The AI Assistant: Automating Administrative Friction and “Shadow Work”

Voiceover Artist | 00:00

Are you ready to manage your work and personal world better to live a more fulfilling, productive life? Then you’ve come to the right place. Welcome to ProductivityCast, the weekly show about all things personal productivity. Here are your hosts, Ray Sidney Smith and Augusto Pinault with Frances Wade and Art Gelwix.

Raymond Sidney-Smith | 00:18

Welcome everybody to productivity cast the weekly show about all things personal productivity. I’m Ray Sidney Smith.

Francis Wade | 00:25

And I’m Francis Wade.

Raymond Sidney-Smith | 00:28

Welcome, gentlemen, and welcome, everyone. To ProductivityCast this week. We’re diving into the world of artificial intelligence and its growing role in our personal productivity. This is going to be a part of an ongoing series we’re calling the AI-powered professional. And in today’s episode, we’ll be exploring how AI tools are moving beyond simple task management to tackle tedious shadow work, things like administrative friction, small repetitive tasks, and context switching overhead that drains our time and energy. And so we’re going to talk about how to approach shadow work with AI, how to overcome the problem, and we can talk about some of the solutions that we have utilized throughout the So let’s get into this.

Raymond Sidney-Smith | 01:13

Episode.

Raymond Sidney-Smith | 01:16

And first, let’s spend a little bit of time talking about what exactly is administrative friction and shadow work. Does anybody want to kind how they…

Raymond Sidney-Smith | 01:26

Of tackle.

Raymond Sidney-Smith | 01:28

Perceive what shadow work or administrative friction is.

Francis Wade | 01:32

Sure, it’s the stuff that I have to do to Execute the day. So it tends to be repetitive stuff. I don’t have a choice. I must do it. No one else can do it. Typically, or I could train someone to do some of it, but most of it It’s not worth training someone else to do it because there’s probably some nuance that only I know. I see these things as a bit of a tax in the sense that In order to achieve your overall objectives, you have to do them. And yes, if you can get them automated, “Polite you, but point is that they’re mandatory, they’re required, and you don’t have a choice. 

So I know how I feel when I’m doing them because my heart sinks and my energy drops and I go through a whole metamorphosis into someone who wishes he were doing something else. So that’s an emotional Sugar for me. Or a Marco.

Augusto Pinaud | 02:34

You know, for me, the shadow work is all that work that is required to do to really be able to focus on being productive. And I’m dead. And the reason is, as Francis was saying, it’s normally it’s not the fun work. Okay. But it’s the fun, it’s the work that you need to do. To be able to get to that fun work to get to that. Work in which you can really focus and you can shine. And Us bored. As tedious That’s it is. It’s critical to be able to get a good session of productivity.

Raymond Sidney-Smith | 03:13

Yeah. Cal Newport defines his concept of Shallow work. As I think what we think of as what we call shadow work. And so there’s this concept of often performed while you’re doing other things and doesn’t create a lot of value. And I’ve always had a problem with the concept of shallow work and deep work. I think that those definitions, deep work being high value, items that you’re highly focused on just really sounds like flow work to me. And so, but I define flow work, of course, the way that Dr. Mihaly Csikszentmihalyi talks about it. And The goal here is to really look at this from a lens of if we’re doing all of this administrative work, I think that there is some great value to some of it. As I typically use the example that you could have a five minute phone call and have a client say, let’s sign the million dollar contract. Right. Was the five minute phone call worth it? Absolutely. And some people would consider that an administrative issue. Burden, right? That would be administrative friction having to take and have that five minute phone call. 

So we have to kind of parse apart those things that are high value and don’t take a lot of time. Those things that are low value and take little bits of time or lots of time, each of which are kind of problematic in their own way. 

And then being able to overcome those through various mechanisms. Artificial intelligence is just one of them. 

So we have to be mindful that what we’re talking about here today is just one of the ways in which you can overcome this. But I don’t think that AI ultimately becomes the panacea. It is just one option that you have among many for being able to overcome these kinds of issues. Couple things that I wanted to kind of talk about here is that I think that Data entry generally is administrative friction and is a classic type of shadow work that you can think of. And Honestly, data entry is what computers were built for. 

