Hey channel insiders, welcome back to channel insider partner POV. I'm your host Katie Bavoso and I'm welcoming two very special guests to the show today. The hosts of our sister video podcast, The Neuron, Grant Harvey and Corey Nolles. Victoria Durgen, managing editor of Channel Insider, also returns to help me co-host. We are so excited to talk to Grant and Corey about everything AI and how the neuron has become one of the fastest growing and most popular AI podcasts in just the past year.
We'll dig into everything from what are the major AI companies and tools to watch, what's influencing AI adoption, and are we in an AI bubble. I am so excited to welcome today's guests. We have a familiar face obviously returning to the Channel Insider partner POV video podcast, Victoria Durgen, managing editor of Channel Insider. But joining us today, we have the two co-creators, co-hosts, co-extraordinaires of our sister podcast, The Neuron, Corey Nolles and Grant Harvey.
Welcome, guys. I'm so excited that you're here and that we get to pick your brains about all things AI today. Corey, how you doing? How's it going over there for you, >> Katie? Thanks so much for having me. Uh doing great today. Really uh excited about this and uh always fun to come talk AI. >> I'm very excited to have you on the guest side of it so that we can get your perspectives on so much today. And Grant, you've already corrected me on so many things when I was like, "Hey, great to talk to you oneon-one." And you're like, "Actually, it's threeon-one, but that's okay." So, I assume you're feeling a little spicy and ready to go today. >> I am not I'm not trying to be combative, unless that's the energy you want, in which case I will bring it. >> I am here for it.
Yeah. If we could sort of have that mid-after afternoon cable news like round table where we all start yelling at some point that'd be great. So >> it's all love here on channel insider partner POV but thrilled to be able to have this opportunity. So thank you both for taking the time. I know you've got very busy schedules because as many people know who are watching this and are familiar with the Neuron, the Neuron podcast was relaunched in 2025 and now you're in the top 100 AI/Tech podcast on Apple podcast, which is wild to see how you've accomplished that in the past year.
Congratulations on all the success there. And I got to ask, what was the goal from the beginning to differentiate the series in this crowded tech space? Obviously, a tech space that I'm in as well, getting more crowded by the minute. So talk to me about the original goal of the podcast and how has that evolved to this point of where we are today with the neuron. >> I would say that you know one of our goals was to sort of create an environment where it's okay to not know everything yet.
You know there are so many people right now trying to understand these tools hearing from their boss that they should be using them should be using them and and don't even know where to start. And when it comes to learning about it, when you go online, you see a lot of really technical podcasts, a lot of really technical YouTubers diving into coding and a lot of things like that. And while we touch on those things, we we try to make sure that everything we cover is as relatable as possible to, you know, someone who works in project management and has no reason whatsoever to know what inference means, for example.
And uh we really try to kind of speak to that niche as much as we can and not be afraid to admit when we don't know something because that happens frequently. >> The Diana of the AI podcast, the people's podcast, >> the people's podcast, >> the people's podcast. Uh yeah, I mean I I would I would echo what Corey said. I would also say that you know like when the neuron was acquired by technology advice um previously one of the co-founders Pete had done a version of the podcast where it was just him talking um giving you the news like kind of a deep dive on news items of the week or one specific item and it was really good and we we were originally going to reboot that but then the minute that Corey and I met each other like the f very first onehour call we were like did we just become best friends?
It was it was very it was very Stepbros and uh I immediately was like oh well we need to make a podcast that's just you and me talking like that would be great. So >> we could do this for hours. Yeah. >> Yes. And we have many days. >> Many many hours now. >> There was also a selfish component where it was just like yeah we kind of just want to do this because it's fun and we think people will benefit from it and yeah it's a fun format. So, so that, you know, props to Corey for pitching that and and and pushing that version of it forward.
And folks who go to the YouTube channel, you can see there's one video I did where I tried to do the the Pete like YouTuber style video of the week and and we very quickly moved on to this format, which is like a lot better. >> It's always when you go back to watch that first attempt or that first episode that you go, "Oh my gosh, thank goodness we we woke up and realized this wasn't going to work." You you also have a lot of people that are featured on your podcast now, the interviews.
Was it that people reached out to you to start being featured on the podcast or were you reaching out to them or kind of a culmination of the two? >> Both. You know, we've always gotten pitches because of the newsletter, I would say. Usually thinking of something for a written article. So, I mean, there there are always people that have been pitched at us. However, since we started and did those first couple episodes, like I I think our first interview was with Microsoft if I'm not mistaken. >> And um >> solid first interview, >> right? >> And and from there it did better.
And and the truth is the better the guests got, the better the uh offers are that come. And we don't do a whole lot of outreach right now. We're getting a lot in that are fascinating. Uh with that said, we do outreach as well. Well, one of our biggest guests actually was the head of Microsoft AI, Mustafa Stimon, and that was we just literally cold pitched their his PR team and we said, "Hey, he just published this blog about is a can AI ever be conscious and why we need to be careful of that.
We'd love to have a conversation about that." And they were like, "Yeah, when are you free?" So, we went with it. Yeah. >> Yeah. Sometimes you just got to ask. >> Yes. As annoyingly simple as that sounds, I think what I've learned in the process of building Channel Insider as a site and everything is like sometimes if you just go, "This is crazy, but I'll just do it." Those are the ones that get back to you the quickest. I'm like, "Oh, >> yeah. >> Okay. >> It really is. >> What are you going to do?
