 3:47that and put each customer in a group,
  115. 3:49like which ones are healthy, which ones
  116. 3:50need attention or at risk of churn. So
  117. 3:53instead of finding out when someone
  118. 3:54cancels once it's late, you already know
  119. 3:56who to reach out to this week. Number
  120. 3:58six is to build back links inside your
  121. 4:00website or your second brain system.
  122. 4:02This one is for anyone with a huge
  123. 4:04website or a big workspace where your
  124. 4:05agents are working in. If you have a
  125. 4:07site, you want your pages linking to
  126. 4:09each other because it helps people find
  127. 4:10their way around and it also helps a bit
  128. 4:12with your SEO. And if you have a second
  129. 4:14brain system, which is basically a huge
  130. 4:16folder of notes, the same links would
  131. 4:18help your agents find their way around
  132. 4:20it. When I tried it on 60 test blog
  133. 4:22posts, it took only two and a half
  134. 4:23seconds to link those pages together.
  135. 4:26And the exact same trick would also link
  136. 4:27up your notes so your second brain
  137. 4:29system would connect itself instead of
  138. 4:31assigning that to a big language model
  139. 4:33to do that for you. Number seven is to
  140. 4:35find buyers in your comments. If you
  141. 4:37sell on social, your buyers are already
  142. 4:39in your comments, but it's just buried
  143. 4:41between everything else. Take a seller
  144. 4:43account on Tik Tok for example. Every
  145. 4:45video they have brings hundreds of
  146. 4:46comments. And what you can do is to pull
  147. 4:48those comments through a scraping tool
  148. 4:50like Appify for example. Then have Jev
  149. 4:52automatically tag each one. Which ones
  150. 4:54are ready to buy? Which one is a
  151. 4:56question? Which one is a complaint? So
  152. 4:57instead of manually scrolling through
  153. 4:59your comments, you would then get a
  154. 5:00priority list with the people ready for
  155. 5:02you to buy right at the top. Number
  156. 5:04eight is about verification of Jev's
  157. 5:06results. Before we go to the rest of the
  158. 5:08use cases, one of the things you should
  159. 5:09learn is how to verify if Jev's
  160. 5:11categorization is correct for you. One
  161. 5:13way to do that is through what's called
  162. 5:15a validation set, which is just a few
  163. 5:17dozen or hundred examples depending on
  164. 5:19your use case where you already know the
  165. 5:21right answer or where a person already
  166. 5:22did the sorting by hand. So what you
  167. 5:25essentially do is you run jev on that
  168. 5:27set without showing it the answers
  169. 5:28first, then count how often it matches.
  170. 5:31You then look at where Jev got things
  171. 5:33incorrectly because that is where you
  172. 5:35can calibrate the model to be more
  173. 5:36correct next time. So, those are the
  174. 5:38easy ones and you can see the pattern
  175. 5:40there is basically you point Jev at a
  176. 5:41pile of stuff you already have and you
  177. 5:43just have Jev categorize those at
  178. 5:45lightning speed. Now, let's get to some
  179. 5:46of the more intermediate use cases. And
  180. 5:48by the way, if you want to learn how to
  181. 5:50build and sell AI systems that
  182. 5:51businesses actually pay for, then that's
  183. 5:53pretty much all we do over at the
  184. 5:55Robberonuggets community, where not only
  185. 5:56do you get access to the Clawed Living
  186. 5:58Master Class, which we update every week
  187. 6:00and takes you from zero to mastery with
  188. 6:02the latest on AI, but you also get
  189. 6:03access to our agents as a service
  190. 6:05course, which walks you through how to
  191. 6:07actually get paid for all these AI
  192. 6:09skills that you are learning. You also
  193. 6:10get to be part of a genuinely great
  194. 6:12community of AI builders. In fact, you
  195. 6:14can see just some of the recent wins our
  196. 6:16members are getting from the program
  197. 6:17right here. So, if you want to start
  198. 6:18earning from AI, then check that just in
  199. 6:20the pin comment below. Now, back to the
  200. 6:21video. Number nine is letting Jev pick
  201. 6:24the right skill. If you use Claude Code
  202. 6:26or another agentic platform, you
  203. 6:27probably have dozens of skills and
  204. 6:29Claude sometimes takes a lot of time to
  205. 6:31load the right one. So, what you can do
  206. 6:33is let Jev pick it. Once your prompt
  207. 6:35goes in, your skills basically becomes
  208. 6:37the menu of options and Jev would pick
  209. 6:39the right one at really fast speeds. In
  210. 6:42my own workspace, I have something like
  211. 6:44145 skills. And across 14 tests, Jev
  212. 6:47found the right skills in about 5
  213. 6:48seconds in total, while Opus 5 took
  214. 6:51around 30. The company behind Jev,
  215. 6:53TypeSafe, also tested this with 182
  216. 6:55skills, and it cut the wrong picks by
