WEBVTT

00:00:00.240 –> 00:00:06.300
Hey, everybody and welcome to a new type of episode. I’m trying called the Friday wrap-up on

00:00:06.640 –> 00:00:11.740
Streamline Solopreneur. A short episode where I talk about three things. What’s on my mind this week?

00:00:11.980 –> 00:00:20.060
Recommended reading and recommended media. This show is designed to help you automate your business

00:00:20.180 –> 00:00:26.240
so you can take time off worry-free and hopefully this curation will help you think more about your

00:00:27.120 –> 00:00:29.980
systems. I’m your host, Joe Casabona. And here’s what you’re

00:00:30.000 –> 00:00:36.040
on my mind. First, I’ve been seeing more types of shows like this, and I’ve always liked the

00:00:36.280 –> 00:00:43.020
idea of a weekly news kind of based podcast. I wanted to start adding recommended reading

00:00:43.700 –> 00:00:51.280
to my full-length Monday episodes, but because of the way I batch them, most of the articles I

00:00:51.480 –> 00:00:59.960
recommend would be a month old by the time they come out. But by the time the episode comes out. So

00:00:59.980 –> 00:01:04.280
this, I’m calling it the Friday wrap-up. I’m thinking I would record it either Thursday

00:01:04.559 –> 00:01:09.980
evening or Friday morning. And it’s just about some of the things that happened this week

00:01:10.180 –> 00:01:16.440
that I’m thinking about. So here’s what’s on my mind. The first is I had a pretty lively

00:01:16.780 –> 00:01:23.860
discussion in a community I’m a part of where someone shared their process for using AI

00:01:24.240 –> 00:01:29.940
agents to basically write long form articles for them. And it is the

00:01:29.960 –> 00:01:36.120
process is, you know, they train this skill or project or agent or whatever in Claude.

00:01:36.800 –> 00:01:46.320
And it comes up with ideas and then the person approves the ideas or provides feedback on the

00:01:46.540 –> 00:01:55.100
ideas. And anything that he approves, the large language model will write the first draft

00:01:56.320 –> 00:01:58.020
and then the person will provide feedback

00:01:59.260 –> 00:02:03.920
and the large language model will rewrite the first draft based on the feedback

00:02:04.950 –> 00:02:07.440
and from what this person says they’ll go back and forth,

00:02:07.600 –> 00:02:13.880
it’s based on voice notes and then the large language model will publish the article.

00:02:15.000 –> 00:02:15.640
And I said,

00:02:17.960 –> 00:02:21.260
at what point would you feel uncomfortable putting your name on this?

00:02:21.320 –> 00:02:23.040
And I feel really strongly about this.

00:02:23.100 –> 00:02:24.520
I’m actually working on a long form.

00:02:25.680 –> 00:02:31.500
episode, like a full episode about this idea and the negative ramifications of it.

00:02:32.530 –> 00:02:43.960
But I just, I feel like if no matter how well you train an AI or a large language model,

00:02:44.180 –> 00:02:50.260
right, a word calculator, if it’s coming up with the ideas and writing all the words,

00:02:50.430 –> 00:02:52.340
it doesn’t matter how much feedback you give it.

00:02:52.400 –> 00:02:53.600
You didn’t come up with this.

00:02:54.580 –> 00:03:02.460
and I strongly feel like if you can’t be bothered to sit down and write the messy first draft,

00:03:03.900 –> 00:03:06.720
then why should anybody be bothered to read it?

00:03:09.940 –> 00:03:14.180
It’s a hill I’m willing to die on, I realized, this week.

00:03:16.180 –> 00:03:18.660
And that’s the biggest thing that’s on my mind.

00:03:18.780 –> 00:03:23.600
It’s dominated the last couple of days of my thinking because I fear that we’re going to

00:03:23.620 –> 00:03:30.440
to see more of this. And there are people who will say, well, I trained it to sound like me. And

00:03:30.470 –> 00:03:36.000
like, that’s cool, but, you know, Jay Moore does a great impression of Christopher Walkin. That

00:03:36.200 –> 00:03:41.980
doesn’t make Jay Moore Christopher Walken. And nobody would hire Jay Moore for a role that they want

00:03:42.440 –> 00:03:49.180
Christopher Walken for. Or they’ll say, well, you know, Claude is a better writer than me anyway.

00:03:50.060 –> 00:03:58.880
And okay, I’m going to guess that’s not true, but also you don’t get better at something by not doing it.

00:04:00.100 –> 00:04:05.040
And so, like, what you’re really saying in that instance is I don’t care about getting better at writing.

00:04:05.280 –> 00:04:07.820
I just care about producing content.

00:04:09.160 –> 00:04:16.320
So I’ll have more on this in a future long form episode, but that’s, if you’re wondering, if you listen to that and you’re like, where did this come from?

00:04:16.579 –> 00:04:17.899
The nexus was from this week.

00:04:18.840 –> 00:04:25.500
Now, the second thing on my mind is I did upgrade to Claude Max over the last week,

00:04:25.500 –> 00:04:27.440
and so I’ve spent about a week using it.