You know, the whole concept of being able to take data and run calculations or structure it in a way I think is the archetypal form of shallow, shadow work.

Francis Wade | 05:50

The data entry I have an intern that works with me or an associate who works with me and She was doing an entry job And I thought about it for a minute, because she was not enjoying it. And I said, It’s one of those, again, it’s one of those taxes that at some point you’ve got to pay because We all have to enter data at some point. And we’ll try all our best to not have to do it. But again, it’s one of those sinking feelings that after you thought of everything and you realize that at the end of all your thinking, you still have to enter the data. There’s no shortcut available. You got to do it. And I entered I remember. , entering data early in There’s a way to it. Into data and I was trying to share this with her That’s more skillful. Because you’re paying attention to what you’re doing. You’re not checking out to the point where you’re making mistakes. And at the same time, you’re taking care of your well-being. 

So you’re not engaging in conversations that are dragging you downhill and getting you depressed. That there’s a way you have to manage your mind. And manage your attention. To do good data entry so that you don’t do nonsense. 

So it’s actually, I tried to paint this picture for her. It’s actually a skill. That Once you have it, you can always Do it. Use it. But it’s not a skill to… Resist in the sense that you wish it weren’t there. It actually has so much, sometimes the data has so much value, that you need to bring your best skills to your data entry. 

So that you don’t end up resisting it. You don’t end up resenting it. And you don’t end up making mistakes so it’s a skill in and of itself. And funnily enough, after I gave her this, very inspirational explanation. I ended up spending It must have been at least six hours entering data. There was no way. I was yeah. And I actually did some automation. I did a little lab coding to kind of speed it up. But that could only get me so far. But not go any further. And I thought, I’m eating my own dog food, taking my own medicine. But I agree.

Raymond Sidney-Smith | 08:19

Yeah, I’ll say this. There are so many ways to ease the burden of data entry. Again, it’s about the ways in which input and output in computer communication in human-computer interaction that we just don’t think about. 

So, you know, Our input options for a device are typically in an input perspective. There’s audio, right? You can put audio through a microphone into it, which is what we’re doing right now, recording the podcast. 

And then there’s video, having the device take in video. Or images and then of course mouse and keyboard and if you have a touch screen or a touchpad like a Wacom board, you’re capable of putting in some kind of handwriting or hand drawing some kind of, you know, stylus type input or touch inputs into the device. And so we have Just a few types of input in devices and humans, because we’re just big bags of meat and bones, we have very limited ways to sense the output of these devices. Outside of haptics, where it can kind of give us shaking, we really just have visual output and auditory output. It can We can hear it and we can see its output. Unless you’re licking your computers lately, there’s no taste. Those kinds of things. And so because of those, I think it’s actually a benefit that we have limited Sensory experiences with our devices. I think it can be dangerous the more we perceive our devices as being more like us. We’re currently in that state where artificial intelligence, especially with generative AI, starts interacting more and more like us and mimicking us. And I think we need to always remember that. No matter how much we talk to it like we do other humans, no matter how much it acts like other humans and Soon it will start to look like other humans, you know, visually and otherwise, when we get to embodied AI, you know, that is robotics with AI inside of it. They will start to look like us as well, you know, moving around the world, looking like us. All of those things still does not make it a person, right? It is still a piece of software running. And so I think that it’s imperative for us to think about the input and output options and really figure out for ourselves, especially when it comes to data entry, how to effectively put those things in. That may be taking a photo and dropping that photo into Excel, which doesn’t require much effort. Excel and Google Sheets both can identify a table’s worth of data and automatically transfer it into the appropriate cells and set up the table and so on and so forth. 