Email your boss and go, "How dare they ask me for my time?" >> Yeah. Right. >> How dare they? Yeah. and and if they put a blog out about it, the the the thing that they want is the ability to go out and talk about that viewpoint, which is such a cool one to begin with. But I do agree there are those like 50 layers of of PR representatives and legal representatives before you get to some of those executives and then it's so surprising at certain companies where you just reach out and you hear back from that person immediately and you go, "Were you just waiting for this?" You just sitting by the computer by the inbox waiting.
So incredible there. I'm curious too, you guys, as you kind of have built the podcast over the course of the year. I'm sure you kind of felt on things as you went. Were there lessons that you learned kind of early on in setting not just topics and kind of Cy to your point, what you wanted to be known for, but also who you brought on to interview? I mean, there's no shortage of people who at least say that they have an opinion on AI right now, right?
So, have you seen that kind of grow and evolve over the last year? And and how have you kind of landed on some of that diligence work before something gets to your audience? And do you have any favorites? >> That is very much like like favorites that we have interviewed already. >> Yeah. >> Oh, several. But first on the other, it's very much a process we are still figuring out like like prior to this call, Grant and I had a call about exactly that today.
And uh and it's a subject that comes up every week because you know the truth is you know in our ideal world we'd have every one of these conversations because they're all interesting. But the fact is there's only so much time. We only have so many resources to get additional things edited and pushed and promoted and pay and attention. They need a little room when you upload these videos to to get some traction before you bury it with the next one as well.
And the thing that's been a struggle there is is we've said yes a lot. And uh as a result, we've had a lot of really fascinating conversations. In fact, our two best performing interviews uh were both things that just kind of came to us that sounded cool to both of us. As far as my favorites, that's so hard. That's so hard. The one I was probably our first interview with Open AI was pretty cool. I remember that. >> Why was that your favorite just because it's Open AI? >> It kind of felt like like a we're making it moment if that makes like this is working.
This is working. Uh because we'd we'd already talked to Google. We'd already talked to Microsoft, you know, and there are still labs we're not making headway with yet, but we're coming for you if you're listening. Anthropic, check your email. >> We're trying. We're >> insider partner POV and not the neuron. We're going to get them all. >> That's right. >> You're welcome. >> Grant, same for you. What is your favorite, Grant? Do you look back at your podcast and and go, this interview, whether it performed awesomely or not, this was a top 10 for me? >> Good question.
Uh definitely the interview that we just did with Zuzanna from Pathway that was a really cool conversation not only because like it clearly resonated with people but because um they're working on something really interesting. It's like a what's called I don't we won't get too technical here but they're calling it a post transformer uh model. So it's like a language model that you know like chatbt but they changed some stuff and in order to try and make it work more like a human brain.
Uh so obviously that's pretty exciting and interesting. And then uh there was actually been some sleeper hits that should get more attention. The OpenAI one was really good because we talked about their super intelligent math math. >> That was Ahmed Elish. >> Yeah, that was a really good one. >> Yeah. And that was really cool. We talked about the future of coding >> and even about what it was like to work at OpenAI when it's exploding like it is.
Like what's that like? It's got to be crazy, right? >> Yeah. >> Yeah. And it was one of his first interviews that he'd ever done, actually. So, it was it was cool to talk to him. >> Oh, good. What a question to hit him with right there. Like, so what's the job like for you there? Talk to me about everything on the inside. >> He was fine with it. >> Yeah. Leak it all. >> He told us stuff that No, >> they told they told us nothing. He was He was >> They told us nothing.
He was very professional. >> Very. >> That is really cool. I I I got to know like too as we you talked about it in the beginning, Corey, where you were listening to what was already out there and it wasn't a whole lot of what the everyday user of AI could kind of take away from the news, from the other podcasts where it was very deep into coding or very deep into prompting and very deep into engineering. So, who are your viewers typically? Like, do you hear back from them?
Do you hear from them at all uh when they tell you who they typically are? So, who's your audience? It's broad is what I would say is is it's very broad. We we absolutely do have, you know, engineers and researchers and executives at at large companies and things like that, but we also have a ton of people who are figuring it out, trying to find their place. They're maybe maybe they're trying to figure out how to make money with AI or how to use it in their job to be better at what they do and and to move faster.
And we do our best when picking guests and uh having conversations and even asking questions to try our best to frame them in a way that is accessible to as much of the audience as we can. Like you know it doesn't work for every question, but as as much as we're able to, we try really hard to make sure we do that. >> I'm also curious thinking about when we talk about AI, there's so many players in the market, right? We have big players in the AI space that have also been around for a long time.
I was talking to Victoria about this before we came on. I kind of separate them in from old school and new school. So when we think about old school, we have AWS, Microsoft, Google, companies that have been big names for a long long time having nothing to do with AI and are now very big players in the AI space. And then we have the new school, OpenAI, Anthropic, Grock. So when you think about the companies that are leading the market right now, who are they and why in your opinion have they been able to influence so much in this space, even if they're not necessarily the Kleenex of the AI space, the one we all may be thinking of for each of you, who do you think is the company that really comes out on top right now? >> I think, you know, without question, as of January 20th, 2026, and Anthropic is the company with the heat. leaving 2025, everyone thought it was Google um because Google had basically a absolutely winning combination.
