Are We Close to Perfection? AI in Content Curation and Beyond

[00:00:00] Jess Bahr: Welcome to another exciting episode of Marketers Talking AI, where Pavel and I get together, talk about the latest AI news.
[00:00:08] Pavel Konoplenko: The latest and greatest.
[00:00:10] Jess Bahr: Pavel, I have a request for you.
[00:00:13] Jess Bahr: You're building, like, all this AI stuff.
[00:00:16] Jess Bahr: Can you please build me a platform that will take these recordings and splice it up into clips for social?
[00:00:22] Jess Bahr: Because everyone that I've tried, I feel like, gives me just, like, weird stuff.
[00:00:26] Jess Bahr: Sometimes it cuts off mid sentence.
[00:00:28] Jess Bahr: Sometimes it's a sentence that isn't compelling to stand on its own.
[00:00:33] Jess Bahr: And I have not been able to find a good AI solution that I could just put my entire recording into, and it will just magically give me all the clips for social.
[00:00:43] Pavel Konoplenko: Yeah, there are a few tools that I used.
[00:00:47] Pavel Konoplenko: I think one of the tools that I tried that literally just didn't work.
[00:00:50] Pavel Konoplenko: Like, I never got the clips from them.
[00:00:54] Pavel Konoplenko: And it sounds like a simple thing.
[00:00:58] Pavel Konoplenko: It's like, just look through everything and find something.
[00:01:01] Pavel Konoplenko: So as I'm thinking through this as to what are the challenges that these tools may be running into, I think what we may be seeing is that this is like an overutilization of AI, because this is generative AI.
[00:01:18] Pavel Konoplenko: And this large language model.
[00:01:20] Pavel Konoplenko: AI is really good at basically predicting the next word and creating this content, generating content for whatever topic you want, for whatever theme, with tones and whatever that then the human can take a look at and curate better with your request and what these other tools are trying to solve.
[00:01:40] Pavel Konoplenko: It's kind of doing the opposite.
[00:01:42] Pavel Konoplenko: It's going in the reverse direction in which you're the one giving all of the generated text through the transcript, and then the AI tool has to try to figure out and curate as to what's compelling.
[00:01:57] Pavel Konoplenko: Yeah, I guess I'm curious, underneath the hood, what the prompts look like?
[00:02:04] Pavel Konoplenko: Because these models aren't good at knowing what's compelling content because they're good at, again, mimicking.
[00:02:14] Pavel Konoplenko: They're really good at mimicking and sounding like an expert in something, but I think they're a lot worse at figuring out, okay, this is like, a good opinion, or this is some controversial insightful thing.
[00:02:28] Pavel Konoplenko: So I'm curious if this is just a technological issue because the AI tools aren't there yet, the technology isn't there yet, or if it's just a matter of poor prompt engineering.
[00:02:43] Pavel Konoplenko: If I had to dive into it, if I had to build this tool out, and if this is a request, I guess I have to go in and try to figure it out.
[00:02:52] Pavel Konoplenko: I think there's a lot of guidance that's going to be needed because the tool wouldn't know what's good content based on how you want to come off.
[00:03:01] Jess Bahr: Do you think that if you had a video on a topic that was if we're talking about best practices to make a sandwich, it would be able to clip that more effectively?
[00:03:13] Jess Bahr: Because it's something there's probably a lot of data on versus something where we're talking about like cutting edge news topics that might not be cataloged by any of the modern LLMs yet.
[00:03:24] Pavel Konoplenko: I mean, there's many ways to make a sandwich.
[00:03:27] Pavel Konoplenko: So again, how would the AI tool know which one to go with?
[00:03:32] Pavel Konoplenko: And the way that these tools work and why they are so convincing is that if it's trying to give you a sandwich recipe, it can just kind of try to and this is how it starts to hallucinate.
[00:03:46] Pavel Konoplenko: Because it can start to go down a certain path in its own language, where it may actually know that it's going down the wrong path.
[00:03:57] Pavel Konoplenko: And I'm kind of anthromoporphizing, you know.
[00:04:03] Jess Bahr: The word that I'm yeah, you're making it an animal.
[00:04:06] Jess Bahr: Are you making it human?
[00:04:07] Pavel Konoplenko: A human?
[00:04:08] Jess Bahr: Yeah, you're making it a human.
[00:04:10] Jess Bahr: It wants to be a real boy.
[00:04:13] Pavel Konoplenko: It can't just go backwards in its past.
[00:04:17] Pavel Konoplenko: So it has to continue to spew out whatever it thinks should come next.
[00:04:23] Pavel Konoplenko: So with the sandwich making, if it goes down, like one particular sandwich style, for example, it'll still sound convincing, but it won't be able to fill in what you're trying to say, if that makes sense.
