Latest / Randy Unscripted / The Future is Collaborative - NaPodPoMo 2025 #15

Transcript

The Future is Collaborative - NaPodPoMo 2025 #15 · Randy Unscripted

• 0:02
  Randy Black: Randy Black.
• 0:04
  Randy Black. Randy Black. Randy Black. Randy Black. Randy Black. Randy Black. Black. He's a troublemaker. Welcome back to Randy Unscripted,
• 0:12
  and I'm your host, Randy Black. And, again, another episode in our napod promo 2025
• 0:18
  series,
• 0:19
  national podcast post month.
• 0:22
  This is episode 15 for November 15 halfway through the month.
• 0:27
  It's all part of my daily series exploring creativity and technology and all kinds of other stuff throughout the month of November. The goal is to get
• 0:35
  30 episodes
• 0:36
  in thirty days.
• 0:38
  Yesterday,
• 0:39
  on our fourteenth episode for Napod Pummel, we looked at doctor Craig Van Slyke's reminder that AI isn't hard. It's human.
• 0:47
  But today,
• 0:48
  we move into the second half of his pod indie workshop where he introduced something that caught
• 0:54
  everyone's attention,
• 0:56
  and that was meta prompting.
• 0:58
  Craig Van Slyke: How do you get from this open iterative thing to the template?
• 1:01
  You use something called
• 1:04
  meta prompting.
• 1:06
  Meta prompting.
• 1:07
  This this is gold.
• 1:10
  This is really, really gold.
• 1:13
  You ask AI
• 1:15
  to write the prompt for you.
• 1:18
  Randy Black: Wow.
• 1:20
  That's where we left off yesterday, and that line made the room
• 1:23
  pause.
• 1:25
  It flipped the whole idea of prompting on its head. Instead of crafting the perfect question,
• 1:32
  you let the AI model
• 1:34
  draft it for you.
• 1:36
  Craig then spent several minutes unpacking the idea before showing how it worked in practice. Like, anybody use deep research?
• 1:45
  Craig Van Slyke: I'm gonna show you meta prompting here in a second using deep research. If you're not using deep research and you need to perform research, you need to be using deep research.
• 1:54
  And I'll I'll show you here in a second. So,
• 1:56
  ChatchiBT has it.
• 1:59
  I think Claude has it. Gemini has it. I think Grok has it. So most of the big players have this thing called deep research.
• 2:07
  I'm gonna use Gemini,
• 2:09
  and Gemini is just gemini. Google.com. We'll see how big we can make this.
• 2:14
  All right. We've jujitsu ed a lot. Somebody else give me another topic. Surfing? Where did that come from? We're in Indiana. All right. Good. I like random.
• 2:22
  Okay. What kind of a deep research? Okay. I
• 2:26
  want to do a deep research report
• 2:29
  on the history of surfing
• 2:31
  and how it's spread around the world. I also want to know about where the best surfing hotspots are and who some of the most famous
• 2:40
  surfers are in history.
• 2:42
  Also, I want to find out what sort of spectacular accidents there have been in surfing. Anything else?
• 2:49
  I don't know where that came from, by the way. What else should we put in there? Anything?
• 2:53
  I also want to know what the chances are of shark attacks while surfing by location and by time of the year. Also,
• 3:02
  I want the deep research report
• 3:05
  to give me
• 3:06
  information
• 3:07
  about the best types of surfboards
• 3:10
  broken down by the skill of the surfer
• 3:13
  and the size of the surfer. I know I ain't using a small surfboard. So alright.
• 3:18
  Please write a prompt that will create this report.
• 3:22
  Randy Black: And that's exactly what Gemini did. It crafted the report. We built this request live, and I clipped parts of that out where you couldn't hear the folks in the audience giving him information and stuff. But we we put this in asking for a full
• 3:37
  research report
• 3:38
  about surfing, created it created it inside Gemini using this very technique he had just described to us. And then we went on, and we did other stuff. We worked on other things.
• 3:49
  And it was probably an hour or so later when we came back to it, and Craig pulled off showing us the report that had been generated.
