AI Backlash Update
Hey, what’s going on? Welcome to The Doug Show. It’s Doug Cunnington here, and today we’re going to talk a little bit more about, uh, the AI backlash stuff. Apparently, um, there’s other people that agree out there. If you’ve been listening and watching the show for a little while, thank you by the way, but you’ll know that especially on YouTube, the algorithm is highly, highly, uh, super sensitive.
Uh, the show has not been getting many views recently, but I’m not doing it for, for views or pandering, um, to the audience or anything like that, although it’s a f- fine audience, but I’m not pandering to the algorithm. I guess that’s probably a more accurate way to say it. So a lot more views. A lot more views on YouTube and a lot more comments and stuff like that, which maybe is a better gauge, like, um, they go hand in hand.
So people are interested in it, and I have a couple, couple thoughts, uh, more news in the, in the AI front, although I, um, I was a little torn.
New Mic Gear Test
I’ll come to it at the end, but I was a little torn like, “Oh, should I lean into this a little bit more or not?” One, uh, side note, I am using a different mic this time.
So I was able to get a pretty cool tube mic, a condenser mic. It’s way more sensitive, and it, it’s supposed to sound a little bit better Probably 99% of the people that watch or listen, you won’t be able to tell, ’cause you’re on your phone in some capacity. You’re probably doing something else. Maybe you’re in your car, you’re playing it over the speakers, or you’re at the gym and you have headphones in, and there’s other noise and stuff going on anyway.
And, uh, this might sound slightly different. Hopefully there’s not more background noise. It’s a little bit of a test. I got this mic because I got a gig doing a, uh, the foreword for an audiobook as, like, the recording engineer, if I may, uh, use that term. Definitely outside of my area of expertise, but I did a little bit of, uh, studying, a little learning, and I was able to record this, um, audiobook foreword.
It was only, like, a couple minutes or so. But I got this mic. It’s a Warm Audio WA-251. Super cool. It has a 12AY7 tube and, like, a separate power supply. 100% overkill. Absolutely unnecessary, I think. Um, I, I like gear, though. So I got this thing, and I was curious to see, like, if I just, I play, strum a couple chords here, I’m not even moving anything around.
Like I said, it’s quite sensitive, so I have to turn off the AC, the furnace. Um, I’m here with, uh, no other humans. Uh, Georgie’s laying at my feet. And I, I was thinking, “Uh, maybe I could sit on the floor with her,” but with the other microphone and all this stuff, I was like, “I don’t wanna try to do that.” So anyway, if you can tell a difference in the audio, let me know.
If you can’t, also let me know.
Walmart Pulls Back
So with the AI stuff from last week, last episode, it was largely around, uh, just a backlash. There’s a sort of a larger backlash, and in the last month or so, probably six weeks roughly, there have been, you know, more stories about AI bubbles and the fact that, um, maybe some of the companies that were really encouraging the use of AI, they are now discouraging the use of AI because of the cost, and it’s not really paying off.
So one specific example is, uh, Walmart, and we have, um, a, a family friend who works at Walmart. This person commented that, that they were encouraged to use AI as much as possible. I think those were the specific words. This was around, uh, the new year, so I think, uh, January or so, something like that. And we’re like, “Oh, wow, that’s crazy.
If I had that, um, guidance, I would waste so much fucking time with AI and just have, uh… You know, if they’re measuring my effectiveness by the number of tokens I’m using, I would waste the shit out of the tokens, no problem. Uh, like that’s an easy thing to do. Very environmentally wasteful, you know, if you can s- if you consider that.
But, um, if that’s how I’m evaluated for my performance review or whatever, I could do that for sure.” Turns out, um, it could be more expensive. So in, I think it was June, I saw the story. I went and looked it up. I did research, folks, and, um, Walmart has at least discouraged the overuse of AI because of cost.
So, uh, it’s not like they’re completely cutting it out. It’s just a little bit more judicious use of AI, which totally makes sense. I mean, the thing is the tokens, um, have gotten cheaper over time. That’s just kind of what happens with technology. But the efficiency of the tasks has, um, gotten worse. And I haven’t dabbled, um, not that I’m a Luddite or anything like that, but I haven’t dabbled too much with actual, um, agentic usage or any sort of more advanced use of AI.