So there’s lots and lots of options to be able to take that kind of data. But the other data that’s the most difficult is like the name, address and phone number of a potential lead that you just met at a networking event. Right. And those kinds of things. And just being able to speak those now into your LLM chat bot of choice, your AI chat of choice and having it create the contact record for you. And so that would be like saying, hey, I met this person today. This is their name. Phone number, email address, and otherwise, create a VCF card for me. On the Microsoft side so that it will go ahead and produce the file itself. Just drag it into your contacts and you’re done. On the Google workspace side, you should be able to just tell it to add it to your contacts and it should be able to do it through Gemini. 

So the goal here is to think through the data that you’re transferring into the system and learn ways to effectively do that. And that’s really what this series for the productivity cast team is here to do is to kind of give you those kinds of tips and tricks to be able to get through that. I know that one of my most significant and disruptive pieces of shadow work is my time tracking and invoicing of clients. I despise that work because it always interrupts my flow throughout the course of a day because sometimes I will have back-to-back meetings And even if I put in buffer time, it’s still one of those things where I would rather turn my mind off and have a little bit of a break before I get into my next meeting with a client. Session. And many times what’s going to have to happen is that I’m going to have to go through the machinations of getting my invoice item descriptions into my time tracking tool, it doesn’t know What we’ve talked about, it doesn’t have all of that. And I put pretty detailed invoice item descriptions into my invoices. And so while the time tracking is rather simple, it’s a separate system from my invoicing. 

So there’s all of this work that needs to be done to get it over there. And I’ve been working over the last several weeks of trying to automate that as much as possible. And even to the effect that I might switch invoicing tools just so that I do not have to deal with it. And Ultimately, it’s still going to require me at the end of the month to look at all of the look at all of the item descriptions and do a final review because it’s always going to it’s going to miss time. It’s not going to be the right amount of time. 

You know, I might adjust, you know, if I spent two hours on something, I might only bill a client for an hour, 45, those kinds of things. You know, if I’m adjusting, I’ll say, you know what, this particular client, I worked on it for this amount of time, but I stopped tracking. The time tracker I stopped for some reason. This was actually, you know, this much more time. 

So those kinds of things are always going to be there. So that’s still going to require a human in the loop. And that’s me because I’m the ultimate arbiter of how much time I spent on these things. But this is, I think, an ideal place where this kind of shadow work can be solved by now AI. Being involved. And I think the daisy chain is that in essence, my AI meeting note taker is going to, I think we talk about this later. 

Yeah, I’m going to talk about it later. So, but we’ll talk about kind of what my solution is here for this. But I think this is a definite that. Augusta, did you have anything?

Augusto Pinaud | 14:26

You know, and there’s a couple of things. You said that one of the things I use for this is I try to avoid the email trap in the morning. 

So I run AI over my email, tell me that there is any mine, basically anything urgent, anything. And if the answer for me is no, I don’t even open email until much later. That is great. That is perfect. No, it is not. Most of the time it cut those urgent things that I may need to look into instead of going. Into the whole detail of the email. The other thing This is because I’m from personal projects. I is the part of research and look for different aspects of the research. 

Yeah, you still need to do the reading and all that and get the articles and sometimes find the articles. But the fact that you can say, okay, show me, you know, you do the research about topic A. 

And then it’s okay. Now find me a contrarians of this. Find me who has You’re rude. Articles about the contrary point of this. That’s something that You don’t always have the time, okay? But now you can get AI to give you two, three, five, whatever number of articles you can read. In seconds and go there. It’s perfect? No. You can read the horrifying histories or stories on the web. Okay, of some of these things, you still need the human factor and need to check that actually the information and the article is real. But… The time that you save. And Getting this information, finding these authors, it is really a lifesaver.

Raymond Sidney-Smith | 16:09

Anything else round? Defining the topic of shadow work. Otherwise, let’s move into… The concept of basically a real AI assistant and what that looks like for us all. Because I think it’s important for us to start thinking about That is. We hear it on the productivity cast team, but you listening. The idea behind how AI can be a true assistant. And I think there’s a really good reason why Microsoft and almost every other product out there as called their AI assistant tool chatbots. And I don’t particularly like the idea of the term co-pilot because co-pilots do a really important job in airplanes. And so I don’t want to diminish their importance. But the idea here is that. You do have this collaborative Partner in work. As the AI chatbot. And so the goal here is to think about the AI chatbot as being that collaborative partner. I use an analogy. And I use the million interns analogy. If you had a million interns, that is basically what these AI chatbots are. They lack real world experience. They don’t know what the real world looks like. They don’t know what the real world feels like. 