[00:04:43] Pavel Konoplenko: My suggestion for these tools out there, whoever's listening and wants some tips, I think there has to be a lot of prompt engineering into giving examples of what makes something compelling.
[00:04:55] Pavel Konoplenko: And this is a similar thing of even the prompt engineering that goes into tone and style when it comes to generating content.
[00:05:04] Pavel Konoplenko: It just doesn't know what is good when these tools, they're trained on data, and when we think of that, we automatically assume it's good data.
[00:05:16] Pavel Konoplenko: But a lot of the content that it's trained on from the web, it's like some random someone wrote some college essay on some random topic.
[00:05:24] Pavel Konoplenko: The system doesn't know what's actually good.
[00:05:27] Pavel Konoplenko: So that's why one of the tricks or one of the earlier tips for prompt engineering is to act as a world class marketing expert or something, or act as a blah, blah, blah, is because you're trying to prime it into a certain direction so it knows, oh, you want a good email?
[00:05:44] Pavel Konoplenko: Okay, I got you.
[00:05:45] Pavel Konoplenko: Because if you just say, write me an email, it doesn't know that it needs to write like, an excellent email, and it doesn't necessarily or like a business email or a business email, because, again, it just tries to predict the next word.
[00:05:59] Pavel Konoplenko: It's not necessarily trying to solve what you're trying to do, but once it goes down the path, then it's kind of stuck there.
[00:06:08] Jess Bahr: It's like the early days of well, I think early days of SEO is such an easy comparison, but I feel like it's the early days of Search, where you used to be aging myself as I say this, but you used to see commercials, they say like keyword mobile or like keyword TDs.
[00:06:27] Jess Bahr: And so when you go on AOL, you put in your keyword to go find the website for it.
[00:06:32] Pavel Konoplenko: Yeah.
[00:06:34] Pavel Konoplenko: And I think over time, we're going to start to see less of a necessity to try to prompt into what you're trying to do.
[00:06:43] Pavel Konoplenko: It'll know what you're looking for.
[00:06:46] Pavel Konoplenko: That's why I also think prompt engineering isn't actually like a long term job for someone, and I think Sam Altman was talking about that too, not too long ago.
[00:06:55] Pavel Konoplenko: But eventually the AI systems will be better than whatever you're trying to prompt engineer it with.
[00:07:02] Pavel Konoplenko: So I think it's a useful skill to have for the next year or so, but eventually it'll be an outdated thing and similar to the keyword with AOL or even when you go to websites, when you see old commercials for websites, it's WW dot.
[00:07:20] Pavel Konoplenko: No one does that anymore.
[00:07:22] Pavel Konoplenko: You don't even need the domain name.
[00:07:24] Pavel Konoplenko: Just put in what you're looking for.
[00:07:26] Pavel Konoplenko: It's already connected to your Google search anyway, so you don't even need to know the exact domains.
[00:07:32] Pavel Konoplenko: So I think we're going to get to that.
[00:07:35] Pavel Konoplenko: But again, the original request of AI curating human content, I think is the wrong use.
[00:07:46] Pavel Konoplenko: It's not the best use of AI.
[00:07:48] Pavel Konoplenko: I think the best use of AI as we've talked about it, is let the AI generate a whole bunch of variations.
[00:07:55] Pavel Konoplenko: Let the AI give you a bunch of drafts, and you as the human, are then curating the direction from where to go.
[00:08:03] Pavel Konoplenko: But if you're trying to use AI to curate, it's a much more of a hit or miss than the other way.
[00:08:12] Jess Bahr: Yeah, I feel like it's been a lot of miss, a lot of weird stuff, but that makes sense to me.
[00:08:22] Jess Bahr: I think in our youth, we were sold this vision of someday computers will be our perfect assistant that will predict what we need, and we'll run the home and we'll run everything and be super efficient.
[00:08:37] Jess Bahr: And sometimes it feels like we're still where we were in the early 2010s.
[00:08:43] Pavel Konoplenko: Yeah.
[00:08:44] Pavel Konoplenko: And I think there's a lot of hype with AI, so it kind of creates this impression that AI is going to be able to do everything.
[00:08:50] Pavel Konoplenko: So then when it doesn't do something well, you're like, oh, it's broken, but AI doesn't work.
[00:09:00] Pavel Konoplenko: We're worried about them taking over the world in terminator style, and it's like, hey, you can't even find a social clip.
[00:09:07] Jess Bahr: Yeah.
[00:09:08] Jess Bahr: How are they going to take over the world?
[00:09:14] Jess Bahr: When I search myself on chat GPT, it says I don't exist.
[00:09:17] Jess Bahr: How are they ever going to take over the world if they don't even know who we are?
[00:09:22] Pavel Konoplenko: Yeah, but we're making progress.