They had some of the best models, you know, for the first time ever that on all of the benchmarks. They released two really good models back toback. One was Gemini 3 Pro, one was Gemini 3 Flash. And then they had all of this um hype about their chips that they were producing as well, which is the the Ironwood TPUs is what they're called. And so everyone was like, "Oh, Google is crushing OpenAI now." But then something weird happened over Christmas break where all of a sudden everybody got to play with cloud code for maybe the first time ever and non-engineers were building software and losing their minds over it on Twitter and other places and all of a sudden it got picked up by Wall Street Journal and all these other mainstream podcasts.
I even heard uh if anyone listens to Bloomberg OddLots, Joe Weisenthal on there was like making software with AI. So it has gone truly mainstream and I think you cannot take that away from the design of cloud code. Like it is just a very delightful tool to use and even though it is um something that's very intimidating. It's a command line interface tool. So you have to actually go into the command line of your computer and with no UI and and and talk to it people people can use it and and find a lot of value out of it.
So I think they're they're winning for sure. But I'm curious to hear if Corey has a rebuttal to that. I would say that that is absolutely true today and that it could not be true by the time we eat dinner tonight. Uh that's kind of that's kind of how the AI space works. Uh and u I would argue that probably OpenAI had a stronger December that we kind of went from Google to OpenAI into the big cloud code explosion that we're seeing right now. Um and the truth is it changes from day to day. uh you know quite often you know one new model drops and suddenly someone else is in third place or fifth and and you don't know uh but I I think we really do a disservice by by thinking of there's such a desire to to have a winner and I I think the truth is we're at a point right now where for the first time since this started they're all good you know whether you're dealing with their their chat interface their API whether you're dealing with their command command and line interface tools.
The fact is every one of them now is good. Like so so incredibly much better than they were 3 years ago when all of this started. Yeah, absolutely. At the moment I'd say it's anthropic and uh and I would put Gemini in third right now. Uh [clears throat] because Codeex is doing really well. CHBT dropped 5.2 which is pretty significantly outperforming Gemini 3. So, you know, I I really think that I would put anthropic open AAI Google right now, but I'm fully prepared for that to change before tomorrow.
And, you know, these legacy companies though that you mentioned, I want to I'd like to address too because some of those are just now getting around. You know, I think I think we'll see some moves out of Microsoft. Who knows if Apple really do anything or not. They, you know, they they've kind of tap danced around it all along. I think we'll see them maybe Meta make a bigger splash this year. [snorts] >> It's interesting because we are not in the kind of weeds of AI that you guys are from like a model standpoint, right?
So when I hear that question, I don't think necessarily in technical capacity so much as I just think in terms of like well what are people using? Like if I go text my friend group chat or if I walk downstairs in my, you know, shared house with my parents and I name drop a brand and they know what that means. That to me is like, wow, that brand is really >> Yeah. >> it, right? And I still think Open AI is like the one name that people across the board know. >> I think that's fair. >> That doesn't mean that they understand how to use it, right? like that.
I think that's where a lot of this starts getting kind of in the weeds is like, >> yeah, >> I don't know [clears throat] if anybody has a shared definition of better, >> right? >> Yeah, that's fair. That's well that that extends up to the definition of AGI, right? Which for people who don't know, AGI is artificial general intelligence, which is theoretically an AI that, you know, can do anything a human can do. It can generalize. So, it can it's smart, but that means it can like figure out things like we can't.
Well, everybody has their own definition of what actually AGI looks like. >> Some of them are crossed. >> Yeah, it's kind of this thing that's impossible to define because everyone's like, well, I, you know, depends on your definition. And it's like, well, can we decide on a definition, please? >> Well, actually, I'm glad you said that though because I feel like so much of the AI conversation, like there are times when we've talked just in passing or whatever at work or you've dropped information in our kind of company Slack chat and I've been like, why do I care about that though? like why is it cool that the model is new?
Why is it important that the inference has changed? Like yeah, >> and I feel like if you know those things, you get really into them and you can answer that question with your kind of top five and here's why they're moving and here's how things are working and I'm much more on the side of the fence. It's like they say they're cool. Okay, I don't really know what I'm doing with them, but like that's great for someone. >> That's a really good point. >> And you know, for me, I think that boils down to the new things you can do.
Like maybe that's the ability to not just generate a photo, but to edit one that exists. Like we're at a point now where I've taken black and white photos of my family and made them color and had them cleaned up where they look brand new and you know and just done that on my phone. And and it's the same with the models that control a chat GPT which which I'd argue is probably going to be the Kleenex of AI in terms of like that's just what people will say at some point.
Yeah, without without question. The Google of AI >> the Google of AIT. Yeah, >> I think some of those other abilities, while subtle, and maybe the average person doesn't know them, it's why I always encourage someone who who you can tell has like a 2-year-old opinion of how bad AI is. Like, you should really go back because the fact is, you know, in the last 12 months, it's night and day what's what's possible and what wasn't. Like, we look at it through a microscope, which sometimes makes it feel like things aren't happening.
