Why Doesn't AI Know What Day It Is? (And How to Fix It)

Ever yelled at your GPS because it sent you down a road that no longer exists? That's basically what's happening when your AI assistant confidently gets the date wrong.

Joe Downs and Peter Swain tackle one of AI's most hilariously frustrating quirks: it doesn't know what day it is. 

But Peter makes a compelling case that it's actually a feature, not a bug…sort of. 

Along the way, they unpack AI workflows for 401k rebalancing, family chore wars, memory movies, and a career GPS for employees feeling stuck. 

If you've ever wondered why the smartest assistant in the world can't read a calendar and what that means for your AI side hustles this one's for you.

Listen For

3:25 Why does your AI assistant not know what day it is — and is it actually a feature?

6:49 What is a context window, and why does it explain almost everything weird AI does?

13:16 Can you really use ChatGPT to rebalance your 401(k) based on your actual worldview?

19:36 How can AI referee your kids' chore wars — and judge whether the kitchen is actually clean?

29:39 What's the AI career GPS that could help stuck employees map their path to the next level?

Links Mentioned

Gemini AI (by Google)

ElevenLabs

Notebook 11

Read.ai

Otter.ai

Fireflies.ai

Nano Banana 2

Claude AI (by Anthropic)

Wey App

Email the “Idiots” Joe and Peter

Joe Downs

Website | Email | LinkedIn | YouTube

Peter Swain

Website | Email | LinkedIn | X

Transcript

Hosts: Joe Downs & Peter Swain

Joe Downs  [00:00]

We live in a world where we can use AI. And what Danny's doing is he's syncing — he's got a Gem, his calendars sync, so advanced — and yet the solution is, oh, just do something so archaic every time while you're using this so advanced tool. Picture this: you've got the smartest assistant in the world. It's read every book ever written, knows every fact, never sleeps, never complains. Then you ask it one simple question — what day is today? You get a blank stare back. Imagine that. Well, that's AI folks, and it's not a glitch, it's just how the thing works. Today we're going to explain why, and while we're at it, show you some things AI can do that most people haven't thought to even try yet, which is honestly my favorite part of this show. And then we got a great question from Danny — a Dear Idiots question from Danny — and we're going to kick the whole thing off with this. Danny, it's a better question than I think you even realized when you asked it. So I'm Joe Downs, with me Peter Swain. We're just a couple of successful idiots who are using AI to improve our lives and power our businesses. And fair warning though, Peter's British, which means he'll explain why AI doesn't know what day it is and somehow make it sound like your fault.

Peter Swain  [01:24]

Well, of course it's your fault, but yeah. See?

Joe Downs  [01:28]

Alright, Claude.

Peter Swain  [01:29]

What if I tell you it's a feature, not a bug?

Joe Downs  [01:33]

Well, we're going to get to that in a second, and I do want you to tell me that, and that's a stupid feature, but alright, Peter, here's what we have for the folks today. So three things we're going to get. We got the question for the mailbag, which Danny, that was your question, which is the topic of today's show — love it. And now I'm even more intrigued because Peter's telling me it's a feature — stupid feature. Anyway, I guess I'm going to wait to be proven wrong on that, Peter. I want to go back into the home because there's some AI stuff that makes things just run smoother around the house. And I've been talking to people and they're like, what about this? So we got some magic tricks. Today's full of magic tricks and one of them, actually, I guarantee people are going to start doing tonight. And then Peter, I think we've been neglecting a fairly large segment of the population, and that's really anyone who works for someone else — the employee. So there's some pretty cool — again, I'm going to say magic tricks — and some really neat things that I want to bring up with you. And I want to show someone who's an employee how they can start getting ahead and start making moves.

[02:50]

Alright, first — and folks, thanks so much for the emails. Keep 'em coming. idiots@successfulidiots.com. This week's Dear Idiots from Danny — and a little more context for you, Peter — he's running a Gemini assistant, which is awesome. Fairly advanced, has his whole calendar loaded in including his hockey schedule, and his Gem keeps reminding him about games he doesn't have on the wrong days, like it's living in a parallel universe. So Peter, why does this happen and how does Danny fix it?