• 3:56
  Craig Van Slyke: This is a
• 4:00
  22 page, 6,502
• 4:02
  word report,
• 4:03
  and
• 4:05
  it is footnoted.
• 4:07
  Our part in creating this was
• 4:10
  three minutes maybe.
• 4:12
  And you you don't have to leave the chat open. You can go start another chat, and it'll work in the background.
• 4:17
  And you can do this on virtually any topic you can imagine.
• 4:21
  Randy Black: The audience that I was a part of and others, we saw this happen.
• 4:26
  We couldn't believe it. A multi page footnoted report generated
• 4:30
  in a matter
• 4:31
  of moments.
• 4:33
  Craig Van Slyke: Anything that's really important,
• 4:35
  you wanna verify.
• 4:37
  I mean, you you just do. Now what a lot of them will do now is they will give you links
• 4:42
  to the source material,
• 4:44
  so you can go out and look to see if it actually says what it thinks it says.
• 4:48
  But they're better. They still hallucinate. They still make stuff up, but they're better. But anything that really matters,
• 4:54
  you you wanna go out and check it before you put it out there in the world.
• 4:58
  Randy Black: Even when the tech performs flawlessly,
• 5:01
  Craig reminds us that accuracy
• 5:03
  still depends on the human in the loop. AI
• 5:07
  is not the be all end all. Things still need to be verified because it still will make mistakes.
• 5:13
  Craig Van Slyke: So we don't need to go on and on about it, but but this is kinda neat. Right? In in
• 5:17
  less than five minutes, we could have an entire season, but you could take this
• 5:23
  and use it to create a bunch of blog posts or newsletter articles or, you know, whatever. And I've had these things, these deep research reports run from about eight or 10 pages to,
• 5:33
  like, 75 or 80, depending upon the topic. The idea wasn't to replace the creativity,
• 5:40
  Randy Black: but it was to use AI
• 5:41
  to build the framework
• 5:43
  so we can then invest our energy into the content itself to find better ways
• 5:49
  to produce content for whatever it is we're doing. Look. This stuff is not hard.
• 5:54
  Craig Van Slyke: Once you start doing it, once you start playing around with it, you'll find your own tricks. I didn't read about this meta prompting thing.
• 6:01
  I just thought, how can I not waste these deep research? Oh, I know. I'm gonna try this, and it worked. So now I do it a lot. If it didn't work, I'd try something else. That closing thought captured
• 6:12
  Randy Black: the spirit of this entire workshop.
• 6:14
  Learn by doing,
• 6:16
  and don't be afraid
• 6:18
  to experiment.
• 6:20
  If you fail at it, it's okay. It goes back to something we've talked about previously.
• 6:25
  Failure is feedback, and feedback is the breakfast of champions.
• 6:30
  So across these two episodes for the last two days, doctor Craig Van Slyke has showed how meta prompting can turn AI into a true creative partner for any content creator.
• 6:41
  And the power it holds in helping you to streamline your workflow, to do things faster, to do things easier is just amazing.
• 6:50
  So this has been episode 15
• 6:52
  of my napod homo twenty twenty five series on 11/15/2025.
• 6:58
  Remember, the goal is
• 7:00
  30 podcast episodes published
• 7:03
  in thirty days. And if you're a content creator, if you're a podcaster, you need to jump on board. It's okay if you jump on board in the middle. Start on day 16 and work your way to the end of the month and keep it going.
• 7:14
  You can find notes about this
• 7:16
  entire event over on my website. I'll make a short link to it so you can find it easier. So just head to randallblack.com/podindy2020five,
• 7:25
  and you can find the blog post that I wrote up with the assistance of AI through transcribing and and things to put that together and then built these two podcast episodes off of that. But remember, until the next time we're together here on Randy unscripted,
• 7:39
  which will be tomorrow,
• 7:41
  Keep learning,
• 7:43
  keep experimenting,
• 7:44
  and keep being curious.