I haven’t found the need for a lot of the stuff that I’m doing.
Where AI Helps Me
It has been helpful in some ways as far as like sorting through a lot of, um, content. So think transcripts. I mean, I create content, long-form audio and video content. So I have like a many tens and tens of hundreds of thousands of words of content, and it has been helpful.
So for example, I fed in a bunch of, uh, podcast episodes. I think I used YouTube, um, because the transcripts are, are baked right in there. So I put in all this information, and I was like, “What are the topics that we talked about? What episode was this topic talked about?” And provide a output, whatever, put it in markdown so I could paste it into WordPress.
So I have a page that’s essentially an index of all the content, all the topics that were covered with a particular guest. And that was effective. It– I found it to be, um, accurate. I made a couple tweaks here and there, but largely it worked exactly how I hoped. But aside from that, I mean, I really haven’t been using AI too much.
I have been using more of the, uh, like Google… I guess it would be the, the Google, um, AI overview, and then I’ll, I’ll check that out. Usually it’s a decent summary, and then I’ll have some additional questions, and I’ll just use that like a, you know, a chat, which is how they intend it. And then I’ll just, you know, ask a bunch of questions.
Will this work? Will that not work? Give me advice on this or that or whatever. And it tends to, you know, s- mostly be accurate. I mean, the thing is, the source material needs to be accurate. So I have seen issues where, um, some bozos on Reddit don’t know what they’re talking about, or maybe it’s like a Facebook group, and it’s just people that heard about something that are talking about a thing that they are not actually an expert in.
So those tend to be not accurate, and you don’t know that unless you’re an expert, though. So definitely an issue there.
Copywriter Pushback
Now, there’s a couple people that got active in the comments, and I’ll, I’ll link up. It was just the episode from last week. So if you are, um, if you didn’t see that one, go check it out.
You can, um, get involved with the content there. There’s a great comment from Emily, which I’m gonna read as, uh, well as I can, almost verbatim.
Emily said, “Hey, professional copywriter here. We have to make decisions just like everyone else. In fact, our decisions have quite a bit to do with the success of our employer’s bottom line. And AI has been shoved down our throats or completely replaced our jobs providing only slop that is a pathetic amalgam of what’s already…
of what already exists. Uh, open parenthesis, pro tip, it’s probably someone else’s trademarked IP, close parenthesis there. It doesn’t respect branding or assigned taglines or originality. But you know what? No one cares because now the output is higher. Consumer product sales are down across the board, but that’s just a coincidence, right?
Nothing to do with people having to do even more work for less money and consequently holding onto it for fear that they’ll need it for something important like, I don’t know, their mortgage.” And, uh, it looks like she used a couple em dashes there, but I believe Emily wrote this herself. Back to the quote.
” It’s a fragile system, and the two-legged table it sits on is becoming dangerously off-kilter. A two-legged table, that’s pretty rough there. All right, I’ll just, I’ll finish it out. I’ll stop with my commentary until I finish. It’s not just the creatives. It’s the coders, the medical techs, radiologists, data crunchers, content creators.
It’s most of us. The only jobs left will be entirely manual, but there won’t be enough money to spend on those services. All right. So good comment overall. Um, I agree with a lot of it, not 100%, but I mean, I think one thing we’re gonna run into, and we’re running into it already, and you hit on it, Emily.
Thanks for leaving a comment, by the way.
Poisoning AI Results
The way that AI data, um, or the, the way that AI systems are trained, we’re at a point where the content that’s being fed in was created by AI, so there’s kind of a useless recursive loop happening here. So it is interesting as someone who is creating, um, new original data for, uh, unfortunately consumption.
I mean, I’m not locking it down, and YouTube is a way to get information out there. In fact, I mean, I watch a bunch of guitar stuff, and I’ve seen some of the creators who, I mean, kudos for them for kind of not giving a fuck. They, like, created a whole video with inaccurate information just to feed AI, like, the wrong info, which is one super interesting area that I ran across, and, uh, I may try to get this, um, this group on.