[00:09:25] Discussion on Chat GPT's new code interpreter and its functionalities

[00:09:25] Pavel Konoplenko: Recently, Chat GPT released a code interpreter.
[00:09:29] Pavel Konoplenko: It's like one of the beta features for the Plus Pro version or whatever.
[00:09:35] Pavel Konoplenko: Really interesting.
[00:09:36] Jess Bahr: What does that do?
[00:09:38] Jess Bahr: Tell me more about that.
[00:09:39] Pavel Konoplenko: So it's extra capabilities, basically.
[00:09:45] Pavel Konoplenko: I've been playing around with it.
[00:09:47] Pavel Konoplenko: One of the use cases that I saw that was interesting.
[00:09:49] Pavel Konoplenko: It can make QR codes, which I didn't think I would be able to make QR codes with.
[00:09:56] Pavel Konoplenko: Chat GBT.
[00:09:57] Pavel Konoplenko: QR codes is one of these things.
[00:09:59] Pavel Konoplenko: It's like there's so many websites that exist that make it.
[00:10:03] Pavel Konoplenko: I don't even have like a go to website, like clients.
[00:10:06] Jess Bahr: Bitly.
[00:10:07] Jess Bahr: Bitly is my bitly.
[00:10:08] Pavel Konoplenko: Yeah, Bitly makes it.
[00:10:09] Pavel Konoplenko: But anytime I've had clients ask me to make a QR code, I'm like, I'm just going to Google it and see who's the most recent ranking company.
[00:10:18] Jess Bahr: QR codes in general?
[00:10:19] Jess Bahr: Amazes.
[00:10:20] Jess Bahr: I think the Pandemic did a lot of good for them.
[00:10:24] Jess Bahr: I definitely remember in 2015, 2016, making fun of someone for having a QR code on their business card.
[00:10:32] Pavel Konoplenko: I actually remember that marketers have had a love hate relationship with QR codes.
[00:10:39] Pavel Konoplenko: If it wasn't for the Pandemic, I don't think they would ever have the sort of popularity they do now.
[00:10:46] Pavel Konoplenko: And I'm curious how it'll be in five years.
[00:10:50] Pavel Konoplenko: What else can code interpreter do?
[00:10:52] Pavel Konoplenko: Apparently it's really good for data.
[00:10:54] Pavel Konoplenko: Well, it's good for interpreting code.
[00:10:57] Jess Bahr: What does interpreting code mean?
[00:10:59] Jess Bahr: Does it tell you in humans speak what the code is?
[00:11:02] Pavel Konoplenko: I guess I'm confused, to be honest.
[00:11:05] Pavel Konoplenko: And I'm not done playing with because it's not like chat GPT couldn't interpret code before.
[00:11:10] Pavel Konoplenko: I'm curious as to why it's like a special feature now, or it's a special thing.
[00:11:17] Pavel Konoplenko: It's functioning similar to the plugins where as it runs the prompt, it runs the code interpreter, which is how the plugins would work.
[00:11:27] Pavel Konoplenko: When you select the plugin to use, you ask the request and then it does the plugin got you.
[00:11:32] Pavel Konoplenko: So it does the code interpreter.
[00:11:34] Pavel Konoplenko: It's also apparently really good for data like data mining and creating flowcharts, or not flowcharts.
[00:11:41] Pavel Konoplenko: Like diagrams.
[00:11:43] Jess Bahr: I love diagramming.
[00:11:44] Jess Bahr: I love flowcharts.
[00:11:45] Jess Bahr: I love visualizing data.
[00:11:47] Pavel Konoplenko: Visualizing it's.
[00:11:49] Pavel Konoplenko: So I created some charts in Chat GPT, which is pretty cool.
[00:11:54] Pavel Konoplenko: And that goes back to the conversation we had last week where marketers one of the top questions that we had was that how can we use AI for data?
[00:12:02] Pavel Konoplenko: So I'm sure if they're asking us that, someone's asking OpenAI that, and they're looking for ways to make it easier for people to interpret data, visualize data, and find the insights, I will be playing with it a little bit more.
[00:12:19] Pavel Konoplenko: I know also there's like SEO capabilities functionalities that you can use it for.
[00:12:25] Pavel Konoplenko: So I'm just excited to see additional improvements to the base model.
[00:12:32] Jess Bahr: Interesting too.
[00:12:33] Jess Bahr: So with SEO, because the current model.
[00:12:36] Pavel Konoplenko: Goes back to 2022 yeah, 2021.
[00:12:41] Jess Bahr: That's interesting.
[00:12:42] Jess Bahr: There's SEO with it because I feel like so much of SEO is sorry, my eyelid, itches there.
[00:12:47] Jess Bahr: But so much of SEO is understanding your current rankings.
[00:12:52] Jess Bahr: I feel like that would be outdated and ineffective.