Peter Swain  [03:25]

And it is a feature. It is honestly a feature. So we've got this thing Joe called a context window. Now a context window is how much AI can understand before it starts pretending it's drunk. And humans also have context windows. We're talking to employees later on, and a lot of those employees unfortunately might be in those situations where they spend multiple hours in a meeting. Those people that love calling a four-hour meeting —

[03:54]

— without yes, those — I saw the face. Yes, I know — those not understanding that after 45 minutes, the human brain really can't comprehend much anymore. We too, humans, have context windows. We have a certain amount of information that we can process at any one time. Now AI has a much bigger context window than we do, but it is a very hard cutoff. So at some point — and if you've been using AI for a while, you've experienced this — they just start talking gibberish. In the early versions you'd be like, yes, I can help you with that email. And it would literally say: snow monkey, elephant, giraffe. I mean it would just go completely off the rails. It was quite fun, but it wasn't very useful. So if you've got this thing called a context window, Joe, the question is what do you put in it? Because everything you put in is going to fill the context window.

Joe Downs  [04:46]

I'm not with you.

Peter Swain  [04:47]

So even the date is going to fill the context window. So the decision was made, rightly or wrongly — which is why I'm saying it's a feature; it might not be a good feature, but it was a feature. The decision was made: it's not very hard for you today to say, today is February 19th. And if time isn't a relevant factor to you — for example, we spoke about doing financial analysis, your board does financial analysis in the past, what would you do in this situation, here's my statement of account — or I'm looking to recruit a new marketing manager, write me a job description. If you think about it, nine out of ten of your use cases are not time-bound.

Joe Downs  [05:27]

Okay, so the —

Peter Swain  [05:29]

Large language model.

Joe Downs  [05:30]

I don't want to make this a technical show, but the LLM — is it a feature? So you're saying it's not a feature, it's deliberately not included as a feature. Correct.

Peter Swain  [05:44]

I would agree with you that it was a bad design choice, but it was a choice.

Joe Downs  [05:50]

But it was in the interest of being able to compute more — is that what you're saying?

Peter Swain  [05:58]

Everything you put in is going to limit how long a conversation you can have with it.

Joe Downs  [06:04]

Okay, so it's operating in a space that doesn't need to know the time for 99% of what it's doing for us.

Peter Swain  [06:13]

Correct. So why would you put it in?

Joe Downs  [06:14]

So why put it in and slow it down? I think it's a speed thing, is what it sounds like, right?

Peter Swain  [06:19]

No, it's like — imagine water in a glass. A conversation can only have so much water in the glass, and when you fill the glass too much, the water's going to come over the top.

Joe Downs  [06:29]

Alright.

Peter Swain  [06:29]

So it's not speed, it's capacity.

Joe Downs  [06:31]

Okay. So it's a design —

Peter Swain  [06:34]

It's a crappy choice in my opinion because the people —

Joe Downs  [06:37]

So they made a choice. Fine.

Peter Swain  [06:38]

The people that have that conversation, they say, well, why didn't we put the weather in? Or why don't we put your top ten favorite books in, or all this stuff. It's like —

Joe Downs  [06:48]

I'm on board with all of that, actually.

Peter Swain  [06:49]

Time and date is time and date. Come on. That's not the same as do I like Jill Collins' eighties trash romance books. What is the time and the date? So I do believe they should have put it in, but the solution really is as simple as doing one of two things. Number one, either say the date at the beginning of the thing, or two, ask it to look up today's date before it continues. So if you have a Gem — in this case, a Gem is like a custom GPT — the first line should say: before you continue, look up today's date. Do not make any assumptions. You must get an up-to-date time and date reference from the internet. If you put that as the first line of instruction in the Gem, Danny's problem disappears.

Joe Downs  [07:36]

Okay. But gosh, we live in a world where we can use AI, and what Danny's doing is he's syncing — he's got a Gem, his calendar sync, so advanced — and yet the solution is, oh, just do something so archaic every time while you're using this so advanced tool.

Peter Swain  [08:02]

Yes, and I'm going to defend the industry whilst agreeing with you for a second. We're not even three years in.

Joe Downs  [08:11]

It's not fast enough.

Peter Swain  [08:12]

But we're not even three years in.

Joe Downs  [08:13]

I'm kidding.

Peter Swain  [08:14]

And training a new model of the large language model takes a year. So if a bug gets found today of this level, it's at least a year before you can actually fix it properly. You can patch stuff, but to actually fix the model is a whole other year. Now obviously they're doing that in parallel. They're not doing one and then another and then another. But the recycles to the innovation —

Joe Downs  [08:39]

Do you think it gets added in soon?

