Subscribe
Share
Share
Embed
You ever think about whether this is all just, a simulation? This week the boys give up RV life to get back to.. something closer to normal, and we’re talking AI! What does it mean? Is it going to steal my job? And what are the toolsets out there today that could actually make you more productive.
Welcome to AutomationTown! A podcast about regular people, building automations for the problems we all share.
First time here? Start with S02E01 "For When You Need A Chatbot"
What are the things you do each day that you'd rather forget?
Chad & Jason explore common pain points for knowledge workers, and track down the people & tools necessary to automate trivial to-do's.
Previously on Automation Town. So what are we gonna do when we get outta here? What do you mean they tried to bury us? If we just waltz out here unscathed that the people I work with do not take kindly to Loose Ends Untitled podcast. Walk me through the rationale again. Bad guys tried to bury us under the grumbling center.
Got it. We got out. Bad guys are probably mad hiding only. Seems like it'll work for so long. Nothing against this RV Davis, but I'd rather not be on the run forever. So we get everything we know in a very public way with the rationale that they can't come after us if we're out there in the public. You know, Amelia, you mean badass?
Kicking down doors. Suspiciously knowledgeable about cellular equipment for a sound engineer, Amelia, that's the one. She's kind of amazing. Oh, Paul, welcome into automation. Uh, automation pod. Automation pod, automation pod. We have got a bit of a story to tell. We do, and we wanna explain why we're concerned that the mayor of our town isn't quite what she.
The podcast comes as it was, fear of the Automation Show team was lost at the Grimley Center collapse. The podcast details, the story of how the team escaped the disaster through a series of tunnels while being pursued by mass individuals. The automation show refer to them as the spooky police they got.
Going on to detail the events of the Automation Town Water Disaster implicating Mayor, good Ways associated with the Mass Individuals. When asked about the events presented in the podcast, the Mayor's office responded with the following statement. It comes as a great relief to myself, as well as the citizens of Automation Town to learn.
Chad Davis, Jason Stats, and Amelia Ledger are in good health after the Grimley Center collapse. Beyond that, I have no further comment at this time. Well, it should be good for downloads. I guess. We've had over 8,000 in the past 24 hours. The Manual Gazette had an article today framing the collapse as a cautionary tale about automation, something about an automated gas valve, and how automation and AI are going to steal our jobs.
Do you guys ever think about that? Automation stealing jobs. It's complex. I think that was a concern with every, like tech development from computers to smartphones to industrial revolutions. Yeah. I'm more interested in how we take care of the people who could be displaced by it since that progress is kind of just inevitable.
MCC Clark's work for everyone. Yeah. I love MCLs. I don't have my wallet. Can you? Every time
watch this. Welcome to mc. Can I you in a Clucky cream for only 99 cents. Hi, would lie, who doubles with, okay. Jeez Chad, please no pickle. Two double cluckers with cheese, no pickle. A big cluck. And a fry fry, please. Will that be all? Yes. Yep. Please pull forward to the first window.
You ever think about whether this is all just a simulation? I feel like if this were a simulation, it would be better. This week the boys give up RV life to get back to something closer to normal. And we're talking ai, what does it mean? What's just marketing fluff? And what are the tool sets out there today that could actually make you more productive?
All that on this week's
Well, we kind of call her now since her shows a podcast. But we can get a lot of questions via email. We can train this AI thing to be sort of like a co-host, like read some questions for us. It does feel on brand. Did you sleep last night? I came out at like two When you were still up? Yeah. It's, the clocks is good.
Yeah, working on it. I ran outta a potatoes on though I kind of can't sleep without it, so I made an ai. That's pretty cool. Sorry to hear that Chad. So what's the plan here? We'll record the. , then swing by what's left of the Grimley Center to see if our cars are there, may as well. We're kind of going back to normal life now, aren't we?
And not a moment too soon. You like the rv? Admit it. There's a smell. There's a real smell. I did notice that.