[00:12:55] Pavel Konoplenko: Yeah.
[00:12:56] Pavel Konoplenko: So now one of the plugins is like a link reader, so it does have some access to the web.
[00:13:04] Pavel Konoplenko: I'm very curious to see how the AI SEO relationship develops because I feel like they're like.
[00:13:14] Jess Bahr: Hundred.
[00:13:14] Jess Bahr: I feel know every person I know who's worked for Google SEO.
[00:13:19] Jess Bahr: They may not say they hate AI, but I feel like they just don't like it.
[00:13:24] Pavel Konoplenko: Yeah, I was listening to something recently and the guy was talking and he made a point that I've said google engineers of whom there are a lot more than of us, and they're more research, more funded, they're aware of what people are trying to do with generated content.
[00:13:47] Pavel Konoplenko: So you can't beat Google at their game.
[00:13:49] Pavel Konoplenko: But then I started to like, does Google actually care if your content's generated by AI?
[00:13:57] Jess Bahr: I don't think so.
[00:13:58] Pavel Konoplenko: I don't think so either.
[00:14:00] Jess Bahr: They care about getting someone a good answer, so they keep coming back.
[00:14:04] Pavel Konoplenko: Exactly.
[00:14:04] Jess Bahr: Which I think is why we see Google like knowledge graph and Google's ability to surface search results continue to increase where back in the day it was just a link click out to another link and now Google's pulling information in.
[00:14:17] Jess Bahr: So I don't think they care.
[00:14:19] Jess Bahr: I think they care about the quality of the answer.
[00:14:22] Pavel Konoplenko: Yes.
[00:14:23] Pavel Konoplenko: And that's the conclusion I'm actually getting to where people are afraid of using AI generated content because oh, what if it hurts our SEO?
[00:14:32] Pavel Konoplenko: But again, I don't think if it's good content, it's good content if it answers the question.
[00:14:36] The Effect of AI Content on Google Search Ranking

[00:14:36] Jess Bahr: You want to talk about a fear mongering headline?
[00:14:39] Jess Bahr: It's not a headline, it's a person talk.
[00:14:43] Jess Bahr: They said it.
[00:14:44] Jess Bahr: I don't know, I'm counting as a headline.
[00:14:46] Jess Bahr: But I have heard so many people say if you use any AI content on your website, your website will be blacklisted by Google and you'll never rank for anything.
[00:14:56] Jess Bahr: I don't think that's true.
[00:15:00] Pavel Konoplenko: It would be a very short term limited decision by Google to do that.
[00:15:05] Pavel Konoplenko: It would hurt them ultimately.
[00:15:07] Jess Bahr: Yeah, maybe it would impact your quality score, your quality ranking for domain authority, but I can't imagine.
[00:15:16] Jess Bahr: I always think it's interesting that there are so many people employed by Google just to work on search when it feels like such.
[00:15:25] Jess Bahr: And I mean, I say that also I've ran PPC for I don't want to say how long, a long ass time.
[00:15:34] Jess Bahr: And I know Google search traffic is so valuable running on Bing.
[00:15:39] Jess Bahr: Also have looked at Brave and a few other browsers.
[00:15:44] Jess Bahr: I almost said Netflix.
[00:15:48] Jess Bahr: No, not netscape.
[00:15:49] Jess Bahr: Firefox.
[00:15:51] Jess Bahr: Wow, that took a minute.
[00:15:55] Jess Bahr: I understand search is important.
[00:15:57] Jess Bahr: I feel like the volume of people who work on search relative to the people that are using it seems so high.
[00:16:06] Jess Bahr: I'm guessing upwards of thousand plus individuals at Google that touch search.
[00:16:11] Jess Bahr: I mean, Google is search.
[00:16:13] Jess Bahr: Google is literally search is search.
[00:16:15] Jess Bahr: That seems crazy to me.
[00:16:16] Pavel Konoplenko: I feel like when people talk about Google sometimes like marketers talk about Google, they might as well be talking about like God.
[00:16:23] Pavel Konoplenko: They're like, if you do this, God will be angry and you will be smitten.
[00:16:28] Pavel Konoplenko: Forever.
[00:16:28] Jess Bahr: Okay.
[00:16:29] Jess Bahr: And if Google search goes away, I don't think it's a big deal.
[00:16:32] Jess Bahr: Go to Bing.
[00:16:33] Jess Bahr: There's other tools out there.
[00:16:34] Jess Bahr: It's not the only option, it's the most popular one.
[00:16:37] Jess Bahr: Now Google.
[00:16:39] Jess Bahr: If Google data center and Google cloud goes away and Google Workspace goes away, it's a different situation.
[00:16:45] Jess Bahr: But I always think it's interesting.
[00:16:46] Jess Bahr: It feels like it's kind of outweighed.