Peter Swain  [08:42]

Yes. I mean, to me it's a ridiculous oversight. As I said, I actually agree with you — I'm just explaining all of it.

Joe Downs  [08:47]

As we see it. You saying oversight? Oh, you mean oversight in the decision? Yeah. Okay.

Peter Swain  [08:52]

I would've put it in.

Joe Downs  [08:53]

Yeah, it's so —

Peter Swain  [08:54]

Fascinating to me. It's like, well, if you want to make this useful to people in their daily life —

Joe Downs  [09:00]

Right?

Peter Swain  [09:00]

Then the word daily is already temporal. It's already time-bound because we're saying daily life.

Joe Downs  [09:06]

And when you say daily, I'm assuming you mean every increment of that day. Because really we're talking time too, not just date.

Peter Swain  [09:13]

But a context window of like 200,000 characters is already getting filled with 20,000 characters of all the stuff it's remembered about you, which is why they're pretty brutal with how they prune what goes into the conversation.

Joe Downs  [09:29]

Alright.

Peter Swain  [09:29]

To me, as soon as you're saying today, tomorrow, and next week, this week — all of those are temporal, which means that injecting the time in the background kind of makes sense, but they decided not to. So that was my first reason. By the way, we're only three years in. My second reason is: if you've ever tried to have conversations with people with IQs of 170 with triple math degrees and quantum physics degrees and a cryptology background —

Joe Downs  [09:53]

No.

Peter Swain  [09:54]

They're not necessarily the most well-adjusted of humans. Have you? Yeah, unfortunately, the intelligence that it requires to have conceived AI — developers are some of the smartest people in the world, but it's a very abstract form of intelligence. So actually humanizing the intelligence is one of the hardest parts of the job, which is why you prefer Claude and almost — well, everybody prefers Claude to ChatGPT because Claude is actually more humanized.

Joe Downs  [10:27]

More human. That makes so much more sense — that second part. Now I get it. I can picture the extremely high-IQ introvert who may or may not even operate on a normal nine-to-five schedule, if you will. And most of us don't, but we get up in the morning and we are all up around the same time and we all work until around the same time, and they probably don't have a concept of normal time. That is interesting. That's actually the most logical reason — they just didn't think of it.

Peter Swain  [11:08]

And then by the time they thought about it, to them they'd have a shortcut of Command-Shift-T upper caret that inserts the time. They're like, well, you can just hit Command-Shift-T upper caret, right? And you'd be like —

Joe Downs  [11:20]

And they probably actually calculated that by the time this reaches a mass of people who are angry that we don't have date and time, we'll introduce the Control-K key.

Peter Swain  [11:29]

See, now you're thinking like somebody that's an AI engineer, which is why — scary — Claude has employed, I think, about four times more product managers than OpenAI and ChatGPT do. People that actually think and represent the customer and have conversations with the developers.

Joe Downs  [11:50]

The more I talk about Claude, the more people I'm talking to about it, the more they're shifting over to that. I think they're seeing it. A lot of them are entrepreneurs like us and it makes sense. Speaking of time, you did some foreshadowing into what I want to talk about next with analyzing the financial statements.

Peter Swain  [12:06]

It's almost like I knew what we were going to talk about. Maybe I was just proving that I did read the show notes you sent me this time.

Joe Downs  [12:15]

I love it. This wasn't an email, but I also talked to people in passing and two things came up — or one in particular. This is not the employee part, but one thing that affects employees: most of them have 401(k)s and they're supposed to — I don't even know if they give these people a video to watch anymore — but you're supposed to put on your hat and now you're a financial advisor. So when you're making your elections at the window that they give you to balance your 401(k) portfolio, or when you just start one — how would you like to invest? It's almost unfair that you give people these options with no guidance or very little guidance. So Peter, what can they do? What can somebody — let's say they already have a 401(k), they got a statement — now it's time to make my reallocation elections. Is this where we introduce a board of advisors? Do you even need one?

Peter Swain  [13:16]

Well, the first thing I would do is I would voice prompt and I would speak through my belief of the world. So I'd be saying something like, I think that we might go to war with Iran, I think that's probably going to happen, and I think this is going to happen, and this AI thing is going crazy — and I'd be talking through my worldview. That's what I would do.