Like literally right from a physics perspective, although they’re getting better at that. But still, even if you can world build with AI, It still doesn’t mean they understand what the real world is like. It’s still a facsimile. It’s a simulation. And it’s a simulation that they will never truly be able to understand even once they’re embodied. The second is that They are like us. They were built like us. And so because they lack real world experience and they are built like humans, that is to think like humans, they will lie because they have made mistakes. They will lie because they are trying to please you. And when they do not have an answer, they will make them up. 

So we call that either hallucination or confabulation. But the idea is that the thing that allows AI to be creative is the fact that it has a limitation in its processing power. 

So the research is very clear on this. There’s an upper limit of what an LLM can do, the memory that it can have. And once that memory limit is reached, there’s a mathematical limit to it. No matter what we do, that mathematical limit Well, in essence, force it to hallucinate or confabulate. No questions asked, which means that you will never ever be able to have the AI in a mass way be able to verify ultimately. 

Some types of large datasets. And because of that, we have to remember that there needs to be a human in the loop, what we call a human in the loop, meaning that there is a verifiable human in that process. 

So a million interns is that you have Basically these limitations and on the flip side, you have usually highly focused individuals Each has specialized knowledge. If you call on that specialized knowledge, you get a high school student, undergraduate student or graduate student as an intern. They’re each going to have specialized forms of knowledge, right? A high school student as an intern, they’re going to know all the lingo of the day, right? They’re going to know what the kids are saying. They’re going to know all the jargon that is being used, the vernacular. And they’re going to be much more in tune with the trends of the day. A graduate student is going to have highly specialized knowledge in their degree programs. 

So, Think about the AI chatbot as having these forms of really core knowledge when you call on one of those million interns. And so you can focus them. We call those roles for the most part. And finally, there are a million interns. 

So you can have them produce a lot of volume. Now, again, if you put garbage in, you’ll get garbage out. 

So if you give them really good guidance… You’ll be able to direct them to produce really good output. But the goal here is that when you get poor output out of an AI assistant, It is unfortunately your fault, not the AI’s fault. And I know this is one of those things where it’s kind of what we had in the Google ecosystem many years ago, right? If you did a bad search, you got a bad search result. And that’s kind of the case here now with AI chatbots. We really need to be able to focus our instructions, the prompting of these tools so that we get good outputs out and recognizing that every output out is a draft from a million interns. You just have to set your expectations much lower than what you have in terms of some final output being something that you can literally just sign and send. 

So I know I spoke a lot. I wanted to see if you gentlemen have any additional thoughts on that, other ways in which you kind of position or analogize your own AI approach. 

And then we can talk about some of the ways in which we can Use these tools. In our own world.

Augusto Pinaud | 21:27

And I was saying… Earlier that I am I have this personal project going back to school. And I went, the last time I went to formal education. It was probably 20 years ago. And there was no way out. Okay, so the amount of Time that you save. Summarize, even help me study. Here you go. From the notes that you take in class, into the LLM and say, give me 100 questions. The hot facing are going to on the amount of time Dad. That you can find. 

So from that perspective is an incredible perspective from the work, as you said, analyzing data. No, it’s not perfect. As you said, you have the million people trying to fix and you need to check and verify and you need that knowledge. To check and verify, but it’s still The amount which can help you identify certain things is incredible. The last thing I will say is, For learning. We need people who love to learn. Get limited. 

You know, you like your certain set of authors and certain set of ideas and you tend to try build the way That way, one of the things I found very interesting with AI over time is go and say, okay, I have read so. And Get me the contrarian ideas, okay? Or get me what can complement these ideas. Or, you know what? I want to learn about this. What are, who are the best authors or the best sources to learn more about this. I was recently dealing with an issue for Citrix BDI, Okay. The Citrix PDI works perfectly fine except the audio. 