[00:16:50] Pavel Konoplenko: That goes into think.
[00:16:52] Pavel Konoplenko: Yeah, I don't think Google hates AI as much as some marketers are led to believe.
[00:16:57] Pavel Konoplenko: And I think ultimately, if your content is good and it answers the question and people are reading, like, that's all Google cares about.
[00:17:09] Benefits of using Bing Search over Google Search

[00:17:09] Pavel Konoplenko: Speaking of Bing, I've been using the Bing chat app.
[00:17:12] Jess Bahr: Are you liking?
[00:17:13] Pavel Konoplenko: Yeah, like, I actually use it quite often because it has the AI thing built into it and I use the mobile app and it has native web access, which I love because right now I'm searching anytime I have a question, I actually go to Bing now.
[00:17:33] Pavel Konoplenko: Not every time, but half the time I'm actually opening up Bing to search the question or whatever I'm asking, and I'm using that as my answer rather than Google.
[00:17:42] Pavel Konoplenko: And then if something doesn't make sense, then I'm like, oh, let me just go Google it.
[00:17:46] Pavel Konoplenko: I still won't use Bing for the search.
[00:17:50] Jess Bahr: Are you using Bing like you'd use Ask Jeeves?
[00:17:54] Pavel Konoplenko: Yes, I guess, right?
[00:17:56] Jess Bahr: Where with Ask Jeeves, you're supposed to go there with the question, how hot is it in California?
[00:18:02] Jess Bahr: Instead of looking at recent temperature.
[00:18:05] Pavel Konoplenko: California like I'm asking for explain the difference between certain things, give me examples.
[00:18:12] Pavel Konoplenko: So I guess I'm using it more.
[00:18:15] Pavel Konoplenko: So like a search, but like a research.
[00:18:18] Pavel Konoplenko: Like a research?
[00:18:19] Jess Bahr: Yeah, research.
[00:18:20] Jess Bahr: Not search.
[00:18:21] Jess Bahr: Like research.
[00:18:22] Jess Bahr: Research.
[00:18:23] Pavel Konoplenko: That just means that I'm searching against isn't that like never noticed that the word searches in research?
[00:18:30] Discussion on AI Innovations and Tools

[00:18:30] Jess Bahr: Welcome to the Etymology podcast.
[00:18:34] Jess Bahr: What was that word earlier?
[00:18:36] Jess Bahr: Anamorphicize.
[00:18:38] Pavel Konoplenko: That's been a hard word for me because I know speaking of Entropic, the company, they released Claw Two recently.
[00:18:52] Pavel Konoplenko: Also speaking of new innovations happening in Chat GPT or in LLMs.
[00:18:58] Pavel Konoplenko: Yeah, they released it I think like three, four days ago.
[00:19:01] Pavel Konoplenko: Also a big fan, 100,000 tokens, which roughly translates to like 80,000 words, like a big book.
[00:19:10] Pavel Konoplenko: So you can upload PDFs, throw it right in there and you can have conversations about the PDF and it can find insights there.
[00:19:21] Pavel Konoplenko: It can rewrite content and do all of that.
[00:19:26] Pavel Konoplenko: Very cool, very exciting.
[00:19:28] Pavel Konoplenko: Their UI is a little lackluster, I would say.
[00:19:32] Pavel Konoplenko: It doesn't have saved conversations.
[00:19:37] Pavel Konoplenko: It's definitely lacking.
[00:19:39] Pavel Konoplenko: It doesn't give you as much options as the native UI for Chat GBT, but great tool too.
[00:19:47] Pavel Konoplenko: Yeah.
[00:19:48] Pavel Konoplenko: I feel like this past week there's been a lot of new announcements with AI.
[00:19:53] Jess Bahr: We had a low last week because when we were looking for our news articles, it felt like there wasn't really much going on.
[00:19:59] Jess Bahr: The last week or so.
[00:20:00] Jess Bahr: It was like a little bit of a lull, but yeah, now Elon Musk, I think is like the biggest news in AI is launching Xai, which I feel like that used to be a domain for something.
[00:20:13] Pavel Konoplenko: Yes.
[00:20:14] Pavel Konoplenko: So X AI was a meeting assistant.
[00:20:19] Pavel Konoplenko: I used to use it back in 2016.
[00:20:22] Pavel Konoplenko: 2017.
[00:20:23] Pavel Konoplenko: The meeting assistant, her name was Amy.
[00:20:26] Pavel Konoplenko: Amy Ingram.
[00:20:27] Pavel Konoplenko: So AI, but you can change it to Male, so it's Andrew Ingram.
[00:20:32] Jess Bahr: Yeah, I remember this.
[00:20:34] Pavel Konoplenko: Yeah, I was a big fan of it.
[00:20:37] Pavel Konoplenko: And it worked.