Okay. You ready to do the thing, Paul? Yeah, I'm, I'm ready to go. Chad, you're gonna prompt the AI guy right? Pardon? We shall call it Pat. Pat, okay, sure. Episode two of Automation Pod in three, two. Welcome in to Automation Pod. I'm your host, Jason Stats here with Chad Davis and oh yeah, pat. Okay, let's circle back to that.
First off, we heard from so many of you well wishes, lots of questions means a lot to us. We really appreciate all the messages. We got a pile of questions about the events surrounding the Grimley collapse, and to be completely honest, we have a lot of questions. What exactly happened, who was behind it?
How it's associated with the mayor. Trust us when we say we wanna know every bit as much as you. And we're doing our best to get to the bottom of it. If you have any info, any tips, please send them our way. In the meantime, we thought, what better way to honor the legacy of automation show than to get right back into it?
Helping people automate everyday tasks, especially as automation seems to be getting a bad rap these days. Yeah, that's right. We got a mountain to mail after the first episode of Pull Out some listener questions to go through, but first, Chad, do you wanna introduce us to your new little friend, Boy do I.
So, given that we're a podcast now, we can't exactly have people call in live. We've started playing with creating our own AI for the show that initially will read out listener questions, but down the road, who knows, could get into other things. So, without further ado, meet Pat. Hi, pat, you. Pat, sorry. I don't understand.
Okay, let's get to some questions first. Caller. Caller, no email. Are they emails? Yeah, they're emails. Emailer. Letter up, pat, Chad, and Jason. Boy am I glad to learn. You are. All right. You too have done such a service to this town. That's kind Please. Repeat request ness, you broke it. Uh, driven that the ley collapse was blamed on an automated gas leak valve to you to have many stories of automations going spectacularly wrong.
Some days I worry that I should leave the stop proper develop. Boy, do I ever. It feels like every day that you're building something, something goes wrong. It's never a like point A to point B to point C, and things just always work out, but sometimes you just screw up so bad that you have to laugh at them.
I've got one story about some accounting stuff that involved a couple different accounting systems. And I needed to create entries back and forth between them. Oh, I remember this. Yeah, yeah, yeah. That's my all time worst automation story. And it's only cuz it ate up like 10,000 make automations. Oh my God.
And I thought it was going well. Oh, you thought it was going well and it wasn't processing them correctly, right. Yeah. So for those of you keeping track at home, It's about 10 bucks. Nothing crazy , but just having that run and, and looking back and thinking like, okay, how am I gonna fix this? Yeah. Big, big hit to the ego.
How about you, Jason? I, well, the best part of that story is, the way that I found out about it was you tweeted. Like what you did to fix it, which was like this novel thing. , you omitted the entire fact that you had screwed up all of this stuff and this was your way of like a novel way to fix all this stuff that you broke
But I loved when I found out the backstory to that , you've gotta have a couple stories like this hit us. Oh, I've had a lot of nightmares. Honestly, I think I. Fortunate. I haven't bungled up anything too big yet. I will say like my biggest fear when it comes to services like, you know, Airtable or something like that, is if I'm doing something mission critical, I'm, I'm literally like an accidental base deletion away from.
Poof. Like all that stuff. Yep. Oof. Nope, I, no, the examples are coming to me now. It's a safe space, Jason. It's a safe space. I think where I've gotten burned more than anything else is air tables, automations, uh, cuz a lot of the automations you can set to. a view. So they run on any records that are inside a view, oh, I know where this is going.
And then you change the filters on a view and like it kicks off a thousand things that you don't want to run, because now that view just completely changed, or you removed the filters on the view and now it's being processed on. Every single record. I think like back in the early Airtable automation days views were kind of the only way to be specific about what you wanted it to run on.
Now, I would say it is not a good practice to set up automations on views. You can lock views so that people don't change that stuff, but like you just forget what you set it up on and there's just a lot to lose. Getting that wrong. I've talked to people who were like managing client lists and sending auto notifications to clients that way.
I think you change a filter and it emails a thousand people, something that's not correct at all emails. It's so worth so bad. Yeah, no, it happens honestly, like people are shocked by the power of the automations and the leverage and all of that. Like, great, with great power comes great responsibility.