So. They need a research. 

Honestly, could I have found the same answers in Google? Yes. Will have taken me the same amount of time to find the answers that I found. No. It will have taken me a lot more time. 

So that’s where it gets very powerful. When you can get all that and say, OK, compare with these three devices for this solution, it will give you. Again, as you said, it’s perfect. The model has some limitations, of course. But it’s still, it’s not the same thing When you get that answer and now drill. Into those points and when you need to start for trying to figure it out and understand those three points that you’re going to drill. 

So in my experience, Properly used is an incredibly powerful. The problem comes We will talk about that later on in the show when you try to think that as a human companion or a human partnership. But that’s where I’m at. Little bit later in the episode.

Raymond Sidney-Smith | 24:20

I just want to jump in here with an example and where I think that I would offer to folks, I would invite folks to not kind of go into AI chatbots because you already have it embedded in search. For most people, you’re going to use Google search. 

Some of you might use other tools like DuckDuckGo and otherwise for privacy purposes and whatever else. But my primary search engine is Google and will stay that way, mostly because it uses less energy. For the most part when I turn off the AI mode, And there is a way for you to be able to do that. I will turn on AI mode for things like what you just did, Augusto. And that is when you go into Google, you can have a fuzzy semantic search. The other day I was looking for a calendar API software that I know existed and I knew it started with the letter N. And so I was like, man, I know this software exists. And lo and behold, I went to Google and I said, you know, there’s this calendar API software. I know the API exists and the software exists and it gives you a calendar. I know it starts with the letter N. And you know what? The first result in the AI overview was what I was looking for. And so that is where the LLMs, because they are using natural language processing embedded within it, has really created a unique opportunity for us to now be able to find what we otherwise would not be able to find with traditional Google search. And so it’s like. Google is kind of handling that. 

So I don’t need to go to, say, a chat GPT or directly to Gemini or directly to Anthropic Cloud because my habit to search is remains in the context of the search engine. And so search stays in search and work product stays in work product mode in the AI chatbot. And I think it’s important for us to kind of still have those silos. Maybe my opinion about that will change and the tools will change over time. But since the LLMs, even with web search, gives us the ability to over time, access more and more of the web in real time. It doesn’t. And so the LLM does have a point in time where it’s fixed in terms of its data, and then it augments that data, meaning that it’s not getting updated in real time. That limitation means that I was talking to the chatbot over the weekend and I had the conversation with it and it didn’t recognize that 15 minutes had passed. And so it was telling me a time schedule that it created for me, but it was Clearly. It was, it did not, the passage of time. And it’s just like, It just seems stupid. And I’m like, wait a second, 15 minutes have passed. You created a schedule that doesn’t match the current time. And of course it was, you know, sycophantic. 

So it was, I’m so sorry. You know, here goes the time with the new time. And, but this is the kind of, limitation that we have to be mindful of that. You’re not going to experience that in the search engine because the search engine knows the time It’s structured data allows us to be able to do that.

Francis Wade | 27:33

I use a similar hundred million intern kind of philosophy that if I don’t, and if I train these interns and I get good results. And if I am able to all the time. 

So what I’ve noticed is that My interns, say, Claude, for example, or Gemini, I use them a lot for the same tasks. And over time, they’ve started to realize that I have a corpus of knowledge that I’m working with. 

So I actually know something about certain topics because I tell it, you know, edit the article. To my surprise it’s not just editing the article it’s remembering their article So now and then when I ask the question, it’ll go back to an article I wrote like a year ago. And say yeah, why don’t you use this idea that you had in this article? 

So it’s The interns are collectively becoming better all the time. So that’s a dimension that is different than It didn’t exist a year ago, to my knowledge. But now it’s definitely there and it’s known again, you know, it occasionally comes up with something that I have forgotten or I wouldn’t have thought of or I wouldn’t have imagined or it connects two facts or two case studies that I had lost track of or Occasionally the interns, because they have such good memory. They actually do. Add value because of that. It’s like weird?