But the truth is when you step back and look at the macro of it all and see the growth like you know this last year OpenAI went from$2 billion dollars in ARR in 2023 to 6 billion in 2024 to 20 billion in 2025. So, like we're talking about exponential growth on the revenue side, on the compute side, and um they're all kind of seeing that, you know, it it appears it's not just them. And I think that comes from new abilities that we didn't know we were waiting on in some cases. >> My question would be for for you, Victoria and Katie, are you hearing people like out in the real world, whether they're, you know, family members or friends, talking about anything other than chat GBT?
Like, are you hearing gentlemen or are you hearing husband talks to Gemini more than he talks to me some days. Like, >> yeah, that sounds funny. He's probably listening to this like, "What are you doing?" Uh, no, but it it is something. He has a an Android phone. Is more specifically a Google phone. So, he is always speaking to Gemini, which is already built into that phone. I don't have that ability with my Apple. I'm not going to say her name because she'll interrupt really rudely, but she, you know, I can't talk to her in that same way.
Um he's asking Gemini and showing it real time images of how to install a vanity in our bathroom. He did, you know, his own plumbing from scratch using that and it was like using tools and having to saw things out of the wall. It was impressive and he successfully did it because Gemini helped walk him through it. So, I mean, I I personally think Google has the best opportunity here when we talk about the future of AI because of the resources that it has and because of the the content and the platform that it already has out there.
Google, as we've just been saying, is the number one search platform and it's already building Gemini and other AI tools into that platform that we all use every single day. I mean, how often are you googling something and you're just reading the top bit, which I'm not saying do this everybody. I know we all have to check our sources, but how often are you just looking for a quick answer and you're getting that top bit of information? So, I think Google has the best opportunity and I would not count them out.
Uh, and yeah, so personally for me, I I see it in my household much more often using Gemini than using Chat GPT or any other model. >> Wow. >> Yeah. >> You know, and that's not a thing I I see much. And and you know, I do think Google has a massive opportunity. You know, the truth is they can burn through money for a long time and still be viable. The the flip side though is they're also the most at risk in my opinion and I say that because of the ad business.
Everything they have hinges on their ad dollars and their ad dollars which are the core of their business model now depend on searches not just the health of search but the growth of search you know like new things need to be being discussed and searched on Google and and it's being used less. I mean, I mean, it's still the market share to have for sure, >> but uh they've taken a real hit over the last two years, especially. And uh >> I'm not saying they're making the best decisions, but the opportunity, >> but Katie has been figured out how to do plumbing, guys. >> Yeah, but like let's really focus on that.
Yeah. No, the other day, I mean, I had a plumbing related issue where I used TrabT where I was basically there was a clog in in my sink and there was a disgusting red algae plume that had come up and I was like, "Am I going to die?" And so I was filming Yeah. I was filming it and I was, you know, talking through it with OpenAI and I was like, "This is what I'm smelling like, you know, am I okay? Should I air it out?" And it was just giving me advice like, "You're probably not going to die.
Like here's what's going to happen. like it's probably this like you you should wear a mask and >> your limbs are going to turn black. >> Yeah. >> Yeah. Was talking me through how do you do webs out of my hands? When is that going to happen? >> Yeah. But it just it gives you that confidence to do something that maybe you wouldn't know how to do. And sometimes that can be a bad thing. But if it knows what it's talking about, it's really really helpful >> if it's electrical work for example.
Uh you know though I I use it for a bunch of things like that. And and Katie, what your husband's doing is something that I feel like the average person is really missing out on right now. Like there are a lot of normal household things that that AI can be super helpful for. Like like how long do I cook these baked potatoes? I've got three of them. They weigh about a pound a piece. And every time I do them, they're never finished no matter what I do, you know?
So I'll deal with it with cooking with a with a car issue. I've I've, you know, recorded audio to let it hear the noise before. work on things like that. And I think that's where individuals can feel really empowered by it in a lot of ways. >> I realize that I could have saved so much money on a car issue recently had I just asked one of these LLMs what to do if I recorded that video. That that actually this question I want to expand on it a little bit more, Corey, because you were just starting to talk about how you use AI in your day-to-day.
So I want to hear from both of you, Grant and Corey, how you're using it professionally and personally. And since Cory, you kind of kicked us off, like what are you doing for your for your dayto-day at work using AI? Are you using it or is it more so in your personal life? >> All day every day. >> Okay. >> I I'm literally uh if I'm being honest, there are probably four windows up like in the morning, I have chat GPT agent up scouring for the last 12 hours of AI news to bring me back a report with links and why it might be interesting to pass a grant to include in the neuron.
I'm using it to help with memos and documents and factcheck and proofread myself. I'm using it to help plan videos. I'm using it to help send videos down the line to the next people. And it's not much different in my free time. Like I use I use Chad GBT's advanced voice mode a lot where I'm talking out loud back and forth with it. And uh a lot of times I'll be going to get coffee and I'll have a conversation where I'm trying to work through an idea or I've got a I've got a plan and I don't like it.
I think it can be better and I'll use that time to have a good back and forth about it or or have it read me the news, you know? I I do that some days too while I'm driving. Like uh >> how does that work? Do you do you start the conversation before you start driving? >> You Yeah, I I don't like start it while I'm driving. I would get in the car and I'd hit it and I'd be like, "Hey, Johnny." And uh >> you wait, did you name it Johnny? >> I did name it Johnny.