00:04:28.700 –> 00:04:33.300
And I use it for computer-based problems.

00:04:34.080 –> 00:04:45.360
I use it to crunch data or review things I’ve written to see if it adheres to this criteria

00:04:45.640 –> 00:04:47.180
that I’ve set up for what I’m writing.

00:04:48.480 –> 00:04:52.820
I used it for a podcast SEO for this show.

00:04:53.600 –> 00:04:57.360
And the way I did that was I took Courtney Elmer’s framework.

00:04:58.340 –> 00:05:03.380
I had it do deep research, a very specific prompt.

00:05:03.400 –> 00:05:08.200
And then I reviewed the 20 shows that it surfaced for me to make sure it was good.

00:05:08.260 –> 00:05:12.000
And I told it to throw out ones that were not good, find other ones.

00:05:12.740 –> 00:05:18.060
And then I uploaded two different sets of analytics for my show, one from my host and one from

00:05:18.599 –> 00:05:25.120
from a third party and I basically said crunch this data and give me your observations.

00:05:26.020 –> 00:05:29.360
And I also reviewed the data and made some observations.

00:05:30.560 –> 00:05:32.080
And it did an incredible job at that.

00:05:32.600 –> 00:05:34.520
And so now I have these 10 keywords.

00:05:35.460 –> 00:05:38.120
I wrote the description and a bunch of other stuff.

00:05:38.240 –> 00:05:44.960
And then I asked the AI like, hey, is this using the keywords without like stuffing them?

00:05:47.080 –> 00:05:50.960
and now I’ll give it copy and I’ll say like, hey, are there enough keywords here?

00:05:52.040 –> 00:05:55.180
Does it sound natural? Never make it sound unnatural.

00:05:55.920 –> 00:06:00.340
But the important part here is that I’m using it to analyze things I’m doing.

00:06:00.740 –> 00:06:02.620
I’m not having it do the things for me.

00:06:03.640 –> 00:06:09.600
But what’s really impressive to me about Claude Max is obviously the co-work side of things.

00:06:11.080 –> 00:06:14.580
I’ve had it do a bunch of stuff like create skills.

00:06:16.620 –> 00:06:24.540
My son plays soccer and the calendar that they provide is not adjusted for time zone or daylight savings time, which is insane.

00:06:25.440 –> 00:06:38.120
So I had Claude update the calendar to include time zone and daylight savings time so that I have the accurate times on my calendar without having to go and look in the portal.

00:06:38.780 –> 00:06:44.720
Something I did this morning was the exact prompt.

00:06:45.300 –> 00:06:47.220
Oh, I did this on my other machine.

00:06:48.520 –> 00:06:52.000
But it was basically, I want you to find the Yankees schedule,

00:06:53.340 –> 00:06:57.940
create a calendar for me, a separate calendar that I can subscribe to,

00:06:58.500 –> 00:07:05.400
and then grab all the weekday afternoon games and put them on my main calendar

00:07:06.420 –> 00:07:10.780
so that people don’t book meetings with me during game time.

00:07:12.140 –> 00:07:15.240
And it hit a little like sandbox snafu that I,

00:07:15.280 –> 00:07:22.520
had to fix, but it did it flawlessly. And again, I checked it. I looked at the Yankee schedule. I downloaded a

00:07:22.600 –> 00:07:30.440
CS, I downloaded a spreadsheet of all the Yankee games and I filtered by day and time. And it did it

00:07:30.740 –> 00:07:36.840
flawlessly. And so like that’s really impressive. That saved me a ton of time having to look up the

00:07:36.900 –> 00:07:43.139
Yankee schedule and then manually add those games. So, and that’s a very computer thing, right? Look at this

00:07:43.140 –> 00:07:50.100
data do other stuff. It’s computer talking to computer. And so I’m excited to continue to use

00:07:50.540 –> 00:08:00.980
Claude. A lot of the tools I use have like an MCP. And I like that it can crunch data for me

00:08:01.820 –> 00:08:09.060
in the background while I can actually focus on the creative work, which is my whole goal for

00:08:09.060 –> 00:08:12.440
this podcast, right? Help you automate things that you don’t have to do.

00:08:13.220 –> 00:08:17.760
so you can do the things that you have to do or want to do.

00:08:18.980 –> 00:08:19.960
So that’s what’s on my mind this week.

00:08:20.960 –> 00:08:22.880
Recommended reading, I have two articles.

00:08:23.760 –> 00:08:24.820
One is mine.

00:08:25.180 –> 00:08:26.620
Based on everything that’s on my mind,

00:08:26.880 –> 00:08:28.360
I feel like I should mention.

00:08:28.820 –> 00:08:31.840
The three question test for using AI effectively,

00:08:33.860 –> 00:08:39.620
I basically go through why most people are using AI wrong.

00:08:40.700 –> 00:08:49.560
Because when the test is it’s doing something so I don’t have to, that’s not the correct thing to test on.

00:08:50.640 –> 00:08:54.800
Because it doing something incorrectly doesn’t help you achieve your goals.