[00:20:40] Pavel Konoplenko: It definitely worked.
[00:20:40] Pavel Konoplenko: It definitely had some hiccups, and that was like, my first foray into using AI assistants.
[00:20:47] Pavel Konoplenko: And I got on a whole Hype train about it.
[00:20:50] Pavel Konoplenko: And then I was looking for a biz dev sales assistant that was AI based.
[00:20:56] Pavel Konoplenko: And at that time, AI technology was nowhere near to what it is now.
[00:21:01] Pavel Konoplenko: And I was kind of like I gave up on it, basically.
[00:21:04] Pavel Konoplenko: And I think well, obviously X AI went out of business because the domain got sold to Elon.
[00:21:13] Jess Bahr: I feel like I used now that we're talking about it, I feel like I used it for like a couple months and then just kind of like, fell off of it.
[00:21:23] Pavel Konoplenko: But I remember I really was a, um cause I use like, keyboard shortcuts, so I had like a shortcut for, like, hey, I'm CC Amy.
[00:21:33] Pavel Konoplenko: And then I added into that shortcut, hey, just a heads up, Amy is a bot, so don't expect a human conversation because sometimes people would talk to Amy, and Amy's kind of like, I'm just here to schedule meetings.
[00:21:51] Pavel Konoplenko: Because the tool has access to your calendar, so the tool suggests time slots, basically.
[00:21:59] Jess Bahr: But then some people, it would go back and forth.
[00:22:00] Jess Bahr: A couple of times it would go back and forth.
[00:22:02] Pavel Konoplenko: There's no response at emails, and you can set how many times it could follow up.
[00:22:06] Pavel Konoplenko: Sometimes it would get, like, aggressive, where it's like, Stop emailing me.
[00:22:10] Pavel Konoplenko: I'm not trying to set up a meeting.
[00:22:13] Jess Bahr: Is there a replacement for Amy now?
[00:22:15] Jess Bahr: Is there another company that does that and there isn't?
[00:22:18] Pavel Konoplenko: And that's one of the other shots.
[00:22:19] Jess Bahr: Product idea, product idea, fuck everything else we're building.
[00:22:24] Jess Bahr: Let's bring Amy back.
[00:22:26] Pavel Konoplenko: It worked.
[00:22:26] Pavel Konoplenko: And I know plenty of people who are paying for it, so clearly like, yeah, call her.
[00:22:34] Pavel Konoplenko: Irving.
[00:22:38] Jess Bahr: I remember this, and I think it was like, hey, Marvin or Alfred or something.
[00:22:44] Jess Bahr: And there was another one, it was similar, but they would use AI to help automate redundant tasks.
[00:22:50] Pavel Konoplenko: Alfred, I think, was a tool for Mac.
[00:22:58] Pavel Konoplenko: There's Merlin.
[00:23:00] Jess Bahr: Merlin.
[00:23:01] Jess Bahr: There's one because I remember I used them because I left something at a hotel in South Beach and I had them call up and get it shipped to.
[00:23:11] Pavel Konoplenko: Yeah, it's like an assistant that it's like a human AI assistant that.
[00:23:16] Jess Bahr: You can yes, it was a human that would interface with you and would make the call.
[00:23:21] Jess Bahr: I think it would but it would use the had it because I was taking a trip to Miami and so I had it find a place to get our nails done on South Beach that was open and had availability during a certain period of time.
[00:23:34] Jess Bahr: I would have it cancel reservations for me and stuff.
[00:23:37] Jess Bahr: And I was like, how does this work?
[00:23:39] Jess Bahr: Because it's priced really well, it was really inexpensive.
[00:23:43] Jess Bahr: Yeah, we use AI to do a majority of the work and the human just does like the last mile they went out of business.
[00:23:53] Elon Musk's New AI Company and Potential Impacts

[00:23:53] Pavel Konoplenko: So Alfred, apparently this is the other thing, there's a whole other thing for Alfred on Indiegogo.
[00:24:00] Pavel Konoplenko: It raised 65,000, but I used a competitor of Alfred, I think it was called Magic.
[00:24:05] Jess Bahr: Yes.
[00:24:07] Pavel Konoplenko: There should be a list of all of the names that oh my God.
[00:24:11] Jess Bahr: We should make a grave.
[00:24:12] Pavel Konoplenko: Jasper, Jeeves, just all the names that have existed.
[00:24:18] Pavel Konoplenko: Clippy but yeah, I think the Elon Musk news is interesting.
[00:24:24] Pavel Konoplenko: Obviously he's connected with OpenAI and he's one of the biggest outspoken names in entrepreneurship today and Tech specifically.
[00:24:34] Jess Bahr: He's talking about Terminator future with the AI.
[00:24:36] Jess Bahr: I feel like he's not but aren't.