Short Circuit was like my childhood robot movie. It was my favorite one ever. So I call it Johnny 5 because that was the name of the robot in the movie. >> He's asking real follow-up questions here. >> I have way questions now. Yeah. >> Well, it just felt weird being like, "Hey, chat GPT." Hey, you know, I was like, I got to have something to call it, so I gave it a name. >> Yeah. >> Gria, what about you? Same question. How are you using it in your day-to-day?
And what what particular models are you using? >> Uh great question. So right now I pretty much live in Claude, but I go to Gemini and Chatbt for different um specific niche use cases. So, uh, the reason I do that is because on, uh, Claude's, I guess, like pro-us user like web interface, uh, I have all these projects saved, and those projects are where I save all of my, uh, relevant information for all of the work tasks that I do in a given day. So, for the Neuron newsletter that we write, I have like all of my style guides saved and all of my writing preferences.
And so, anytime I go to that folder, I can basically just say like, "Hey, I'm working on a story on this topic. here's the information like let's write and it will automatically know exactly how to write it the way that I like and so obviously when you have that and that's primarily what I do most days that's where I go to most of the time um so then the question is okay so what other things do I need to do that um claude can't do as well as let's say openai or chatbt number one thing claude doesn't do well is web research because anthropic hasn't signed all of the licensing deals that openai has so openai can read a lot more websites than um than Claude can.
So anytime I have a search related task, I almost always go to ChatgBT. Then for image generation, I almost always go to Gemini and Google. And the reason for that is because they have this model called Nano Banana Pro that is really, really good at generating images in all kinds of different styles and editing images. It's not perfect because Corey and I were just testing it earlier today. if you want to tell that story, Corey. But like for example, I can now create headers in the style that um I like for the newsletter and I can we can create thumbnails in the style that we've already predefined.
So the way that I use that those tools is basically you set up a style that like you've created with your own human ingenuity or like our video team has created like a thumbnail look and aesthetic that we like and then we then train don't we don't train it on that but we have a project folder again on Gemini they're called gems and in there there's like you know five to seven examples saved so then anytime I can say hey make a thumbnail for this you know topic with this person's headsh shot boom it can create it exactly the way we want it to so it saves a ton of time doing it that which is where we stumbled earlier. >> Yeah, >> I'm sure you're asking which cat would go well with this headline today in our newsletter. >> Oh, we right now I'm just letting the cats roll. >> Basically, we just I upload the story and I say make a thumbnail for this and then it does it >> and it reads it and understands the context and and it's shockingly good.
It really is. And uh and and I would like to add that four months ago it wasn't. No, you know, >> four months ago, you could barely get uh legible words out of out of an AI image model and uh Nano Banana did it and then chat GPT's new image models came out and now they can both spell, they can both put out an infographic pretty reliably. There's a lot of stuff like that that has just in this last short window of time really improved significantly, especially around images and video.
What was really technical? Oh, no. Go ahead. >> Sorry. I wanted to know what the side story was, Corey, about the image generation that you were attempting to do. >> I have this photo of me that has been floating around at work for three years that I hate with a passion. It is absolut It's It's me at my lifetime largest and like everywhere I go, it follows me and it haunts me. And I I decided I'm going to end this image and and I went and I've got this new one and it's fine.
But I went and realized Grant's like, you should swap that in your signature on your email. So I was like, oh, great call. >> Cuz I know he hates the picture that's there currently, >> but it was facing the wrong way. And I was like, I need to flip this horizontally. And he said, go to Nano Banana. And I was like, I don't know. I'm going to I'll try it. And we went in and I shared my screen. I was like, cuz I'll use both of these. And and I I feel like chatgbt edits better, but that nano banana will generate a first image better sometimes if if that between the two. >> So I go into nano banana and I throw the image in there and I say flip it horizontally and it gives it back to me exactly like it was uh when I when I uploaded and I was like hey no.
And I told it another way and it came back I think the exact same the the second time too. So then we went over to chat GPT and the first time I entered it and said flip it horizontally, it turned me 90 degrees sideways and it's like there's this head. >> Yeah. And it was profile my hair was shorter. >> Frankly, I've never seen that part of my head. Finally, I went to a new chat and and I and it did it on the first try really easily. Um >> that's actually good advice for anyone who who needs to use AI for work or for for fun.
A lot of times the solution to your problem just start over. Start a new chat. Do what's called the best of end. >> Do what's called the best of end technique where you spin up like four chats and you've asked the same question four different times and then you just pick whichever one is the closest to what you're trying to do and then continue >> that conversation. Yeah. >> Yeah. >> It's because the model is still learning too. It's still trying to figure out who you are and what you want and everybody else in the world.
So, that is a really good suggestion, Grant, because I've totally asked it the same thing a number of times in the same chat and gotten the same response and then just the next day start up a new chat and it does something different. Um, really interesting there. Something that I think we've all kind of laughed about and touched on throughout this and Cory, you said it pretty explicitly when when we were talking about who you think is winning right now and you were like, "Well, what could change by dinner?" As these tools keep advancing, right, we're probably at a stage now where for at least for a while it's going to feel like this. >> Yes.