Like you can also screw things up at an astonishing rate that wasn't possible before. Look, there's a silver lining on this, which is there will always be people smarter than you. And I remember talking to one of those people before and they said, look, you can fix this by just being smarter, by being better.
Like, thank you. That really helps. Like what do you mean they're like, Just put filters so in case this actually happens again, create another step in the automation that says, you know, if a date is older than two weeks ago, don't process it. Yeah, yeah, yeah. Overfil it over. Parametize it. Think of the use cases that can be catastrophic and then create.
Environments that are conducive to that. I was like, oh, that's really smart. Thanks. He's like, yeah, that's like coding. Yeah. You newbie . Yeah. The guardrails I ended up building in for those status notifications, so the client would get an email. At specific statuses was basically just a checkbox gets checked in a field if they've gotten an email for that status so that they never then get another email for that status if something changes.
Uh, yeah, you kinda gotta like over-engineer it. That's a good question. Uh, what else you got first Pat? New Explain and then layman's terms exactly what AI is and what ML is. Conflicting information and sometimes wonder if it's just marketing. Hi. Okay, Chad, simple as that. Layman's terms, what is AI and ml?
And 20 seconds. Go. Think of machine learning as logic and limited in what it can do, input based, that kind of stuff. AI is like an intelligent system, so it tries to deal with really complex things based on networks of data that has, you know, no specific small, limited scope. So ML is build stuff, it follows a build, and you deal.
AI is, it's like a neural network. It's like your brain. It could think for itself. And that's pretty good. I kind of wasn't serious when I said 20 seconds, but watching you go for it like that was pretty solid. Well, there's probably like way smarter people are like, no, you got it wrong. Oh yeah, . They're like, no, I just wasn't gonna explain it cuz I don't know if I could, but I think your framework's good.
So machine learning is like throw a pile of data at it and it develops kind of its own model of like what this data's supposed to be, or like it learns based on the more. That you feed it like some sort of, I guess it has to know how to take some sort of like action or develop an expectation for the properties of that thing that you're feeding it or something like that.
Yeah, I saw an app that was advertising as like the next generation, AI in Financial X, Y, Z. Yeah. I was like, so tell me about the ai. And they're like, oh, well we've built it. Oh, so it's ml like, ah-huh. Yeah. But we use AI for marketing. . Yeah, . There's a lot of that going on. Yeah. So like the mainstream AI we see now as like, You know, GPT three and stuff like that, which is general ai, where it's trained on, what do they say, like two thirds of the internet or something like that.
You've just got this enormous training data set that it has so that you can ask it questions and it can tell you things based on that training set of data. I've gotta ask you, I mean, I didn't know very much. Complete Nobe never dealt with anything technical with AI a while ago, and I saw people using words like G P T three and Da Vinci and Open ai and Jarvis and all these things.
How could you explain AI and all these words in a way that you're seeing it for the first time? Like what does it really mean when people throw these words around? Uh, it's hard to say because AI is being used in a bunch of very specific applications, so, Open AI's. G P T three is a thing that we are seeing a lot of, and it's like a text generation implementation of ai.
So like text completion or text insert, like adding text to the middle of something or editing something. But it's able to do it in this amazingly natural way and even. Like from its own knowledge, like build on the ideas or introduce new ideas and stuff like that. So what does AI do? Like the value of AI is that it can kind of do anything and it's being plugged in in a whole bunch of different applications from, you know, generating a slide deck based on your bullet points to like just about anything and everything.
So it's hard to pin it down. One specific thing because the whole premise of it is kind of this general knowledge, whereas machine learning I think is more targeted. It's like we wanna nail down this one specific thing, and by feeding it a whole ton of instances of this thing, it will then know more about that thing sort of thing.
Does that make sense? Yeah. , next caller. Next caller. We're probably in over our heads on this one. Uh, who you got? Chad and Jason, a question about artificial intelligence stealing our, um, yeah, that's not good. Can we pause on the recording, Paul? Yeah. Gimme a minute with this. Wow. So I heard back from Amelia, wait, you bounced back from the other day.