[00:24:38] Pavel Konoplenko: He building isn't Tesla building a robot?
[00:24:40] Pavel Konoplenko: Didn't they robot?
[00:24:45] Jess Bahr: If anyone is creating a terminator, it's Elon Musk.
[00:24:50] Jess Bahr: He is going to bring the apocalypse forward and he's sitting here acting like he's I didn't, I haven't watched Terminator movies in a long time, but he's sitting here talking about saving us from the future.
[00:25:01] Jess Bahr: While he's building Skynet, he is going to build Terminator and if anyone causes the destruction of humanity, it's going to be Elon Musk.
[00:25:10] Jess Bahr: He's going to be in Mars, save us and he'll be in Mars.
[00:25:13] Jess Bahr: His bitch ass is going to go to Mars.
[00:25:15] Jess Bahr: He won't even be here for it him and his employees that he's been.
[00:25:19] Pavel Konoplenko: Impregnating, all twelve of them.
[00:25:22] Pavel Konoplenko: Get the twelve kids or something now.
[00:25:24] Jess Bahr: Yeah, because we're just going on a total tangent right now.
[00:25:30] Jess Bahr: Elon Musk is of the school of thought that intelligent people have a social and a moral obligation to reproduce.
[00:25:38] Jess Bahr: And so Elon Musk and there's articles about like I'm not being conspiracy theory nut here.
[00:25:43] Jess Bahr: Elon Musk has been intentionally reproducing and typically reproducing twins with employees, with executives at Tesla.
[00:25:54] Jess Bahr: He's been having babies, making babies, because he feels that it's an obligation as a smart person to reproduce.
[00:26:05] Jess Bahr: Yeah, I said it's like the greatest threat to humanity birthrights.
[00:26:10] Pavel Konoplenko: Yeah, that is a threat, but I could see how you reverse that ideology.
[00:26:19] Jess Bahr: And you get eugenics 100%.
[00:26:23] Jess Bahr: Yeah, no, it's fucking creepy.
[00:26:27] Jess Bahr: And also it is because he's saying more people like me need to reproduce, not people who are currently reproducing.
[00:26:32] Jess Bahr: Yeah, and there's a lot of issues with it, but I just think it's so funny that he's the one who's saying AI is going to cause the destruction of the world, but I'm here to save it.
[00:26:43] Jess Bahr: And I really genuinely believe if anyone creates a technology that is going to destroy the world, it's going to be Elon Musk.
[00:26:49] Jess Bahr: And he will do it while he's trying to create a technology to save the world.
[00:26:53] Pavel Konoplenko: Yeah, we could have a whole podcast about Elon Musk and this bill Gates.
[00:26:59] Jess Bahr: Is just chilling there.
[00:27:01] Jess Bahr: Yeah, Bill Gates is just chilling going like, hey guys, it's not going to be that the future.
[00:27:05] Jess Bahr: Isn't that like, calm down, let's get our feet grounded a little bit here.
[00:27:13] Pavel Konoplenko: Yeah.
[00:27:13] Pavel Konoplenko: I'm curious.
[00:27:15] Pavel Konoplenko: Where X?
[00:27:16] Pavel Konoplenko: AI or X?
[00:27:18] Pavel Konoplenko: Actually, I don't know the exact name of the company.
[00:27:20] Jess Bahr: Is it just called x.
[00:27:22] Jess Bahr: I think it's just Xai.
[00:27:24] Jess Bahr: The company is X AI.
[00:27:26] Jess Bahr: Xai, which is associated with but not a part of X, which is Twitter.
[00:27:32] Pavel Konoplenko: Yeah, the X core, which is their whole thing.
[00:27:37] Pavel Konoplenko: I think it makes sense that he's doing this.
[00:27:40] Pavel Konoplenko: He's clearly involved.
[00:27:41] Pavel Konoplenko: I'm curious which direction they go into, but I think ultimately it's definitely good for innovation, for there to be more money and more research and more smart people working on it.
[00:27:53] Pavel Konoplenko: And not just like when I say smart people, not like Elon Musk per se, but the team that he's hiring.
[00:27:57] Pavel Konoplenko: Not to say he's not smart either.
[00:27:59] Jess Bahr: But we're saying it.
[00:28:01] Jess Bahr: Elon Musk is a dumb dumb.
[00:28:05] Jess Bahr: The opinions we share are not reflective of our employers or the show.
[00:28:10] Jess Bahr: Disclose that.
[00:28:14] Jess Bahr: Did you ever follow Neuralink?
[00:28:17] Pavel Konoplenko: Yes, I've been following it.
[00:28:20] Pavel Konoplenko: They got the FDA approval.
[00:28:22] Pavel Konoplenko: I have a whole other take on that too.
[00:28:26] Pavel Konoplenko: I think it misunderstands the very nature of thought and language and how they're related.