Do you think that the concept of, you know, people getting ahead of the curve or even businesses being ahead of the curve to some degree, is that kind of dead in the water now? I mean, is anybody ever, this feels very nihilistic and I don't mean it that way, but like are we just in a perpetual state now of somebody having no clue really what the current state of AI is? And and if that is the case or however you feel about that, where do people go then?
Right? cuz it's it's fun to sit here and talk about this, but there is also something in the back of my mind that's very like so I got to devote 10 hours a day for the rest of my life to figuring out what this technology is or else half of the world thinks that I'm some kind of lite because that doesn't seem a fun way to live for anybody. >> Well, I would say the first thing you do is you go subscribe to the neuron who we're speaking to. But second, I think we're going to see things changing.
But what I will say is, you know, lots of people are upskilling right now and and it's not as easy across the board as it kind of seemed like it was going to be in the beginning. Like early on there was this idea that we're going to create these great prompts. We'll give them to everybody on the team and it'll just work. And that has pretty much proven to just absolutely not be the case. So people are going to have to upskill and companies are going to depend on that. companies who haven't started finding their place in AI by now I worry about but frankly at the pace of change there's time to catch up I think as long as what you don't want is to be you know building AI models and in the lead of the race I would say that there are we're going to be in a perpetual state of here's what's new here's what's new here's what's new for a while and even if that stopped today where we're at already you know there's still years and years of development to go on with the models as they exist this afternoon.
Like it would slow down, but it would still continue if there were no new developments, I believe. >> Yeah, I'm going to add some stuff to that. Um, and slightly disagree, but agree in principle. So, that this is a great question. I I love this, by the way. the idea of being constantly behind. I think that will that will continue like Corey said and I think there's just nothing you can do about that because no matter what the AI labs themselves are going to be working with a model that's already better than you have access to today and they're going to be creating new techniques and new ways of working with it um for months in advance before they release it to us.
So, no matter what, there's going to be something new that is coming out that at least some people will have like secret access to. Um, but I don't think you need to spend 10 hours a day learning some alien technology for you to have the skills you need in this environment. I read this really great article from Alberto Romero where he basically broke down how you can get caught up to speed in a day, like in eight hours with AI with basically three techniques.
And um I'm going to butcher all of them, but the main one was just get in was just get in there. Start messing around with it. Start trying to actually write what you need to to to do and continuously iterate on the context that you give the model. So it's not enough to say like, hey, I need to write a business plan or or whatever it is that you need to do. you need to really get specific about what it is that you want it to help you do and give it all of the information possible in order to accomplish that task in order to then see if it's capable of really doing it.
Like you have to basically assume that the model can't read your mind and try to dump your mind into the machine. >> If a new employee started today and walked in the door, >> it needs all of the context you would give that new employee. And then if you once you have given it that you know that experience of of doing it you know do that like we said best event like four different times five different times and see what kind of things uh happen because it isn't a deterministic model it's probabilistic.
So that means that it can sometimes respond differently depending on how many other people are sending prompts at the same time as you are. Yeah. Um which is something that they realized recently and hopefully they've they've all fixed. But prior to now it was like, yeah, if you just if a lot of people were sending chats directly to ChattBT at one time, your model could be go completely off the rails. Like your answer could be way bad just because a lot of people were sending it at the same time.
So yeah, the the number one thing is like just get in there, try to give it as much information as possible and see what it can do. And then there's really only two or three tools that matter that you need to know right now. Obviously, everybody knows the chat interface, which you know, your mileage may vary there. The other one is the command line terminals which are actually just really good general purpose agents and those general purpose agents can kind of do anything on your computer.
So it'd be worth it to just mess around with like a claude code for example which is very userfriendly and you can literally talk to it and say hey Claude what can you do and just like actually try to get to know it >> and if you don't know where to start ask the chat be like hey I want to try this but I don't even know what a command line is. Let's start from there, you know, and >> because you just did that recently, right? >> Yeah. >> Yeah. >> That's one of the coolest things about the opportunity with AI is that if you don't know how to use AI, you can actually ask the AI, "How do I use you?" Obviously, it has to be better prompted than that basic question.
But even if you just type in that Yeah, it's a good point. That's a good point. I think those are great points that you made, Grant, too, about how uh citing that article about how you can get started with it. Um, it makes me wonder though, and I'm gonna ask the question that I was kind of sitting on for a minute there. When it comes to AI, do you think we're in a bubble right now? All of this is super exciting. It's been super exciting for the last two years, but now we do have folks asking, are we in a bubble?
Are we dependent on LLMs too much? Why or why not? And that question is posed to both of you. >> There is something that could happen that would slow down the development here. Actually, there's multiple things that could happen. Number one thing that could happen is like a bubble pop situation where a lot of people debated that last year. And I think the the reality is it's not a bubble yet, >> but things are overpriced, if that makes sense. And they have to figure out the economics of this before Anthropic and OpenAI go public. um because otherwise it's uh you know it's not sustainable for them to to continue at the pace that they're at now.