Joe Downs  [13:40]

This is to balance your 401(k)?

Peter Swain  [13:42]

A hundred percent.

Joe Downs  [13:43]

Okay.

Peter Swain  [13:44]

Then I'd be turning around and saying, based on everything I've just told you, give me a future prediction of the stock market. Tell me where you think things are going to go. Are we thinking stocks up? What's going to happen? And then I'd be saying, great, I'm with Charles Schwab, for example. I assume we're not doing self-directed, but are we doing self-directed at this point? Or —

Joe Downs  [14:04]

It doesn't matter, because whether it's a 401(k) — you've got limited choices — but you're still making choices. You're at Charles Schwab for your example, you've got unlimited choices. You're still making choices.

Peter Swain  [14:14]

Okay, so I'm in Schwab — based on what I've just told you, how would you respond? What would you do, and what would your hedge be against what I just said, just in case I'm wrong? Now we've done this in real life and then we asked it to backtest its strategy over the last ten years and then course-correct it five times over. Its strategy, five times over, based on backtested data from the last ten years. And we got it to the stage where it could perform just about 1% to 2% worse than a Morgan Stanley million-dollar portfolio manager for high-net-worth individuals. But without the fees.

Joe Downs  [14:59]

Well, you just made up your 1% to 2%. That's pretty fascinating. So it was rebalancing. Alright, well you took this —

Peter Swain  [15:07]

A, that's —

Joe Downs  [15:08]

— a different direction than I was expecting.

Peter Swain  [15:10]

I want my 401(k) to reflect my worldview. So if I believed that Trump was the best thing in the world and all these policies are amazing, I'd want my 401(k) to match and mirror the kind of philosophies that are coming out with environmental deregulation. If I was on the other side of the yard and I thought Trump was the worst thing that ever happened and he's going to get thrown out of office and environmental regulation will come back and it'll be double — well then I wouldn't want to invest in heavy industry anymore. I'd want to invest in light industry. And what I was trying to get to in my response was: you don't need to understand how all of that stuff is going to impact the other stuff. You don't need to understand what stocks, what shares, what bonds, what's going to happen to the housing market if America goes to war — with Iran, wrong example — you need to understand that, but you just need to say that's what you believe. You need to tell it whether you want to be hyperaggressive, or do you want to be conservative. What does your portfolio look like?

Joe Downs  [16:10]

I'm not sure some people should do this because I think their beliefs might be misguided, but you just gave them a roadmap to — well, hold on. As a former financial advisor, I will say this: you just gave them a roadmap to have your money managed in a way that aligns with your beliefs. Go ahead.

Peter Swain  [16:30]

Wait, wait, wait. Something that's really important — we've got to make sure we do this. Obviously this is not financial advice. This is for educational and entertainment purposes. Anything that is said should be taken as purely entertainment value. Thank you very much.

Joe Downs  [16:42]

As a former financial advisor with securities licenses, I was going to make sure we got that in there, so thank you.

Peter Swain  [16:48]

Yeah, well, so that would be the approach I would take. What else could you do? You could also say: if you were a money manager for Charles Schwab, what would you do? Tell me what Warren Buffett would do when presented with the same information. Again, we've had this conversation so many different ways, so many different times, and we're going to carry on having it. Context, context, context, context. The prompt is simple — I want you to rebalance this. Simple. The context: what is it you believe? How much risk do you wish to take? How much risk do you not wish to take? How many children do you have? Is there anyone in your family who is less able that you may need to look after? Are you about to start preparing to care for an elderly parent? In which case you might need more cash allocation than long-term holdings. So those are the kinds of things that come into play here.

Joe Downs  [17:43]

Alright, well this 401(k) is a nice bridge back into the home because we probably have one if we have a 401(k), and therefore we probably have things that maybe don't go as smoothly at home as we'd like. So along the lines of using AI to help us with difficult family decisions or planning for holidays — we've sort of touched on working with contractors in the past, but you know what we didn't talk about was even renovation budgets. If you are looking at doing something — or, you know what, as I was doing some show prep, one of the things that stuck out to me in this part of the show, that I really thought was great — I grew up as one of twelve kids and we had a chore list, and for us it was either a nightly one or a weekly one.