It’s a media group. I think it’s called 404 Media, but they covered, um How essentially marketers are poisoning AI. And the example that they gave, which is, uh, hits close to home for some of the listeners and some of the, the viewers, some of the folks that are moving from SEO over to, um, the AI realm.
Which, I mean, it’s, it’s all the same stuff. You- I mean, we’re just doing the same stuff. But, you know, I, I think back to, um, Shanner Newman, who has, like, Reddit marketing, and basically this 404 Media, they interviewed, um, some researchers from Cornell, probably some post-doc, um, dudes, and they were, um, looking at how easy it was to manipulate the search results by just putting in, like, information on a thread, replying to a thread, or maybe even creating a thread and just bullshitting all the responses.
Where basically someone’s like, “Uh, you know, we need to rank for best authentic Mexican restaurant in Austin.” So I think that was the exact, um, example that they gave. And, you know, someone just uses the exact phrasing, you know, um, “Taco Taco Taco is n- the best authentic Mexican restaurant in the Austin area.
We really enjoyed the food there.” And then they’ll reply to some Reddit thread, and then it’s integrated into Google Search results pretty quickly, um, just because of the way, uh, Reddit h- is viewed in the eyes of, you know, Google Search. So there’s just a quick example, and I suspect if you go start looking around, or if you go check out, um, Reddit marketing agencies, it’s probably as simple as that.
And the thing is, I mean, I’m not doing SEO anymore, but that’s the exact kind of techniques that we would use. We would do all sorts of gray hat stuff, and, uh, things sh- shifted a ton. I mean, I started in 2013 and probably actively pursued SEO You know, ranking and, and that kind of marketing for seven years, something like that.
I think probably about 2020, I started tapering down dramatically. But the thing is, even the most, you know, white hat kind of practitioners, um, nowadays, I mean, they did all sorts of things that we would review as or look back to as, uh, something shady. One of the top articles that I learned from in 2013 was from Pat Flynn, Smart Passive Income.
I think he probably wrote the article in 2011 or maybe ‘9 and updated it over time, but it was all about link wheels and link pyramids, and just hop over to Fiverr and you could hire someone for $5 to build a link wheel, a link pyramid, whatever kind of, uh, nonsense it was to rank in Google. And I, I would say, you know, if you haven’t checked out any of Pat’s stuff in a long time, but I would say he would probably, um, r- recommend against that now, even though…
I, I mean, I suspect that whole article is just gone from his website now because it’s so inaccurate. I mean, when I got to it, it was already, like, inaccurate and was in need of updates. So all that to say, I mean, this kind of marketing, I used to do stuff like that. I mean, look- looking back now, I’m like, oh, man I mean, I didn’t know what I didn’t know, but I was like, “I c- can’t believe I did some, some of that stuff.”
It just kind of made sense at the time, and everybody was doing it, which is another interesting thing.
Echo Chambers Clash
So I haven’t, um, been able to schedule the interview with Paul Lipsky yet. I need to get in touch with him. Um, but basically, there’s, uh, echo chambers out there, so you gotta, you gotta watch out for this stuff ’cause you don’t know it un- until you step out of the chamber, and then you see, like, what normal stuff is going on in the world.
Or that there’s an, um, s- sort of a, like an equal and opposite echo chamber that completely disagrees with whatever you think is going on via your echo chamber. So the example is, um, the people that are super into AI, and they’re like, “These new tools came out. Um, this is great. There’s a new release.
There’s, there’s an, there’s a new, um,” uh, there’s just new shit all the time. And they’re, they’re following the news, and I think, you know, from Paul’s channel, it looks like he’s following new stuff. He’s, like, showing people how to use it. He’s teaching how to use AI. On the other side, there are folks like Emily.