  217. 6:57more than half. Number 10 is to let Jev
  218. 7:00pick the right model. So, this is the
  219. 7:02same idea, but for model routing. So,
  220. 7:04what you can do is to make a /jv command
  221. 7:06in Claude. And when this is on, Jev
  222. 7:08would read every message that you send
  223. 7:10and pick the model for it in about a
  224. 7:12third of a second. So for example, when
  225. 7:14I ask it to find a file path, Jev handed
  226. 7:16that to a haiku helper, which is the
  227. 7:18cheapest model in Claude instead of
  228. 7:20Opus. And when the job is hard, it goes
  229. 7:22to the bigger model. So you only pay for
  230. 7:24the big model when the job actually
  231. 7:26needs it. And this same model routing
  232. 7:27idea works even if you are not using
  233. 7:29Claude. That also works in Codeex or
  234. 7:31other agentic platforms. Number 11 is to
  235. 7:34put Jev in front of your AI agents. if
  236. 7:36you have an AI agent answering your
  237. 7:38emails like these, you're paying a big
  238. 7:40model to read every single one, which
  239. 7:42includes even spam. So, what you can do
  240. 7:44is put Jev in front of it. In one test,
  241. 7:46Jev sorted 1,700 emails for 18 cents,
  242. 7:49where Jev tagged what needs a reply,
  243. 7:51what's a brand deal, and even what's a
  244. 7:53scam or a spam. Then only the ones that
  245. 7:56actually need writing or replying go to
  246. 7:58Claude, so your agent reads less, and
  247. 8:00your bill for tokens also drops. Number
  248. 8:0312 is to build Chrome extensions. You
  249. 8:05can also build Chrome extensions using
  250. 8:07Jev. And one fun example I saw is about
  251. 8:09cleaning up your X feed. Someone built
  252. 8:11one that checks every post as it loads
  253. 8:13and Jev answers one question if this is
  254. 8:16AI slop or not. If it is, the post gets
  255. 8:19folded away. So what's left is the stuff
  256. 8:21that is worth reading. The same trick
  257. 8:23works anywhere, by the way, where
  258. 8:24there's a feed like for example LinkedIn
  259. 8:26or any page. Or let's say if you want to
  260. 8:28build an extension where you want to
  261. 8:30hide the ads and cookie banners in a
  262. 8:31page like this Chrome extension called
  263. 8:33Unclutter. So you can launch plugins
  264. 8:35like these that reshape what you see in
  265. 8:37the web using Jev. Number 13 is search
  266. 8:40by meaning features. Now when you do
  267. 8:42Ctrl+F that only finds usually the exact
  268. 8:45words that you type. But with Jev you
  269. 8:47can actually build a find feature where
  270. 8:49you type roughly what you're looking for
  271. 8:50and it would highlight the right part of
  272. 8:52the page even when the words don't
  273. 8:55match. It would split the page into
  274. 8:56paragraphs, asks Jeff which one matches
  275. 8:59best and scrolls straight to it. It's
  276. 9:01fast enough to feel real time. So, if
  277. 9:03your app or your website has a search
  278. 9:05box, this is one of the easiest upgrades
  279. 9:07that you can give it. Number 14 is a
  280. 9:09related one, which is searching your
  281. 9:11images by what is in them. In my
  282. 9:13workspace, I generate a lot of images
  283. 9:15and videos, and searching them by file
  284. 9:17name barely works. So, what I built is a
  285. 9:19search powered by Jev, where typing
  286. 9:21keyword, let's say Claude, would find
  287. 9:23the images that are actually about
  288. 9:24Claude, not just the ones with Claude in
  289. 9:26the file name. The one thing you need
  290. 9:28though is some written words about each
  291. 9:30image which we call metadata. In my
  292. 9:33case, I log the prompt for every image
  293. 9:35that I generate. And so that is my
  294. 9:37metadata. But if you don't have that, a
  295. 9:39really small cheap model can write a
  296. 9:41short caption for each one. What you can
  297. 9:43try out is Gemini Flash Light because
  298. 9:45captioning a,000 images costs about only
  299. 9:4850. So those are the intermediate ones
  300. 9:51and they already make your setup faster
  301. 9:52and cheaper. Now let's go to some of the
  302. 9:54advanced ones. Number 15, a live checker
  303. 9:57for what people say. Because Jev answers
  304. 10:00in a fraction of a second, it can check
  305. 10:02what people say while they are still
  306. 10:04saying it. Someone actually built a live
  307. 10:06BS meter called Jev meter, which scores
  308. 10:09every sentence of a debate as it's
  309. 10:11spoken. You can also build a plug-in,
  310. 10:13let's say, for your meetings where every
  311. 10:15sentence gets sorted as it is said into
  312. 10:17decisions, which ones are action items,
  313. 10:20which ones are risks, and which ones are
  314. 10:21questions. So, if you're on a sales
  315. 10:23call, you'll get easy notes to recap as
  316. 10:26you go. And it works for interviews and
  317. 10:28even podcast conversations as well.