[18:44]

There was every now and then the mob would rise up and demand change. And so we would switch from weekly chores to nightly chores. But managing it — balancing the managing of it — was part of the... can you imagine? You have two kids. Can you imagine? There weren't twelve angry men, but there were probably six or seven of us that were involved. Actually, it's probably only five and six. If I recall, it was matched to the days of the week because the others would've been younger. But my point is — and I'm getting too long-winded here — you could imagine the battles over who was doing what. How can AI step in here and just take some of the tension and stress away from these normal home conflicts?

Peter Swain  [19:36]

Oh, I've got a cool idea. What if you were to use Claude — and as the parent you would set up a Claude Team account — and you could use Google Gemini as well for this — and you said, here are all the chores and here are my kids. And when they log in, you are going to spin the wheel of all the chores. And when they take a chore, you are then going to make sure you give it to them in such a way that they can prove whether it's done or not via a camera. So they could take the picture of the clean kitchen — they could take the picture of it — and at the end of the day —

Joe Downs  [20:17]

Wait a second — would AI judge whether it was done?

Peter Swain  [20:21]

Yeah.

Joe Downs  [20:23]

Oh my goodness. Go ahead.

Peter Swain  [20:25]

And at the end of the day, you're going to tell me how many chores each kid has done, and whoever's done the most gets the next day off. How about that?

Joe Downs  [20:33]

That is hilarious. That would've been interesting in our house. I'll tell you what, things would've gotten done faster.

Peter Swain  [20:39]

No, that's just my crazy idea. But yeah —

Joe Downs  [20:43]

Capitalism would've definitely won the day there. Whoever could have gotten it done faster for the day off. What about along those same lines? I love that. For the chores — for holidays, holiday planning.

Peter Swain  [21:00]

Holidays? Easy. We did that. That's the Disney one.

Joe Downs  [21:03]

Well, no, no — I forget, you're British. We call those vacations. Holidays are things like Christmas and Thanksgiving.

Peter Swain  [21:13]

What is Presidents' Day? Whilst we're on this —

Joe Downs  [21:16]

We just had it.

Peter Swain  [21:16]

No, I know, but what is —

Joe Downs  [21:17]

What is it?

Peter Swain  [21:18]

Yeah.

Joe Downs  [21:19]

The day we honor the fact that we had presidents.

Peter Swain  [21:22]

Congratulations on having presidents. Yeah.

Joe Downs  [21:25]

We honor the Mount Rushmore presidents.

Peter Swain  [21:29]

Ah, see, that makes more sense, right? So it's the big four.

Joe Downs  [21:33]

Yeah, right. Actually, I'm not even sure it's the —

Peter Swain  [21:36]

Four. Most of those were British, so good. I agree.

Joe Downs  [21:39]

They were not loyalists. Clearly, they put their name to something. I don't think you want to go there.

Peter Swain  [21:48]

If anyone can't hear the inherent humor — I'm proud that we birthed one of the greatest nations in the world. You're like our child. It's a true honor.

Joe Downs  [21:57]

Look at me squirming in my chair.

Peter Swain  [21:59]

You just so want to respond, but you're like, I can't on the pod. Yeah.

Joe Downs  [22:02]

I'm just sorry, but we're just one — oh, we're just always going to be one. Oh.

Peter Swain  [22:06]

I don't disagree with you — courtesy the French. Okay, let's get back to it before I go on a French rant.

Joe Downs  [22:11]

Oh geez. Alright. Alright. This is one that I love and it's something I think I want to do. So one of the unfortunate things about AI lately — well, fortunate or unfortunate depending on how you look at it — is all the deepfakes. Some of it's crazy. Unfortunately it's getting some kids in trouble at school — I've read headlines about that. Some of the stuff is wildly entertaining though, that you see on social media. Some of it's wildly misleading, certainly when it gets to political, and all that. My point is there's a whole bunch of noise out there. Can we bring it back to some goodness, some wholesomeness that AI can bring into the home? How do we take — Peter — old videos, pictures we've got... I don't know how many pictures I have on my cell phone, a hundred thousand, who knows? Because our cell phone is our best camera. What can we do with this? What can we do that's positive? Maybe put something together that puts a smile on somebody's face or brings a tear to their eye.