[00:28:31] Pavel Konoplenko: Language is the formation of thoughts and you can't just bypass language.
[00:28:39] Pavel Konoplenko: Now we're getting into like, philosophy, but language carries with it thought.
[00:28:45] Pavel Konoplenko: So if you remove language from the communication, you can't have, I think, one without the other.
[00:28:51] Pavel Konoplenko: So this idea of like, you're going to just pass your thought without language seems misguided.
[00:28:57] Pavel Konoplenko: But I also feel like this is a very human thing.
[00:29:00] Pavel Konoplenko: We want to try to reduce the friction we have with communication, but the very nature of the universe is that the closer we get to discovering there we have it.
[00:29:10] Pavel Konoplenko: We figure out the atom is actually other things, and then these other things are actually other things and it's a never ending process.
[00:29:18] Pavel Konoplenko: So even once we get good at this neuro language interface, we can actually figure out, hey, there's some other pathway of thought that we have to solve.
[00:29:27] Pavel Konoplenko: So we'll never even get to this point of frictionless communication.
[00:29:31] Pavel Konoplenko: And I do think that friction of language is necessary because that friction is the spark of it all.
[00:29:44] Pavel Konoplenko: Yeah.
[00:29:44] Pavel Konoplenko: The universe has a way of being mysterious beyond just being able to capture it completely.
[00:29:52] Jess Bahr: I think his AI stuff will tie into Neural link.
[00:29:57] Jess Bahr: I don't think he'll publicly talk about it, but I think it'll tie in.
[00:30:02] Jess Bahr: It'll be something related to trying to leverage AI, unlock 100% of your brain power.
[00:30:13] Jess Bahr: I think there's going to be some tie in with it.
[00:30:15] Pavel Konoplenko: I think there's definitely going to be a tie in to Tesla because for self driving cars, like being able to understand and predict, you need artificial intelligence there.
[00:30:26] Jess Bahr: Actually, they have some really cool technology in their cars when you're driving.
[00:30:33] Jess Bahr: So a lot of vehicles for self driving vehicles, it will see the car in front of them.
[00:30:38] Jess Bahr: Tesla actually uses lasers that bounce off the ground to see multiple cars in front.
[00:30:45] Jess Bahr: As much as I just called Elon Musk a dumb dumb, it's actually pretty advanced.
[00:30:50] Jess Bahr: Yeah, 100% tie into it.
[00:30:52] Pavel Konoplenko: Yeah.
[00:30:53] Pavel Konoplenko: But I think, again, that's the nature of AI.
[00:30:56] Pavel Konoplenko: It's a very foundational technology, so you can use it for everything.
[00:31:01] Pavel Konoplenko: On some level, we're going to have.
[00:31:04] Jess Bahr: All these kids that don't know how to butter toast.
[00:31:08] Jess Bahr: They're going to be, ah, what do I do?
[00:31:10] Jess Bahr: And then Al be like, get your peanut butter out.
[00:31:12] Pavel Konoplenko: Well, the robot will just make it for you.
[00:31:16] Discussion on the Shutdown of an Amazon Warehouse Robot

[00:31:16] Jess Bahr: Q, did you see the video of the Amazon Warehouse robot that just shut itself down?
[00:31:22] Pavel Konoplenko: No.
[00:31:23] Jess Bahr: I'm sure it encountered, like, an error, but what people are saying is the robot realized that its job was just to move boxes and so it killed itself, which I think is, again, anthromorphicizing.
[00:31:36] Pavel Konoplenko: Look at you, robot.
[00:31:43] Pavel Konoplenko: That's funny.
[00:31:43] Pavel Konoplenko: The Internet will say that the Internet.
[00:31:46] Jess Bahr: Always goes to like they'd rather believe a conspiracy theory in some outlandish doom and gloom case than the idea that the code had a bug and it had to reboot.
[00:31:56] Pavel Konoplenko: Yeah, it's the equivalent of when your laptop freezes.
[00:32:00] Pavel Konoplenko: You're like, oh, I guess I shouldn't be working on this project.
[00:32:04] Jess Bahr: Yes.
[00:32:05] Episode Wrap Up: Open Discussion and Future Plans

[00:32:05] Jess Bahr: On that note, we're actually at time for this episode.
[00:32:09] Jess Bahr: Thank you, everyone, for joining us.
[00:32:11] Jess Bahr: If there's any topics you want us to discuss, any news, send it your way.
[00:32:14] Jess Bahr: We're happy to give our opinion on things.
[00:32:16] Pavel Konoplenko: We have plenty of those.
[00:32:19] Pavel Konoplenko: And if we don't, we'll ask Chad GBT for some opinions.
[00:32:22] Jess Bahr: See you all next week.
[00:32:24] Pavel Konoplenko: Bye.