She said she listened to the first episode of the pod and feels bad that I got caught up in the middle of all this. So is this turning into something more? Oh, that look on his face. I know that look. So how's it been going? Can I. Okay, I don't see much here. Paul. I listen to the first episode of the pod and feel badly that you got caught up in the middle of this.
So that's it. Then. I'm not sure what to say back. This just brings me back to high school when, when I was in high school everyone had like just started texting. Oh, it was an exciting time. Some things never change, young love. Are we ready to go here, Davis? I think I got it started. Standby. It's a local number.
Hello? Jason Cheese. Louise Gil. I know Gil. Hey, am I relieved to hear the two of you are? Alright, Gil, we're actually recording. Good question for you. Of course. About live flow. Big surprise. So the other day I saw Google release Simple ml A let all no Code machine learning extension for Google Sheets. Hang on, I'm gonna put you on speaker.
Gil Chad, you were talking about this the other day. Simple ML for Google Sheets. Hi Gil. Chad. Hi Gil. So simple. It says it'll make predictions in your data, find potential errors, find patterns, lot of things that could be helpful for my clients. Makes sense. So check this out. I'm pretty sure I can use this on my client's accounting data by syncing the data out to Google Sheets with live plug.
Oh, now we're talking then using simple ML on the accounting data. Gil, I love that. Honestly, this may be my favorite part of live Flow, that when you use it to sync data from your accounting file out to Google. You then have the entire Google Sheets ecosystem at your disposal because you definitely don't have access to machine learning in an accounting file.
That's right. So what I'm thinking is I use live flow to sync transaction data out to Google Sheets, month over month financial performance. Then you simple out to analyze the data. That's a great idea. Gil, it's great marketing for you as well. How many virtual CFO shops are actually using machine learning to help your clients?
Where would I be with Audi? Is that the uh, rhetorical automation part? Summer Gill's Best work right there. I like it. Okay. Are we ready to get back to recording? I've got some food in the fridge. It's getting old. Any updates from Love Bud? Over there? Still just crippling indecision. Got it. Okay. Let's do this.
Paul, live again in three. Give it a go. Pat, Chad, Jason, A question about artificial intelligence stealing our jobs. I know it won't be as simple as in the AI when my job disappears. But how does an I developer like myself explore how to make my own life easier with ai? Oh, pat, coming from an ai, we're in meta inception here with this , the age old question, will AI replace and eventually just destroy us all?
Oh, it's so true. This is about accessibilities. That's why I kind of asked earlier about like what does it all mean? Like what is G P T? What is Da Vinci, what is OpenAI? What is Google cooking up? What is like all these other things dealing with and like, how can we just know what's possible? And I feel like it really wasn't an unlock for me until I got like a tool in front of me started typing in it and was like, whoa, this is gonna change Everyth.
And I think the first thing was like an image generation that was using like the Dally two AI stuff and I was like, whoa, that's really bad . But it's pretty great at the same time. What about you? When, when was your first kind of unlock getting into this? I feel like it was the image generation stuff that like kind of captured everybody's attention and how it was kind of janky but also pass.
So like in certain contexts it was totally fine and so like it crossed that threshold where all of a sudden it mattered and it almost felt like you could miss the boat if you weren't paying attention. How'd you navigate that landscape of, okay, there's image AI here, and then in the news you see all these people saying you're gonna lose your job,
How do you connect the image to those statements? I dunno, there's a, there's definitely a conversation around like, uh, what this means for artists and I think neither you or I are qualified to have an informed opinion on that. I will say just on the general discussion of this new technology, replacing people and displacing them from their jobs, there's always an element of truth to it.
There's always an element of exaggeration. And that's just kind of always been the case from the first computer to spreadsheets, to calculators and all that stuff, right? Like people re-skilling, like moving on to different tasks. I see AI as a continuation of that. I think as a society you have a responsibility to look out for the people who are displaced and what happens with them, but at the same time, like, what are you gonna do?