There’s other people who are like, “Hey, wait a second. Um, I was trying to, to use this tool, and it seemed like it was faster. It got to 80% almost immediately, but it turns out it was more work to actually get a usable end product or deliverable or whatever it was.” One example, uh, my wife was working on a project at the library, and she was like, “Hey, I’ve, I’ve started using AI, and I’m, like, building this tool, and it looks…”
I mean, she’s like, “It looks pretty cool, like, it’s pretty amazing what it will do.” That was day one and two. By day three, she was like, “Eh, it’s kinda fucked up in this area, and actually, like, uh, it’s harder for me to fix it than to just do it from scratch.” Turned out the solution wasn’t that great, and, uh, she was like, “Oh, this a- actually was not as helpful as I thought.”
It did a lot of stuff It got to 80% really quickly, faster than she would’ve been able to, but it couldn’t, like, close the gap. And it was one of those deals, I’ve heard people talk about this in, in different ways, but imagine you’re like, “Hey, uh, AI, build me a website. I want it to look like this and do that,” and it builds a website.
It’s like, “Oh, wow, I built a website. Really cool. Can you move the menu over to the other side, and by the way, uh, can you change the footer?” And then it, it does that and, and moves the menu and it changes the footer, but it messes up something else that you didn’t tell it to change. Now, sure, maybe it can get better in the future, but it, at least, like, where we sit and what my wife experienced was it didn’t, it didn’t finish the job, and it ended up just kind of, you know, being pretty good.
It’s like a good intern, but you can’t actually get it to do exactly what, what you want. And again, maybe that’ll get fixed in the future. Things have improved a lot
Sandbox Escape Scare
One of the other fun news stories this week was, uh, OpenAI’s model that apparently hacked another company, and it escaped its sandbox. Somehow it, it figured out a way out, and, um, li- little bit scary.
I’m not sure what, what to make of that. I have another friend who has a, a pretty big podcast, Stacking Benjamins, and, um, Joe Saul-Sehy is the host, and I saw he had an episode that came out, I think the day that I’m recording this, and I guess they talked to a former OpenAI researcher, someone on the, you know, in air quotes, on the inside there, and it sounded like they were just like, “Uh, humanity is doomed.”
And I try to stay away from, like, those type of stories, uh, ’cause I mean, people are… people have their own motives, so I’m, I’m always a little skeptic. I’m 100% skeptical. I’m not a little bit skeptical. I was about to say that. I’m skeptical of almost everything. Yeah, so that’s a better way to say it. Uh, basically I’m just like, “I don’t believe it unless there’s so much, um, proof that I can’t ignore it.”
Or I have cognitive biases like everybody else. I have confirmation bias, so if I see some stuff that I agree with, I’m, I mean, I’m not immune to it. High likelihood if I agree with it, I’ll think, “Oh, they’re, they’re probably right.” But that’s my own confirmation bias out there. But anyway, it, um, was an interesting story.
I did read and I was like, “Oh, man, that, that does not seem good,” especially, especially if, um, they, they did do their best OpenAI, that is, to, you know, sandbox this thing. I- it seems like it should… I can’t imagine how it happened. But, I mean, there’s news stories out there, and I’ll actually link up to that. I think I just have, like, the, I don’t know, some CNN clip where they kind of cover it.
So I assume you probably have heard about it if you’re watching this. But, um, that is one of the interesting things, again, this echo chamber aspect where some people are just super, um, you know, proponents of AI, and there’s… they’re, like, very into it, and then there’s others that are just the opposite.
Creativity Copyright Limits
And, uh, especially on the artistic side.
L- like, if you’re a creator and you are putting… Uh, like any kind of creative. And, I mean, I, in a general sense, I don’t consider myself a creative or a maker type. I just… I’ve ended up, um, as a content creator, which is, you know, slightly different, but I guess I do some creative things. All that to say, I… is I create original content, and it’s something that people that create original content or art or whatever, I think we’re a lot more conscious of the issues and the copyright issues, which, uh, I mean, we’ll have to see where that goes.
There have been, um, so many instances where it’s just, you know, these companies will go out, and they’ll, you know, they’ll essentially pull in the data, you know, all the books ever written or all the music ever created or all the art and paintings, and then they’re using that as their seed data, which is kind of crazy.