So the the one way that things could slow down would be if the bubble like has like a a mini pop, let's say. The other way that things could slow down is if there's any sort of government uh regulation or backlash at the moment. That doesn't seem like that's happening in the US, but we have an election coming up this year and that could change fundamentally change things. The third thing is it could just be that the model companies get exhausted or scared of what they're building and they they have a gentleman's agreement to like, hey, let's not re let's release these things like twice a year.
Like you know how right now iPhones, you get a new iPhone basically like once every fall or whatever. >> Yeah, the early September developer conference. Yeah. >> Yeah, it might go that way where they actually just slow down. I think if people it it depends because this whole thing is a race to AGI and like we've said like could be that we've already achieved it could be that it's way far away depending on your definition. Uh but once that is achieved I think there's going to be a lot more questions about how to make sure that it's safe h when and when not to release it.
Um you know and so they might actually slow things down to a cadence where they're like hey we don't want to release this because we need six months to test this before it goes out. I would say that yes, there's is definitely some element of a bubble here and simultaneously I'm not that worried about it. Um, and I say that because we've never had a technology with the potential to touch as many things as this does. Like this isn't just about technology.
This isn't just about how AI is going to impact business. I mean, we're talking about healthcare andarmacies and education and agriculture. We're talking about banking and politics and warfare and we're talking about infrastructure. This is going to touch everything in an increasing way as we move forward. In my opinion, I think what we're just now starting to see is people in individual industries really kind of learn, okay, there's a spot here where I can own this niche.
I can I can own what this looks like in design. I can kind of own what this looks like maybe in accounting or law. And we're going to see that continue to happen across everything from manufacturing to boating. I mean, it it's it's going to be there in all of it. And I think that the opportunity is so big that people will continue throwing money at it. And and I don't even believe that the companies currently involved could necessarily fail in the way we think of a company failing.
Like if OpenAI goes broke today, Chad GBT is not going to go away. There's going to be a bidding war where people are going to pay the most obscene amount of money for a defunct company they have ever paid in the history of our world. And and they're going to be immediately propped back up, you know. And Nvidia is an interesting player. Awful big. uh really really big, really really smart, made a lot of smart moves a decade ago, 12 years ago that that have led them to this position where they're at now.
And honestly, I feel like they've got years of headway in front of them still. When you look at the new architectures they've released and what that means and everything they're trying to do, I Yeah, bubble, no, I'm not really that worried, I think, is is really my takeaway. >> Yeah, I'll add I'll add just two more things to that. So number one is that the spending at the current level is like still not at the level of the telecom boom bubble adjusted for inflation.
So it's I think the statistic that I read most recently was like it's like 0.7% of GDP versus the telecom boom was 2% of GDP at its height. Um so we're far away from it being like that level of of intense. And then also a lot of people were saying if there is a bubble it's an infrastructure bubble and it's not a financial bubble at this point because the infrastructure is um you know what's overpriced at this point and it's not necessarily that the stocks have been inflated.
Now once the companies go public that's when you need to start paying attention to the financial bubble because who and their grandma is not going to buy open AI stock when it goes public. >> I have a few questions about people's grandmas knowing how to buy stock in general but yeah we >> they will figure it out. They will figure it out for they'll ask me. >> They'll ask GBP. >> They'll ask Chad GBP. How do I go buy stock today? >> If OpenAI is smart, they'll have like Robin Hood uh app in their store and that's like super easy.
It's like buy OpenAI stock here. >> Yeah. >> Actually, that's brilliant. Yeah. >> Yeah. They'll probably do that, >> right? Like the stock conversation gets so tricky because then you have the, you know, grandma probably doesn't know what OpenAI is, but grandma's Roth IRA investment firm probably does. >> Yeah. So, grandma's probably like, "I don't know what that means, but >> I don't know what I have, but but it's making money." >> I wouldn't I wouldn't rule out the grandmas, uh, Victoria, because there are some grandmas who are reading, uh, the neuron and they're learning. >> Listen, I'm sure they are, and I'm big fans of those grandmas.
But I do think it's interesting to hear you guys because I think a lot of this conversation just lacks any kind of nuance on the internet as many conversations on the internet do, right? So most people who run an AI podcast I think would be like no everything's great you're crazy if you have a question and I don't think people are crazy for asking where is this going >> I think we're looking at something significantly larger than the dot when we when we compare this like the closest thing we have to compare this to in my opinion would be like the general commercialization of the internet >> and there are so so many parallels between now and when consumer boomers and businesses started getting online.
Like, people were scared to death of the internet. They thought the internet was bad. It was it was evil. It was going to be this big problem. They didn't want computers in their darn business. They weren't going to have it. And now, can you imagine any business operating like, oh, we just we just don't use the internet, you know? I mean, it's ludicrous to even think it. And I really >> probably do. >> I'm sure there are. I'm sure there is. I'm sure there's a blacksmith with a hammer and a forge banging out knives all day.
It's like, I don't need the internet. But but for the >> word of mouth. >> Yeah. Port of mouth. Yeah. And uh I'm word of mouth. But I really think that we're looking at something of that magnitude or greater in terms of of what we'll see. And you know, there are obviously going to be companies that fail along the way. They're going to be investors that lose their tail. there going to be people who make investments that looked good and and made sense that just just don't come through.