Stop. Like, not like stop progressing. Like I don't know how much Star Trek you watched as a kid, but in the next generation. Man, everything's good. Like they've got it figured out on earth, like people aren't having to work that hard. Like I didn't watch, but now I feel bad for the first generation. Yeah, I mean there were like, if you're like bumping around on starships, like things were pretty spicy, but it looked like for the people on earth things were pretty chill.
Did we answer the question here? Even remotely? The question was about stealing our jobs. So like, I like what you said about, there's always been new technologies come and this is just the next one. Yeah. Like any new technology, you have to start somewhere. So whether it's putting a prompt of like your dogs on a beach with a hat and a mimosa and realizing that, okay, spoken like a man that's literally written that prompt before, but yes, continue , and then realizing that, okay, the skill that I'm horrible.
Is giving instructions. Yeah. What are my resources to giving instructions? Well, it's copying people, . It's, it's seeing what other people are doing and then just experimenting. Yeah. So as long as you're not hurting anyone with that experimentation, and you're learning and sharing, and you're getting resources from things like Twitter or from FAQ sections or from videos on.
You're ahead of 99.9% of the other population out there that's not trying these new things and like that's where we live today. I think an interesting thing about AI that has, like a society interfacing with computers haven't really encountered before is these AI are trained on more things than any human can possibly know.
And so these AI have like these abilities that we don't. actually, no. Or haven't discovered yet. And so researchers talk about this. They call it capability overhang. So these are like hidden skills or hidden dangers even of AI because we don't totally know what they're capable of. And that's where a lot of the fear comes from, is if we don't totally understand what they're capable of, then are there risks that we're not even considering?
But to focus on the upside of. I think it's interesting now how we're starting to see that working with AI is really a learned skill and not everybody's going to get the same thing out of ai. You know, the early examples of this that we saw with, like Jasper, for example, the G P T three Writing Assistant.
Mm-hmm. , I think the headline is, oh, you know, we just put copywriters outta work, or people writing blogs, that sort of thing. When in reality the best version of that, An assistant to those people as a tool for ideation. Mm-hmm. , I'm seeing AI right now as a tool to have in your tool belt, something to pay attention to because two people will wield them in very, very different ways and there's still skill involved in getting good information out of them.
So right now I'm not like, I'm not seeing the doom and gloom, but I am seeing it as like a very valuable tool to be leveraging in certain situations and kind of a skillset to be investing in. It comes back to the age old adage of adaptability. People have been adapting to new technologies and new environments for as long as there have been people and like we can't answer the question, which jobs will be safe or, It's like impossible.
We could guess, but we do know that the people that embrace the change tend to work out a little bit better than those who don't. So that's what I'm looking forward to, is just surrounding myself with those people and moving forward. Yeah. All right. Let's do a couple more. Pat, what do you got? Hey guys.
I've seen more and more ai, Jan January. Learned online. Are there black takeaways I should be using this technology in my. So Jason, have you ever made fun of your friends and created any images of them for make believe blog posts before? That's a very specific reference. , I have, I'll tell you what, in the last month or so, I've been like diving into the AI generated art thing.
I just think it's really fun. I would also say in the last few weeks it reached this inflection point where a lot more. Become possible than like the early days of Dolly where it was like, oh my gosh, that's a cat and an astronaut. He in space. Wow. Like that's like, like cute, impressive. But now it's to the point where it's like, wow, like that's like kind of passable and you can actually like put things in specific environments and like, It's just gotten a lot better.
Mm-hmm. , I'm a mid journey noob. Right? I just reached my uh, 25 prompt imagine, uh, limit. Oh yeah. the free limit. Yeah. And, and I'm just about to press pay on the, the $10 a month plan. I know you've been playing with it for a lot longer. And for those like that, want to try it, mid journey's great because you don't have to be an expert, cuz when you go into the discord to start using it, which is the, the platform in which they're using to generate all of these.
you can see the prompts of everyone else who's also either really pro or really nub and like that's the unlock is that you get immediate insight into like, what are the prompts that other people are using? How have you used those prompts and those types of discords to, to increase your learning? So if you wanna be totally blown away, and we could link this in the show notes, the mid journey, like gallery, like showcase.