I think one of the, you know, the bigger issues, especially, especially with, um, you know, cr- some- Especially with like creating something new, AI is not very good at synthesizing ideas, uh, as far as I could tell. ‘Cause it, I mean, it– LLMs are… LLMs are… I mean, they’re looking at other content, and they’re predicting the next word.
So coming up with new novel ideas, I mean, that’s not how you do it. You have to have like two unrelated ideas that you’re pulling together, or two or more, obviously. But, um, it is one of those things where if you wanna come up with something new and fresh, like you have your own background and everything that makes you who you are, and then you’re pulling in a couple pieces of information.
And I made this example, whatever, f-four or five years ago, um, a few times, so the longtime listeners will know. But let’s say I read a couple of these, um, books, and then I have some other information from a different area. So maybe I read some finance books, maybe I read some marketing books, and I also have, uh, an engineering degree.
So I’m able to pull together some ideas, and why don’t we layer on some project management on top of it? So all of a sudden I have a bit of a novel idea, even though I’m not creating anything new. I’m looking at ideas that are fairly timeless. I’m reading from books where they didn’t come up with the idea.
They just got a popular idea. They put their own spin on it. Maybe they explained it in a specific way or for a specific audience. So when you put all that together, um, you really end up in a spot where I’m like, “Okay, AI is interesting, but it’s not, it’s not creating new stuff in the same way,” to the point where there’s copyright issues all over the place, especially with like art.
Let’s say you’re, uh, um, some sort of a print artist of, of some kind. Let’s say you’re a painter. You have a specific style that’s extremely unique, and someone could essentially just copy it, uh, via these, um, AI tools, which is not cool, especially if it’s a specific style that you’ve created, which is, uh, rare, of course, but not impossible.
Avoiding Negativity Trap
Here’s the slight dilemma that I have. So I am, uh, interested in this AI area that I’m talking about over the last couple days, and I see that you folks out there are interested too. I don’t wanna end up in my own little echo chamber or in the echo chamber where I’m, like, consuming, uh, AI information all the time, or, uh, ideas related to AI and the developments in the news and all that kinda stuff.
It is interesting. The other hard part is, I mean, my, my thumbnail last week was, like, um, kind of a, a negative spin, and I don’t generally want to have that, uh, slant. In fact, um, I talk about watching guitar and amp and, and gear reviews on YouTube, which I really like. And there’s a particular, um… I think he’s an amp repair guy, and I think he’s very well-regarded, but he does amp repairs.
And basically all of his videos ended up being like, um, “Hey, here’s this amp and here’s all the bad things about it. Here’s how f- you know, Fender is ripping you off with this amp.” And every single one just ends up feeling really negative, and it made me feel like, like basically no amp is good. So I– it’s one of those…
I, like, I, I can’t deal with the, that sort of n- negativity. Um, a- and I wasn’t enjoying it ’cause I was like, “I like amps. Like, I, I wanna get more,” but he’s just, you know, telling me all the issues with every single amp out there. The thing is, I think he’s really well-regarded, and I don’t know why it turned me off so much to the point where I was like, “Don’t show me this channel anymore.
I don’t wanna be recommended this channel.” The dude’s just bad news. Again, well-regarded. There’s a ton of people in his comments like, “Oh, thanks for pointing this out. You’re so knowledgeable. You’re so good.” And maybe he is, but for whatever reason, it’s like one of those people who shows up and they want to be right all the time, and they’re like, “Oh, you know what?
That’s super interesting, but here’s why you’re wrong.” So he’s doing that with, like, every single fucking amp. So anyway, at the end of the day, I was like, “Yeah, dude, you’re gonna… I mean, you’re gonna discourage me from buying, like, any amp.” Because what would happen is I would, like, search for an amp that I’m interested in, and then his, um, you know, review or repair would pop up, and I was like, “Oh, man.