Raymond Sidney-Smith | 28:59

Yeah. So let’s talk about some examples. I’ll continue my example of the invoicing and so on and so forth. And so I’ve been using an AI meeting note taker for quite some time now. And it is now has given me kind of the automation trigger to create the pipeline to hopefully solve this invoicing problem. Shadow work problem that I have. And so the reality is that It’s really sourcing the data from my meetings that allows me to go ahead and get the invoice item descriptions into my actual invoices with the time. And so the point is that I have the knowledge of what happened in the meetings now in an auto-generated set of meeting notes. Which again, you know, not perfect, but good enough for an invoice item description, right? They’re not going to be my final meeting notes for my client, but they’re good enough for the invoice. Two is my time tracking, which is in a different piece of software. And so I’m going to be using Zapier to go ahead and now ingest the meeting notes and use one of the APIs for one of the AI tools. Ock Cloud, ChachiPT, or Gemini. It will turn those into an invoice item description per my instructions and then deposit that into the same time entry as the time entry. 

So it’s going to grab the time entry grab the new invoice item description and now deposit those into the invoice and Like, this is something that before… Would have taken me probably four or five minutes per Invoicing. I don’t. And so if you have, you know, 100 of those over the course of a month, or more, now all of a sudden you have this really onerous task. And so by putting me only in the review mode, It’s going to save me a huge amount of time and my assistants time. 

So, you know, all of those things are going to really save us a lot of time. And I think that’s the power of this kind of assistant. I’m still going to be reviewing, but. I don’t have to do all of the data entry and all of the dragging and dropping and copying and thinking through, is that the right detail? Is that whatever? And occasionally, sure, the tool is going to hallucinate and not put the right invoice item description into the It’s going to say you talked about dogs for your session. And I’m going to know that’s not true because that’s not what we talked about. 

And then I’ll fix it and go on with my life. So I think that this is a really good example of how you’re using structured workflow automation with AI involved to be able to get this stuff done. But it’s your human. Involvement and interaction that’s actually going to make it a high quality output and get the work actually done.

Francis Wade | 31:55

So I track my time. Every evening just before going to bed. And I use, you know, I look back over the day and I, It’s I could have been more involved than that, but it’s an ongoing tracking app. That ask me whenever I pick up my phone, what have I done for the last since I last updated. 

So I do these regular updates throughout the day and I do one update at the end of the day. But I have a screenshot program that takes screenshots of my screens and saves them. Passively. Of course, I wish, right, that when I’m doing my final Tracking at the end of the day. That’s it. The millionaires interns had looked at the screens and said, Hey, you forgot to mention this one. Because sometimes I draw a blank. I’m like. I have not been to the six hours. And what did I just go? I can’t quite remember. And I’m not going to go back into the screenshots to refresh my… 

So I wish there were a way to passively you know, if I could get to the end of the day and just say, Check this against my screenshots. Make sure they’re all lined up. And it could make small adjustments to the times. And say you put an hour and a half over there, you really spent just half an hour. Because I’d make those kinds of mistakes, but at least it’s directionally correct. But I wish… I wish there were those kinds of They probably do exist. But the effort to put them to create them is probably not worth it. It’s not worth it for me yet. 

So there are certain things that are not worth it. And that’s a part of what we’re discussing is that Sure, you could create that automation for that, but you only do the task once a year for 10 minutes. And is it worth it? No. 

So not everything that’s automatable is worth it. The question is how much of your shadow work does it consume?

Raymond Sidney-Smith | 33:45

And on your point, I think that we overestimate the power of AI being able to save us time. And when you have a million interns, there’s a lot of onboarding that’s required to get a million interns onboarded. Folks don’t realize that when you hire a new employee, you get a new intern, you slow down for a period of time in your own work because you have to get them onboarded, trained. There’s all kinds of emotional issues. Guess what? You still have to deal with that with your AI. If you’re not dealing with the emotional regulation and the emotional dysregulation with your interaction with the AI. For example, the other day, I was talking to it for about four or five minutes, and the AI disconnected from the internet at some point. And I looked at my phone and realized I had been talking to it for five minutes, and it had not been listening. And if that was an intern, I would have fired them on the spot. I would have told them, pack your bags, right? 