00:08:56.040 –> 00:09:07.180
So speaking of this person at the top of the show, I’d love to know how many of those articles that he publishes actually help him.

00:09:07.900 –> 00:09:09.140
Whatever his goal is,

00:09:09.410 –> 00:09:09.860
what rank,

00:09:10.240 –> 00:09:10.400
you know,

00:09:10.480 –> 00:09:12.280
maybe rank in SEO is fine,

00:09:13.100 –> 00:09:14.860
but get people to join his mailing list,

00:09:15.020 –> 00:09:16.160
get people to hire him.

00:09:17.220 –> 00:09:18.520
What’s the point of the writing?

00:09:19.800 –> 00:09:20.920
And if it’s thought leadership,

00:09:21.220 –> 00:09:22.460
is it really thought leadership

00:09:22.700 –> 00:09:23.980
if you’re not doing the thinking?

00:09:25.480 –> 00:09:27.540
So I have a three question test.

00:09:27.900 –> 00:09:29.780
I will link the article

00:09:30.740 –> 00:09:32.580
and everything I’ve mentioned here in the show notes.

00:09:34.200 –> 00:09:37.820
And then the second article is by Jason Snell

00:09:37.840 –> 00:09:44.160
colors. It’s called rethinking RSS newsletters and how I read every morning. This really

00:09:44.400 –> 00:09:49.500
resonated with me. This was kind of the impetus for me starting this format episode.

00:09:50.720 –> 00:09:54.920
Because I constantly think of this. I’ve tried a bunch of RSS readers. I currently have

00:09:55.020 –> 00:10:02.760
Feedly and I don’t really like it. It’s like a fire hose. And so I’ve moved to FeedBin with an

00:10:02.660 –> 00:10:10.240
app called Current. Feed bin, like a killer feature for me is being able to forward newsletters

00:10:11.160 –> 00:10:16.720
to my RSS reader so that all of my reading is in one spot. And so I just think this is a really good

00:10:16.920 –> 00:10:22.180
article. Jason Snell obviously thinks about this stuff a lot. His process, I heard about Current here

00:10:22.330 –> 00:10:29.420
in this article. He doesn’t use it. Spoiler alert, but I like that he thinks about this because I

00:10:29.740 –> 00:10:32.620
think about this a lot. Like what’s the way for me to most effective?

00:10:32.640 –> 00:10:38.740
read and consume news in a way that’s not overwhelming that doesn’t feel like a to-do list.

00:10:38.980 –> 00:10:43.640
So really enjoyed that, rethinking RSS newsletters and how I read every morning by Jason

00:10:43.860 –> 00:10:44.779
Snell on Six Colors.

00:10:45.900 –> 00:10:49.140
And then the recommended media, this one really pushes all of my buttons.

00:10:49.620 –> 00:10:54.600
It’s called, I made a network of every home run in MLB history.

00:10:55.560 –> 00:11:01.640
It is a YouTube video that visualizes every home run ever hit.

00:11:02.500 –> 00:11:07.680
And even if you are not into baseball, if you are into data and data visualization,

00:11:08.920 –> 00:11:12.560
this video is well worth the watch.

00:11:14.060 –> 00:11:19.820
I think it’s so interesting his approach to how he did it and the kind of conclusions

00:11:20.460 –> 00:11:25.060
and like hidden history that he can find in this.

00:11:26.440 –> 00:11:28.360
and some of the

00:11:30.340 –> 00:11:31.020
some of the

00:11:31.880 –> 00:11:33.740
stories that are told with the data

00:11:33.950 –> 00:11:34.800
I really like.

00:11:35.760 –> 00:11:38.180
So I made a network of every home run in MLB history.

00:11:38.340 –> 00:11:39.020
It’s over on YouTube.

00:11:39.280 –> 00:11:40.420
Again, I’ll link it in the show notes here.

00:11:41.160 –> 00:11:43.140
But that’s it for the Friday wrap-up.

00:11:43.760 –> 00:11:45.540
If you liked this episode,

00:11:46.960 –> 00:11:51.620
let me know by going to streamlinedfeedback.com.

00:11:52.640 –> 00:11:53.660
Because ultimately,

00:11:55.120 –> 00:12:02.920
The thing that will motivate me to keep doing these is if people are actually interested in them.

00:12:04.280 –> 00:12:14.620
And if you want to learn how to create time and space so you can actually focus on the work that matters, that is the main thing I do.

00:12:14.740 –> 00:12:16.620
It’s the main thing I talk about on this show.

00:12:17.360 –> 00:12:19.080
You should check out the Solopreneur sweep.

00:12:19.280 –> 00:12:22.200
It’s my process for understanding how you work.

00:12:22.880 –> 00:12:26.720
and you can get that over at streamlined.fm slash sweep.

00:12:27.600 –> 00:12:29.280
But that’s it for the Friday wrap-up.

00:12:29.780 –> 00:12:31.100
Thanks so much for listening.

00:12:31.900 –> 00:12:35.900
And until next time, I hope you find some space in your weekend.