Unbelievable will absolutely blow your mind. When it comes to thinking about like the level of detail that goes into these things, like they're just incredible. The hardest thing is to get exactly what you want, like get the picture in your head to be generated on the computer. The things that it can generate are incredible, but it's not always gonna line up necessarily with what you actually wanted.
But when you look at that gallery, you can see the full prompts that got them there. Now it's not quite as simple. Running a single prompt, and it just gives you that because for any single prompt, there's an infinite number of variations of that. And then with Mid Journey, you will select specific ones that it proposes and then build on that and make variations to kind of that core one.
So when you see the showcase, people didn't just drop in that prompt and they got this majestic thing. It was like, Took a lot of massaging to get it there and that's where like kind of the skillset comes in as each of these kind of initial suggestions that it makes are then a seed and you can make kind of deviations on this seed and like that's honestly what's kind of been fascinating.
To me and and I, my impression of ai, which is kind of the surface level impression you get from like news stories and stuff like that, if somebody just types in one prompt and they get this thing, and that's the news story, when the real value of the tools is like going deeper and like iterating on these prompts and like massaging the result that you got into something that's even better.
And it's what kind of first took me down the path of realizing like, wow, like people are actually developing a skillset to make this stuff really useful. For them. So to get back to the question here, are there practical ways I should be using this technology in my business? It definitely depends on the business.
I think I still wanna say it's not a replacement for like pulling an artist where you need an artist. Mm-hmm. . And much the same way that I don't think, no code, displaces code, I think it actually creates more dev than there's ever been. Mm-hmm. . So everybody's had ideas in their head for this app that they want, they would be.
And it never came to life because they're not gonna drop six figures on a developer to develop a mobile app, but with no code. It's accessible. You can do that sort of thing. And so I think the best version of AI art generation is bringing to life things that would not otherwise exist, not as a replacement for going out and commissioning this thing that you would want, because.
It fundamentally isn't gonna do things necessarily the way that you want it to be. I guess an artist isn't going to either, but I think the best version of this is creating more of that, not necessarily displacing where it's already being done by actual artists. Yeah, I agree. And BA and back to you thing, like people are going to always find the novel uses first.
Sorry. The blog posts, the website photos if they, if they want to try to be funny, sending it to your friends, uh, newsletter images for thumbnails, including some image generation based on Airtable data that you're trying to create. Like that's kind of where we're all kind of starting and fumbling around here.
I'm actually kind of excited. You talked about how artists. Are still required. Imagine this tool in the hands of artists, . Mm-hmm. and like what they could conjure up and make better. A lot better than what we could probably. So I'm excited for what this looks like and who knows where it goes. Yeah. So if you think about how most things come to life, whether it's a scene and a TV show, or a concept art for a comic book, it's a normal person like you and I sitting down and doing this awful sketch.
That's right. And that's a very low resolution way to bring an idea to. , and so I, I think there's a ton of applications there for putting better tools in the hands of people that can't actually do that themselves. The context that you started this question with me writing a blog post the other day. In the blog post, I was kind of referring to this fictional character, and I was like, wouldn't it be funny if I could make this totally absurd off the.
Version of this character. And normally I couldn't, that wouldn't be a thing that I could breathe to life, but with a few mid dirty prompts, I had this shirtless muscley guy with this handlebar mustache, but he was still wearing a tie and it was in an office environment and it was like, Perfect and the context of the story and as this example of a situation where I just would've had nothing, uh, had I not pulled this in and it like made the entire thing better.
Best thumbnail of the year. Okay, next caller. Pat, what have you got for us? Hey guys. Well said. You're creating a tenant to hell. I think that's the end of Pat. Do you spell smoke? I think the question was what AI apps do we recommend? Okay. So to kind of clear up a lot of the terminology we've been using, which we probably should have done in the beginning, tried.