  318. 10:31Number 16 is a chatbot with zero LLM
  319. 10:33calls. This one sounds implausible. A
  320. 10:36chatbot that never really calls a
  321. 10:37language model, but here's how it can
  322. 10:39work. In this demo app, you can ask
  323. 10:41which of our lesson teaches a certain
  324. 10:43topic. Let's say the person is curious
  325. 10:44about Jev as a topic. In the back end,
  326. 10:47Jev the model then quickly answers that
  327. 10:49in real time because it has access to a
  328. 10:51list of my videos that are already
  329. 10:53pre-transcribed, informing Jev of what
  330. 10:56each video is about. So, if you need a
  331. 10:58narrow chatbot that can answer things
  332. 11:00quickly from a given data set, like
  333. 11:02video transcripts in my case, or even
  334. 11:04company documents in most corporations,
  335. 11:06then Jev is a potential tool for that.
  336. 11:09Number 17 are smarter designs for apps.
  337. 11:12Here's a simple example for apps that
  338. 11:14rely heavily on visuals like icons. When
  339. 11:16a user types something, Jev can pick
  340. 11:19from a given set of icons. So, let's say
  341. 11:21in a habit tracker, you type drink more
  342. 11:23water and it picks the water icon for
  343. 11:25you. In a mood app, you write how your
  344. 11:27day went and it picks a custom mood
  345. 11:29icon. A designer actually showed this
  346. 11:31off where as he's typing, Jev is pulling
  347. 11:33up the icons relevant to that text. And
  348. 11:36the thing is before Jev to do this
  349. 11:37effectively you need either an LLM which
  350. 11:40can be costly or at least much costlier
  351. 11:42than Jev or to build out something like
  352. 11:43a reject based matching system which is
  353. 11:46sometimes not as effective versus a
  354. 11:48language model. But with a decision
  355. 11:50model like Jev this can be done for so
  356. 11:52much faster and for so much cheaper.
  357. 11:55Number 18 are web pages that build
  358. 11:57themselves. Now Jev can't write code but
  359. 11:59it can pick parts. Someone actually
  360. 12:01built a demo where Jev would get a
  361. 12:03library of readymade pieces like forms,
  362. 12:06buttons, signin boxes, and fonts, and it
  363. 12:08decides what each visitor needs and puts
  364. 12:11the page together on the fly. And the
  365. 12:13page shows up about as fast as a normal
  366. 12:15page load because Jev is only choosing
  367. 12:17parts and not actually building them.
  368. 12:19So, every visitor could get a page
  369. 12:21that's built for them in real time.
  370. 12:23Number 19 is about finding your own Jev
  371. 12:26use cases. This last one is a bit of a
  372. 12:28meta use case because you can actually
  373. 12:30use jev to find your own use cases.
  374. 12:32There's a site called shipwithjv.com
  375. 12:35which lists Jev use cases by category.
  376. 12:38So what you can do is to open claude,
  377. 12:40paste in that site along with a file
  378. 12:41about yourself and your business and
  379. 12:43just ask based on everything you know
  380. 12:45about me which of these use cases would
  381. 12:47help me the most. You can then ask
  382. 12:49Claude to use Jev in order to categorize
  383. 12:51which of those use case are most useful.
  384. 12:53So you get a starting point. Another
  385. 12:55option is to ask Claude to look back
  386. 12:57over your past sessions and actually
  387. 12:59point out the steps where a quick Jev
  388. 13:01call would have done the job. That way
  389. 13:03you would find use cases in the work
  390. 13:05that you are already doing. So there you
  391. 13:07go. That is all of them. And as a final
  392. 13:10note, OpenAI in their dev day a few days
  393. 13:12ago actually just launched their
  394. 13:14decisions API where you give it a
  395. 13:16question and a list of allowed answers
  396. 13:18and it picks one in about 150
  397. 13:20milliseconds. And the reason why that's
  398. 13:22familiar is because that's basically
  399. 13:23their answer to Jev. And the point I'm
  400. 13:25making is even if you don't end up using
  401. 13:27Jev itself for any of these use cases,
  402. 13:29getting good at this way of building
  403. 13:31makes sense now because for sure the
  404. 13:33other big labs starting with OpenAI,
  405. 13:34probably Entropic 2 in the future will
  406. 13:37be copying something like this too. As
  407. 13:39usual, thanks for watching till the end.
  408. 13:40Again, that free PDF guide with every
  409. 13:42prompt and resource from this video to
  410. 13:44get you started is just linked in the
  411. 13:45description. And uh I'm curious like
  412. 13:47which of these would you build first?
  413. 13:48And do you have other use cases in your
  414. 13:50company that I haven't covered yet? Let
  415. 13:52me know down below and I'll see you all
  416. 13:54next time. Cheers. [music]

About this transcript

This page contains the full transcript of These 19 Jev-Claude use cases are blowing people's minds (with prompts) by Jay E | RoboNuggets, generated from the public captions YouTube serves with the video. The transcript has 3,086 words across 416 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.