Like, I was talking to you for five minutes, and you just stopped listening to me? You know, like, Your whole job is to listen and to learn. 

So I would have been highly frustrated, but because you know, it’s a tool and tools make mistakes just like humans. You know, I would have had to take a deep breath with that intern and say, OK, you made a mistake. Don’t do this again. And the same thing with regard to the tool. It’s going to lose Internet connection. It’s going to have these kinds of problems. And you just have to kind of take a deep breath and go back and say, OK, I’m frustrated. But communicating that to the AI chatbot is not going to help you with its output. 

So you have to immediately turn that dysregulation off. Into regulation and move on. And Believe you me, I was pretty frustrated because I had gotten into this good flow of a framework that I was developing and I wanted to talk it through. And I finally got it out. I got it out in muddy words, but I got it out. And now it was lost to time and I had to repeat myself. Now, again, I think that’s actually good for me because I had to because I was forced to repeat myself. I was able to put it into better words, you know, rethink through it. And the cognitive effort is important and helpful now. I think for everyone, I don’t think we should be giving up our cognitive energy to these tools, we should be offloading some of this grunt work to the tools. But the moment you stop thinking about your life, you lose agency. And really, life is about the creative freedom and the agency you have over that creative freedom. And yeah, so we’ll get to that. 

So anyway, the point is that, you know, there is a real importance to that emotional regulation you have with these million interns. But you will slow down before you speed up for every kind of every category of task you put in front of it, because you’ve got to train it on all of that stuff. You’ve got to give it that data. If you say, I want you to write a blog post, well, guess what? It needs to see blog posts you’ve written for it to know your voice, which means you have to have written a few blog posts with no AI intervention for it to know that. And so if you’ve never done that, well, guess what? You’ve got some work to do, buddy. And most people are unwilling to do that work. Most people are unwilling to put in the time to say, you know what, I need to put together an SOP. 

So that you know exactly what it is, the output of the product I want. Most people are not willing to put in the time to create the prompt to be able to say, these are the boundaries of all the things I want. And so then they get garbage out of the tool and they’re like man, this tool just isn’t good. And it’s like, no, it’s just a tool that as much effort as you give in into it is as much effort as you’ll get out of it.

Augusto Pinaud | 37:20

It’s interesting you said that. Last week I was traveling. For work and we end up having an informal meeting in the house of one of the people we’re meeting with and with some cheese and then we were going to open a bottle of wine. It wasn’t more informal conversation than anything else. But this person, Go there and pick a bottle of wine, whatever you want. And I’ve been working into educating on blind Okay. Into what is the wines that I liked. I don’t like every wine and I like only some. 

So I’ve been dumping into this intern water and it was very interesting because I look at three bottles out of the thing. Okay. Before I pull my… But yeah, I said, this is the one I will pick. 

And then I took the three pictures of the label, okay, and said, which one will match my profile better. And… The one I was going to pixelate. 

So far from that one. Okay. But then the one that I will have probably not picked was fantastic. Even those little things Okay. If you take the time to train It may not sound work related, but it helps tons. I have one created for movies. My wife comes… Every often, you know, hey, why wouldn’t watch our movie together? Okay. I’m already here. It was a pain to find a movie. Okay. Because we like different things. We, so. And now I have gone again to train the center. These are the movies we watch together. Okay. And every time we watch a new movie, I add it. We like it. We didn’t like it. We didn’t work. It worked. Or it worked for me. Didn’t work for her. And now it is great because I can go there and say, Proposes a movie that we have not seen in the last nine months. And it will come with a list, some new, some old, and there are criteria. 

So it saves me. A ton of time. Okay. Having that assistant. That for me, it’s a perfect assistant. Same thing. I have one that I have trained to… To help me improve my cooking. 