I feel like we've completely bungled our explanation of all this. So common names that you're, you're seeing around open ai. It's an organization that is releasing a whole ton of different types. AI technology, gpt three being the most mainstream example. That's kind of the text completion or generating text that looks like a human wrote it sort of thing.
Within GP T three, there's different models like Da Vinci and several others that work a little differently. Mm-hmm. , Jasper, the writing assistant that we referred to, like that's built on top of a version of open AI's, GPT three. It's like their own tuned model of that. You've got Google. Back almost 10 years ago, eight or 10 years ago, bought.
Deep Mind. Right? Alphabet bought that Deep Mind. Yeah. So we don't know what Deep Mind is going to be like publicly, but internally for Google, it's been pretty good so far. So like we've got that to look out for as well. So in your day-to-day, Chad, like what are the couple. Can't miss AI tools that people should be using.
I think we're all, well, if, if you're playing around with AI these days, you're using chat G P T, that is the free tool where you're not restricted by words or what they call token limits for these AI calls, you're probably not playing around with the APIs like Jarvis is in the background. Right. You're just using the tool like millions of other people are right now.
Mm. The stats are that chat, g p t got to a million users in what, five days? Yeah. And all of the big boys and girls after like Netflix and Facebook were all like multi months or years to get to that exact same situation. So I find myself with chat, G P T, rewriting things and using things, but because. Has a connection in, and it always did through the HTTP module, and Zapier has a connection in now too.
I find myself creating prompts through make automations to come back with data manipulation or creating CSVs or things like that that I couldn't get typically through chat. T p T. Those are the two that I'm playing around with right now. A lot more. Blowing my mind, I'm now fixing Google app scripts and creating them from scratch when I had no idea how to do them before.
It's pretty fun. How about you? I know you're getting deep into some other types of stuff. What are you playing around with, I should say? If you're going back and listening to old episodes now, this stuff is all changing so vast, so our answer's three months from now will probably be completely different.
Mm-hmm. , I would say right now, the two biggies for me. Definitely chat. G P. Which I think is a big deal because you've got an interface where anybody can message it, train it on simple tasks and say, now do this task for me. Like that's, that's novel because it's putting AI. Trained AI into the hands of anyone.
Second, I'm gonna cheat and this is kind of gonna be a twofer. If you wanna have a play with something amazing. Check out Mid Journey. It's spelled just the way it sounds. It does image generation, but the really incredible thing in the last few weeks, we got a new version of Mid Journey that was a big step up.
But the amazing thing is that you can combine it with a different image generation service dolly, and that's the one that is open eyes and do what's called out. So, for example, I wanna generate a badass looking pirate, and so I go to Mid Journey. It gives me a sweet pirate and I love it. But the hard thing with AI is to like get it exactly in the setting that you want.
Or maybe I want the picture to be bigger or a different aspect ratio, like for a cover picture on my Twitter or something like that. You can take that picture from Mid Journey, go over to Dolly on OpenAI, and this is like, I'm doing all this with free level accounts right now. You can upload that image into.
And then out painting lets you extend that picture and generate more picture around your picture. . So you basically draw a box outside the boundaries of your original picture and then type in the prompt and say, show a mountain. Yeah. A, a set of mountains or a, a steampunk pirate ship or something like that.
And it will draw it around your existing picture and a seamless way. There are no seams, like it perfectly continues it. It's not always perfect. I had one yesterday where the edge of the picture ended at an arm, like it was, it was almost to the hand and what it connected on the other side was like halfway up to the elbow of some other person.
And so it was like this joint arm and never ending. Yeah. So it's not always perfect but, and the example of that kind of joke picture we were talking about with the guy in the office environment, mid Journey just gave me the guy cuz it doesn't really do like a person and a scene. And then I used out painting to create the rest of an office environment around him.
Honestly, it's amazing, like the fact that you can just like continue that canvas infinitely like it's so cool and makes that thing so usable. So, My two things, Chad, g p t, and then uh, mid journey for Image Generation. Oh, you see so much of this on TikTok, like if you really wanted to get a crash course in what people are doing with Chad G P T, mid Journey and all the stable diffusion stuff, and like writing, just type in AI as a hashtag inside of TikTok and you'll.