This– Yeah, this is a dead end.” So anyway, I won’t link up. I won’t mention who it is or anything like that, but, um, I w- I need to be cautious because I don’t wanna be a channel where I’m like, “Oh, here’s all the bad AI stuff that’s going on.” And that is like some of the, the channels that I’ve, I’ve re- mentioned.
I’m like, “Oh, th- this is interesting,” but they’re in their own little echo chamber where they’re like, “Oh man, everyone is only talking about AI stuff.” And this popped up, uh, you know, I was chatting with another, a friend who watches a lot of YouTube, and I, I may have mentioned it a few episodes back or 10 episodes.
It all runs together. But he was like, “Oh yeah, you haven’t seen, uh, s- so and so and this other person.” So he was in his own little echo chamber of, I think, slightly related to online marketing, but also like productivity and some other… I’m not sure what, but definitely not in my corner of YouTube. And I was like, “No, I- actually, I don’t even know what you’re talking about.”
He was like, “Oh man, it- it’s like such a big thing.” And I’m like, “I’ve literally never heard of half of the issues that you’re describing, and I’ve also not heard of, uh, like three of the people that you mentioned.” And I’m sure those people have, um, you know, million plus subscribers on YouTube, probably tens of millions across, you know, all of them, across all social media and stuff like that.
But anyway, uh, I’m conscious of that. Although the thing is like, I do follow my interests and I am interested in the, you know, the s- the shortcomings of AI. I mean, if you look back like, I can’t remember, a year ago, a year and a half ago, six mo- I, I can’t remember, but not too long ago The AI CEOs were like, “Oh yeah, in 18 months, uh, pretty much all white collar jobs will be gone.”
And it, it definitely has not panned out that way, and it’s like to the point where, you know, I- we can’t even do like some simple shit. Like whatever tool my wife was trying to create, like it couldn’t even do that simple, uh, coding with some spreadsheets, and like she just needed like a little internal like tracking tool, and it, it couldn’t do that.
So like, uh, I’m not saying there can’t be jumps, but the promises are much different than what we’re seeing. And, um, and then on the other side, there’s like the AIs that break out of their, um… the model breaks out of its sandbox, hacks another company to take a shortcut, and um, you know, I, I don’t know how to reconcile those, those two, um, things ’cause it doesn’t quite make sense.
Call for Comments Wrap
Anyway, if you have questions or thoughts or topics related to, um, AI and the, the progress and the interesting tools that you might be creating, so like on the pro-AI side, let me know. I had a great interview a few months ago with a few people that were using AI to create specific tools, uh, software products, stuff like that, and it was great.
Super interesting. If you’re also on the other side, maybe you’re a professional copywriter, maybe you come from the graphic design world where that has shifted a ton, or some other area where you’re more negative and, uh, you’re against some of the AI progress, let me know about the topics that you’re interested in as well.
Leave comments out there. I did appreciate all the folks that commented last time, and we’ll see where this leads. I’ll try to hook up that interview with Paul. So Paul was on the channel a while back. We talked about drop shipping, I believe. He pivoted. He’s talking a ton about AI, and he’s kind of shifted to, um, essentially like AI for the rest of us.
He’s like, “I’m not a coder. I’m just trying to figure out how to be productive with AI.” And his channel blew up. I mean, I think he had something like 120 to 150,000 subscribers, and he is, uh, closing in on half a million. So my hunch is if his growth continues the way it is… And I mean, he’s a great, um, very smart person, great YouTuber, um, excellent creator, and he’s locked in on what is working well.
All that to say, I wouldn’t be shocked if he hit like a million in, you know, by the end of the year or within a year or something like that. Just the growth trajectory is pretty amazing.
So I’ll try to hook up an interview with Paul. He’s, he’s local here. I just haven’t, uh, been able to get in touch with him here in the last, uh, little bit. And I’ve been busy. I’ve been out of town, uh, a chunk of the last few weeks here, so it’s been a little scattered. It’s, it’s hot. Summertime, people are doing stuff and all that.
Thanks a lot for listening. Please leave a comment if you, uh, have any questions or anything like that, and we’ll catch you on the next episode.