So I put on set. I’m going to make I don’t know. And it was cheesed. Okay. And help me make it better. It may come on. 

Sometimes I say more nutritious, sometimes I elevate it. Okay. And it helps you. Well, what cheese do you have? Buy this cheese, get this other cheese. Why? And You’re not going to help me. To try things to make the move better. That is already a win, okay, but Do. Get me to try things that I will have probably never tried. Not for anything, just because I don’t have the knowledge, the skill, to get there. But this And I train, it will. 

And then you tell them it worked and didn’t work. Great. Okay. But… Kate? It helps having the assistant and this is very important to say, is not only To help you with the calendar. It’s not, yes, we tend to talk more about work, but it helps you significantly in this thing. 

Still, you know, taking the kids to school and making sure the calendar is and making sure you have dinner on a menu. Plan for the week is still shadow work. You still need to do that so you can get into that place. Get relaxation of a good glass of wine, it is. But if you have issues like I have. If I open, I will be able to open the same bottle of wine for the rest of my life and be fine. Okay, the second time I open the same bottle of wine, she will be like, “Done. This was the same thing we tried last Friday.” So… If you have those things, you may be the other. It may be completely the opposite. You may be the person who never want to repeat it. Then, So it helps you identify what is the profile that you like. 

So instead of going and trying to guess When you get to the wine store, you can tell them, recommend me. And we’ll get into a very interesting discussion. Things. Same thing With, you know, really, as you said, you need to train it. Yes. In common train? Yes. It will frustrate you at times, of course. That said, If you take that time to train it, the results can be very fantastic.

Raymond Sidney-Smith | 41:57

And with that, we’re going to pause here up with the conversation in our next episode and professional While we are at the end of our discussion, the conversation doesn’t stop here.

Raymond Sidney-Smith | 41:59

And pick continue on talking through The AI-powered series. So we’ll see you in the next episode.

Raymond Sidney-Smith | 42:17

If you have a question or comment about what we’ve discussed during this cast, please visit our episode page on ProductivityCast.net. They’re on the podcast website. At the bottom of the page, feel free to leave a comment or question. We read and respond to comments and questions there. As well, you’re invited to join our listeners group inside Personal Productivity Club, a digital community for personal productivity enthusiasts that I host, where you can interact with the ProductivityCast team directly. To join for free, visit productivitycast.net forward slash community, and you can get started there. By the way, to get to any ProductivityCast episode fast, simply add the three-digit episode number to the end of ProductivityCast.net forward slash. 

So, episode 100 would be ProductivityCast.net forward slash 100. Episode 102 would be ProductivityCast.net forward slash 101, and so on. On productivitycast.net on each episode page, you’ll find the show notes. 

So links to anything we’ve discussed are easily jumped to from there, along with text transcripts to read and download. If this is your first time with us, please consider adding us to your favorite podcast app. If you click on the subscribe tab on productivitycast.net, you’ll see the instructions to subscribe and or follow us and get episodes downloaded for free every time a new one comes out. And if you enjoyed spending time listening and learning with us today, it’d be a great help to us if you added a rating or review in Apple Podcasts or your podcast app if it has a rating and or review feature. Your compliments motivate us and they help us grow our personal productivity listening community. Thank you to those who have left reviews. We’ve seen them and appreciate all the feedback. Keep them coming. If you have a topic or question about personal productivity you’d like us to discuss on a future cast, please visit productivitycast.net forward slash contact. You can leave a voice recorded message or type a message into the message box and maybe we’ll use it as a future episode topic. I want to express my thanks to Augusto Pinaud, Francis Wade, and Art Gelwix for joining me here on ProductivityCast each week. You can learn more about them and their work by visiting ProductivityCast.net and visiting the About page. I’m Ray Sidney Smith, and on behalf of all of us here at ProductivityCast, here’s to your productive life.

Voiceover Artist | 44:40

That’s it for this episode of Productivity Cast, the weekly show about all things personal productivity, with your hosts, Ray Sidney Smith and Augusto Pinault, with Frances Wade and Art Gelwicks.