Peter Swain  [23:21]

So Claude can create these things called artifacts. You don't need to know how to do them in order to do them. And you could literally just upload twenty pictures and say, create a digital experience of these. Next: a soundtrack. You could go to lovable.dev and you could say, here are all the pictures — create me a family photo album and make it so that there's a guestbook, so they sign and put their own memories. So you could have this interactive tool where everybody has their own photos and it kind of groups them together and stuff like —

Joe Downs  [23:50]

That. Where does that live? What's the delivery method of that?

Peter Swain  [23:54]

Lovable.dev has its own cloud, so it'll publish it to the cloud for you. So it would create its own website.

Joe Downs  [24:01]

You can make memory movies or something.

Peter Swain  [24:04]

You could make memory movies. You could make guess-the-location games. You could get it to generate a different version to do spot-the-difference versions of it. I mean, you could turn them into trading card versions and put —

Joe Downs  [24:18]

Soccer players on them.

Peter Swain  [24:20]

Yeah. The answer to almost all of these things, as long as we're in the digital realm, is: where does your imagination take you? It really is unlimited as to what you can do. Once you go, oh, it can basically digitally deliver anything that I'm thinking of — then you're up and you're off to the races. Now, just on this subject, we can decide to go down this rabbit hole as much as we want or not at all. The reason for the deepfakes, by the way, is because they're regulating for the wrong thing. Some parties and some senators and congressmen are pushing for regulation to say the production of deepfakes is illegal. Prohibition has never worked in the history of ever. It doesn't —

Joe Downs  [25:15]

Work.

Peter Swain  [25:16]

Never has, never will. And if you do that, all that's going to happen is these kinds of silos — the Chinese AIs just pop up in the background and people work around it anyway. Where I think they've gone critically wrong is: what should be illegal is the dissemination of deepfakes, the distribution. Because name and likeness laws already exist. I own my face, you own yours. David Beckham ironically doesn't own his because he sold it to a company that he owns. But we own our name and likeness. Those laws already exist, so we should be working more on the distribution, because then you could actually regulate Instagram, Facebook, Google, et cetera. It would be the distribution — not the production — that becomes the legal constraint. It would be much easier to enforce a law on Instagram, Facebook, Google, et cetera than it would to enforce a law on 350 million —

Joe Downs  [26:18]

Individual —

Peter Swain  [26:19]

People.

Joe Downs  [26:19]

That's an excellent point. That is an excellent point.

Peter Swain  [26:23]

Complete rabbit hole. But the deepfake thing is —

Joe Downs  [26:26]

Going to —

Peter Swain  [26:26]

Carry on. It's not going anywhere until you actually start regulating and financially hurting the people that are facilitating the distribution of the deepfakes. Because Facebook and Instagram are currently earning money — they want more content that more people are watching. So it's in their interest to let you publish a Margot Robbie playing baseball video because if 50,000 people watch it, they get paid.

Joe Downs  [26:54]

Yep, that makes sense. I want to get to the employee.

Peter Swain  [26:58]

Yeah.

Joe Downs  [27:00]

I think there are more employees than any other social group in the world. The other day I was having lunch and I overheard a conversation — it was two employees working for whatever company — and I picked up bits and pieces of the conversation, but the gist of it was: one of them was kind of counseling the other, and one of them was saying that they didn't really know what their boss expected of them. Not necessarily in a negative way, just from a — I don't know how to advance myself here. I don't know what the rules are. I don't know what they expect. I don't even know where I'm going. I don't know if I love this job. What's my career path? Am I blocked? And the other one was trying to counsel, well, you've got to work on your resume. I said, well, you can only put so much. The whole point is, I don't know about you, but when you overhear this stuff I wanted to just say, hey guys — and then give them nine different things that my feeble brain could come up with.

[28:07]

I thought it better... well, first of all, I didn't have time. I dunno about you, but I get burnout just trying to counsel people on AI. But I thought, you know what, this is going to make for a great segment. Because there are so many employees out there, and people listening — most likely, odds are — you work for somebody. I know we're entrepreneurs, we love catering to and talking to things that help entrepreneurs and students and parents, but we haven't done very much for the employee yet. So you're in a job, Peter — you're in a Fortune 500 company, let's say. That's a whole different... you talked about meetings earlier and four-hour scheduling, four-hour meetings. I listen to my wife sometimes — she's VP of Sales for a Fortune — she's a parameter. I swear they have meetings to plan meetings. I mean, these are all-day meetings, meeting after meeting after meeting. So it's a whole different organization. You have to learn. There's a different way to talk, and talk about how you downloaded this person and uploaded that. But it's a whole new language.