Hundreds of thousands of videos of just people being blown away of what they're doing. So whether it's writing or responding to emails or creating quick legal agreements or whatever, like it's there. So I'm gonna add a third option to mind, which was like spending a lot of time on TikTok, just seeing what other people are doing.
Yeah. Inspiration. Cool. Well I think that's all we got Time. Today. Thanks for tuning in. Uh, if you are more confused about AI and ML than you were at the beginning of this, we are too. A quick reminder as we still work to get to the bottom of the last week's events. If you have any tips or any info you can share, please send it our way.
Thanks for listening to Automation Pod. Have a question you'd like answered by automation. Look for my RV around town and use the mail slot on the door. Automation Pod is hosted by Jason Stats, Chad Davis and Pat, and edited by Paul Mara. Nice work guys. Okay, can we start heading to the gro to see if my car was squashed?
Oh, I hope my car's okay. It wasn't insured, Paul. Oh boy.
Boy, I did. We just completely buble that explanation of AI and ml. I feel like we did, but probably I'm a little embarrassed, but still it doesn't matter because it's so exciting to play around with it. Hopefully people will just go out and check it out. Well, and the real problem is like three weeks from now, all that information's gonna be out of date.
You said free chat G p T, and I'm like, is that even gonna be a thing that's free forever? Like are they just gonna like send it back to the API so there is no public chat? Yeah, I mean, wait till people find out how expensive it's gonna be. . Yeah. Do you see the new chat Sonic, where it's taking chat G P T stuff from the old and then combining it with the new Google results?
If it's too old and it can't deal with it. That's expensive. So I think we're all gonna get pretty sad when people have to start buying and they can't write their entire novels because they'll have to pay, you know, 6 cents every time they run it. I'll tell you what though, it's been great marketing for Chad, G P T.
Holy cow. Did you see that? They also got a little scared and they said they're not gonna release the research cuz it's too scary. Which research? Yeah. So all the research that they're doing, right. From the millions of prompts that people are putting in, it's putting out some crazy research in the back end.
And CEO just said yesterday that he's not planning on releasing that research cuz it's too transformative. What does that even mean? I don't know. I'm scared. You'll find out about all the fan fiction that people are riding with it. Mm-hmm. . I can't wait to find out. Well, I don't know. Like I know the whole.
Manuel tribe and all that, like it's, you know, AI's gonna put you outta work. I think it's very easy for it to become the boogeyman because it is an unknown. Yeah. But at least we have Pat . At least we have Pat. And honestly, that was a, a little bit of a rough debut for Pat. It was. It'll get better.
Hopefully it doesn't have to cost us too much money to make him better. I'm up for the challenge. Let's do this together. Oh, there's the grumbling center. Even if your cars weren't both crushed, which by the look of it, they most certainly were. I don't think you're getting through this chain-link fence guys.
It's the mayor. She's out as mayor. Is that, put that up on the tv. Why is Jake there? Great things are in store for the future. I will always look back on this period of my life with fondness for the new relationships we forged for the incredible people I was able to meet along the way. Know that I will cherish those relationships forever.
She's out. I think she resides. Yeah. You'll be in great hands with Mayor MCC Kringle Barry with that, I'll hand it over. Jake, this isn't happening. While I can't say this is what. Expected. I will say that it is both in honor and a relief that you've turned to the people of Manuels Burn for support in the coming months.
Automation town finds itself at an inflection point. Do you continue on the path of change, of pursuing the dangers of the unknown, or do you clinging to the safety of the way It's always. The people of automation town have been put through a lot in the last two months, but I'm here to assure you that what lies ahead for automation Town is stability, comfort, more of the same, and it's with that eye on the past and that steady hand that I look forward to serving the people of Automation town as your new mayor.
Easy Automation Town is written and produced by Chad Davis and Jason Stats edited by Paula O'Mara. Keep up with the characters of Automation Town on Twitter @AutomationTown.