But right now, the amount of money coming in and with little reason to think that's going to slow in the near future, I I think people are just going to continue grabbing up everything they can grab right now that says AI on the name. And, you know, hopefully it doesn't all fall at once, you know. N well uh V Victoria to your point about the the nuance that is the tone that we try to strike with the the neuron AI explained and the newsletter is we try to uh look with clear eyes at what is actually going to happen or what's likely to happen and what is actually the most helpful to you right now.
Um do we always do a perfect job at that? No. But that's what we're trying to that's the tone that we're trying to set is like hey we are the middle ground between like AGI is going to like save us all and take us into a realm of abundance um which may happen but the the caveat is achieving AGI and then the like doomer version which is like AI will kill us all and you know we you know there's no way we're going to be able to control it and it's going to take all the jobs.
Um there is there is legitimate concerns there with the way that it's rolled out, but I mean those are like society level concerns, right? They're not it's not just on the technology side. >> Yeah. And in fairness, the people building it did for, you know, the past three years scream at the top of their lungs like, "You guys need to be ready for this." And we were all like, "Oh, look at this crazy guys in tech." Oh, no. >> They're nuts. Look at that.
And and you know, it's it's an interesting time and there will definitely be interesting times ahead of us. But I think that that is something the neuron does really well that that I'm proud of and I know Grant's proud of is that ability to not just say here's what happened today, but to say here's what happened today and here's what that means. Here's why you care. That's that's what we're trying to do in the newsletter. That is is different like like you know if Grant's talking about claude code, it's not just here's a cool tool, check this out.
It's here's a cool tool and here's why you should try it. Go do this. It's not that hard. here's a way to do it that's really simple and clear and and Grant does a masterful job of explaining that kind of stuff in my opinion and u that's what we hope we can offer that's different is the fuller picture and hopefully in you know 10 15 minutes of your morning or an hour of your time on YouTube [snorts] >> I think realistically that's that's an excellent place to end it but just for the sake of pointing it out I really like what you actually touched on there, Grant, where you talked about we're having an election year.
We are going to have a lot more conversations around AI and just because you can, should you? We're looking at a lot of situations happening that have made mainstream media and made parents concerned, made business owners concerned, made women concerned. Look at everything that's happening with Grock right now, over sexualizing women's images. Look at environmentally where data centers are popping up everywhere and people don't like it for e x amount of reasons.
You also consider the safety concerns and the privacy concerns and we're having conversations for the first time in over a decade about whether or not children should be on the internet or on certain platforms because now they're being powered by artificial intelligence in some way. Australia just shut down Instagram for children under 16 and it is so important that we're having these conversations now. So, I do think another blockade to progress, and I I'm using air quotes when I say progress because I think it's very important to consider society and and morally where we're going with this, that it's not always a conversation about losing jobs, but losing a piece of ourselves within this, that we need to be responsible with AI.
We need to be responsible and good stewards of technology to each other, or else we're going to ruin it. We're going to ruin the possibilities of what it can be used for things like cancer research, for things like environmental research, for things like just using it for day-to-day like we're using it right now. That could easily go away if we go too far as we've seen so many times in the press already where people have used it and have either not been able to get the help that they need for their mental health, not been able to stop images being run rampid into into honestly digital porn.
Um, and it's scary that we can see it go in that direction. But again, to bring it back to a place of positivity, I think that's why it's so important that a podcast like the Neuron exists where you can talk about it not just in the context of I am so high above you and can talk about this really cool thing that is very complicated. It's you're bringing it to the people in a way that they can understand it and make their own decisions about it. But I would say I encourage everybody use AI responsibly.
That's on all of us to do for sure. Well guys, if we want to, you know, subscribe to this neuron uh newsletter that you've talked about or this this YouTube page that you've talked about, where can we go? What can we do? How do we get involved? >> So, you can go to uh the neurondaily.com and sign up there. Uh and on YouTube, we are the neuron AI. Check that. >> I think that's right. At the neuronai on YouTube and uh yeah, check it out. We're also everywhere else you might listen to a podcast.
Spotify, Apple, you know, take some time out, check it out. You know, we got a ton of content going out right now from shorts to long form to we're doing weekly lives that are that are a ton of fun. Uh, you know, so if you ever have questions, you want to try something, come show up at the live, just ask. It's very much sometimes just us playing with toys. >> Yeah. >> Well, Grant, Corey, Victoria, thank you so much to each of you for joining me today.
This has been such a fun discussion. And honestly, the only reason I'm ending it is because we actually have to hop over to your studio at the Neuron to talk about Channel Insider and all things channel. So, I'll meet you over there. But, thank you so much for your time today. And I'm so excited for everything the Neuron has going on. And I hope you all at home take a look and subscribe. Thanks so much, guys. >> Thanks everyone for having us. Appreciate it. >> Thank you so much to Corey and Grant and Victoria for joining me today.
And thank you for watching or listening. You can check out every episode of Channel Insider Partner POV on channelinsider.com or come watch us on YouTube at youtube.com/ channelinssider_news and trends. You can also listen to us as a podcast wherever you get your podcasts from. Don't forget to like, subscribe, and follow wherever possible so you never miss an episode. Come connect with channel insider and me on LinkedIn and X. Once again, I'm your host Katie Bavoso and I'll see you next time. >>
This transcript was generated automatically from the
video's captions and may contain errors.