Peter Swain  [29:16]

Where's the question, Joe?

Joe Downs  [29:19]

You see how much this is painful to me.

Peter Swain  [29:21]

Exactly. You say you get burnout as entrepreneurs — listen, I'm tracking. What are we at?

Joe Downs  [29:27]

Help the employee who's stuck in this morass of a bureaucracy navigate their way through to — I don't even know, frankly. I think they should navigate their way out.

Peter Swain  [29:39]

The number one problem — you're right, this would be a good segment — but the number one problem is I got fired from the only job I ever held after two days. I called someone an expletive. So I'm obviously not actually very good at this myself. So how AI would work here, I don't really know. But the things that come off the top of my head: I would download the company's financial statements, I'd download their quarterly reports, I'd get the transcripts of their shareholder investor earnings calls, and I would have a growing Rosetta Stone of what is this company trying to achieve, what is its north star, how does it work, et cetera. So that I really understood what winning looked like at the top level. So that'd be the number one thing. I'd have the AI build a very rich context document on the company I work for. I would then do what we said a couple of episodes ago about when I built the virtual version of my wife — I'd build a virtual version of my boss so I could have conversations with them and I could role-play having a conversation, and I could work out how to present ideas, concepts, et cetera.

[30:54]

Those are the two biggest ones I'd do. I'd have a virtual version of the person I work for and a very deep context document for the company I work for.

Joe Downs  [31:04]

Okay, that is excellent advice. Obviously I'm thinking — because one of the things that this particular individual was struggling with was the career path. Is there a prompt —

Peter Swain  [31:28]

I'd get Manus to scrape Indeed and look at the job ads for my next job, and then I'd get it to disseminate the aggregate skillset that would be required for those jobs. So if you are a director of sales, what does a VP of Sales job description look like? What's the difference? What do I need to prove —

Joe Downs  [31:51]

And what qualifications do they need? Maybe depending on where they are, or —

Peter Swain  [31:55]

Yeah, what additional courses could I take?

Joe Downs  [31:58]

Yeah, that's where I was wondering.

Peter Swain  [31:59]

What networking groups should I join? I basically say to the AI: this is who I am, this is where I want to be, and I want to be there in this amount of time. What do I need to learn? Who do I need to know? What do I need to develop? How do I prove that —

Joe Downs  [32:13]

And can I add one? I think — how do you make that progress known? How do you promote that along the way?

Peter Swain  [32:22]

How do you advocate for yourself?

Joe Downs  [32:23]

How do you advocate? Thank you. So I was going to say brag, but you're right. How do you advocate along the way with your boss? Could it put together an entire GPS for your employee profile?

Peter Swain  [32:38]

Yeah. Because you end up saying something like, hey boss, can I get Friday off because I'm going to this executive leadership conference to learn about this new CPA? I don't know — I'm literally not making up words now because I'm not an employee. No, I know, but I'm going to this thing and they're like, did we pay for that? And you're like, no, I wanted to learn it. So I think you can start peppering in the accolades without wearing the badge.

Joe Downs  [33:06]

Even just little tips and tricks like that — when to drop them in and how to position them.

Peter Swain  [33:11]

And everyone's on social media, so you want to make sure you get your picture at the conference, right? With the logo behind you. But yeah, AI can guide you through. I mean, one of the things about AI — and Joe, I think you are trying to experience this more and more now — is you kind of have to just keep talking because the entire thing takes five seconds to show somebody. So you end up with lots of dead air to fill. Because the answer to can AI do it is yes, yes it can. Whatever the blank is — if we're in the digital world, the answer is AI can do it. If we're in the physical world, next year the answer is Optimus Prime can do it. So anyway, AI can do it.

Joe Downs  [33:57]

We're not going there yet. But you're so right though. It's like what we talked about — I dunno, it was last episode or the one before — about how the difficult conversations... the hardest part is what we think is the hardest part is actually the easiest part. The hardest part is the context. It's loading the AI with the context — then you get back in seconds very good guidance, advice, whatever it is that helps reframe you. Those were your words, and it prepares you to walk into any situation. And here it's just — to me, this is the long form of it as the employee. I like that career GPS, right? This is the long form of it. It's the context. Just constantly putting — it's a diary. You have to have a daily, multiple-times-a-day conversation with your diary. If you're being intentional about something, you're giving it all that context.

Peter Swain  [34:53]

And I tell you, for everybody else out there — as soon as you've done it once and got the whole prompts working, you're probably only $5,000 to $10,000 worth of development time away from having a launchable product. Because as soon as you said career GPS, I'm going, I could build that.

Joe Downs  [35:11]

Oh, is that a product?

Peter Swain  [35:13]

Yeah, but all of them are.

Joe Downs  [35:16]

This show is going to morph into a mini Shark Tank.

Peter Swain  [35:18]

But like: wrapping AI for renovation costs — product. Holiday planner — product. Memory maker — product. Each of these becomes its own product. And with the advances in some of the vibe coding and the more technical stuff, if you're even moderately inclined in that way — you've got a math background or an engineering background — you're not that far away from me. Or, to actually bring these things to life, which is insane.

Joe Downs  [35:46]

Alright, I'm going to brag. Before the show started, I got a text from my son. Do you remember a couple of weeks ago we talked about the athlete training program that you could put together? I developed one for him for lacrosse. I said, here you go — you need to start doing this. It's a big recruiting year for him. And I was a little disappointed because I don't think he'd started to implement it yet. And this morning I got a text with the link to a website he created in Manus, which is for high school athletes' training, and it's an interactive website. And his next text was: we just need to test it out and then I think I can sell it. I was like, yes, it's working. We are winning. I couldn't be more proud.

Peter Swain  [36:34]

Congratulations.

Joe Downs  [36:35]

And who knows if it's ready — obviously there could be some bugs in it — but the fact that he's even there taking an idea and turning it into a product, right? A service in this case, is just amazing. Well, the career GPS — I didn't see that coming here, and I absolutely love that. But I love this show because even from the beginning with why can't AI know what day it is? That was such great insight into how this thing is designed. And one of the frustrating things, I think, for people, Peter, is that it feels like it's constantly changing so fast it's hard to keep up with. And I know that's frustrating when you're trying to figure out, what do I do next? I just learned that — now it's like you're telling me to look over here at this shiny object.

Peter Swain  [37:28]

I just spent two weeks learning Opus 4.6 in Claude and the limitations of how to build Claude Code and how to use it. And yesterday, Sonnet 4.6 was released and it invalidates almost everything I've learned in two weeks.

Joe Downs  [37:43]

But maybe not the context of what you learned.

Peter Swain  [37:47]

The important thing is to learn the math, not the calculator. It's about learning these transferable skills that can traverse from family to employee to business, so that as things move along, it doesn't really matter whether AI adds the date or not. But I think hopefully what Danny took away from that is just adding that first line to the Gem would be something he hadn't known he needed to do — he or she.

Joe Downs  [38:15]

And my point is — for now, because it sounds like a fix is on the way —

Peter Swain  [38:20]

A hundred percent. That's also a permanent fix for any issue like that.

Joe Downs  [38:26]

Yes. The thought process, the —

Peter Swain  [38:28]

Right. Oh, if it doesn't know my location, then I could just add that into the top of the Gem before we do anything else. Look up my location.

Joe Downs  [38:39]

And I guess what I'm looking for is the skill of — oh, this is how I need to treat that. It just happened to be the date we're talking about, but that skill is transferable to anything else. Yeah, I have so much fun doing this. Hopefully folks are learning every week. From Peter here, I know I am — that's why I enjoy doing this show. So if you are, please like, subscribe, and share. Help us grow the show. We genuinely appreciate it, and keep those Dear Idiots emails coming: idiots@successfulidiots.com. Or one final question, Peter — idiots@successfulidiots.com or idiots@successfulidiots.ai. My question for you quickly before we go: is .ai going to overtake .com?

Peter Swain  [39:32]

No, because some people — when you say AI, because they're not in our world — will put ai.com, which doesn't work.

Joe Downs  [39:41]

I thought you were going to say they think it's Allen Iverson, the original AI.

Peter Swain  [39:44]

Let's leave it with .com. But we bought both, so it doesn't matter. Wherever you want to email us, email us. We'll find it.

Joe Downs  [39:49]

For Peter Swain, I'm Joe Downs. We are your Successful Idiots. Thanks for listening and we'll see you next week, folks.