← Marketing Against The Grain
Loading summary
Transcript65 lines
- [00:02]
Kieran
On today's show, we are going to break down how you avoid the AI content slop. We're going to show you how really great AI content is made. We've got a leading expert, Darius Lam from NEXT joining us today. Let's get right into today's show.
- [00:31]
Kip
Okay. We are joined by Darius Lam, who is the CEO founder of next, an AI content marketing system. And he is going to take us through how he has figured out to turn AI into a great content creator, something I know every marketer and listener of the show truly cares about. Darius, welcome to marketing guest degree.
- [00:51]
Darius Lam
Thanks Kieran.
- [00:52]
Kip
I thought we could start with like we want to get into. You have a really cool tool to create AI ads for CPG brands and there's a bunch of things you're going to show our audience. I do think we should start with one topic we were discussing, Mike, because I think it would be interesting to our audience, which is should your content plan today be a content plan to make your content imperfect? And so we were discussing like, hey, we can no longer use EM dashes in our text based content because that is a notorious pattern that LLMs create. And I was talking about the fact that I am like a Grammarly power user because I cannot spell or have very good grammar, but actually maybe those things make me more human. And so I wonder if LLMs can cause us to kind of go back to creating imperfect content. I'd like, I wonder what your thoughts are as a person creating a content AI tool platform in this space.
- [01:47]
Darius Lam
Yeah, this is a really funny question because we have found that the best ways to reach our customers has been to go more direct, be more authentic and to use a lot of writing tools less. Right. So we might use these AI tools for ideation, we might use them to help us write drafts, but the actual end result has to be written by us. And as you mentioned, there's this really funny phenomenon where I used to love to use EM dashes. I still use it a lot in my journaling and stuff, but these days I replace those with a double dash or replace it with something else, some other delimiter. So it's just really funny how AI has kind of changed the way that we create content, not just on the ideation side, but also to make it a little worse in a lot of ways. Right.
- [02:38]
Kip
The problem with AI is it takes patterns that people have figured out that work and then just make them really easy replicated to the world. I was doing a funny post earlier on like that everyone's LinkedIn feeds more are going to be filled with ChatGPT5 high post. When you're on X, there's like this one that always works that I just, like, don't. I'm so sick of seeing. I'm sure Kip is on X, which is like, this guy literally shows you how to build an entire AI agent business. And then it's like a AI generated one pager and, you know, 99% of people know Excel, but they're all using it wrong. I just, like, you see these patterns all of the time. And I think at some point, just like ads, we're going to start to tune out patterns that AI is creating en masse. And so you're going to have to be forced to be more original than ever. And I agree. Like, I think one of the funny things is people are going to talk about, hey, AI is disrupting 90% of all professions. And I'm like, it also took my EM dashes. Don't forget about the EM dashes. That's a. You know, it's a tough loss.
- [03:34]
Kieran
Well, it's crazy. Maybe the biggest arbitrage, if you're like a counterculture marketer, if you're just willing to do crazy, like, you could go and just make a billboard with a bunch of typos about it not being AI and it would blow up the Internet.
- [03:48]
Kip
That's actually pretty grats. That's a good one.
- [03:50]
Kieran
Well, it'll definitely work. We should go do it, like tomorrow.
- [03:52]
Kip
You could do that.
- [03:53]
Darius Lam
This is the double problem with AI, Kip, right? Which is once you made something that works and goes viral now using AI, everybody else is going to copy it, right? And the cost of copying is basically.
- [04:05]
Kieran
Going to 0 a hamster on a hamster wheel who was like hepped up on amphetamines.
- [04:12]
Darius Lam
If you can capitalize on this ad idea right, and make it like your own, then obviously that's awesome for you as a brand because now you're NF1, right? Everybody's following you. I think that's gonna be cool.
- [04:24]
Kip
Yeah. Yeah. I love that actually, like, AI causes originality to be more important than ever. And I do love the brand. Like you could imagine in San Francisco. Kip, you have a billboard of like Hubbot, Hubsplot, the number one. Just have a complet Ron, not brought to you by AI. That actually would work really well. Okay. I do think we should get in Darius to maybe give us some context, like, why AI content platform? Like, why did you decide to do that? That's a tricky space and then maybe give us a little bit of the story of the tool itself. Like, what have you figured out? What does it do?
- [04:59]
Darius Lam
So, I mean, maybe this is a great time for me to do a quick introduction to myself. My brother and I, you know, we're co founders. We started NEX really to support CPG brands. We saw that, you know, these brands are producing the goods that we are using every day, right? Food to beverage to even, like consumer staples. But that every existing AI tool for these brands was failing them in two key ways. The first is that none of the content, when you talk about images and videos, was actually product or brand consistent. And that's like a huge deal. And, you know, I can talk a little bit about some examples and I'm sure, you know, the listeners of this show will, you know, resonate with that. And number two, that we believe that this isn't going to be solved with building like a wrapper around existing technologies, so to speak, right? Where there's a lot of folks in the space that are building these, like, agentic workflows and things like that. I think there's a ton of promise. There's. But the real solutions have to be fundamentally technical solutions. And so we spent a lot of time and effort, for example, to set up our own computing infrastructure overseas. We've collected our own data set, we've like built our own diffusion models at large scale to be able to help these brands. And I think the results, you know, you'll see will hopefully speak for themselves. And, you know, a big part of that is like the process that we take to create this content where you can't just go to a chat box and say, like, hey, create a lifestyle ad for my product. Right? Like, that's not going to cut it. There has to be a process. And our belief is that that process is, you know, a variant of this OODA loop, you know, from the military, right? Observe, orient, decide, act. And in the content space, it's ideate, create, distribute, optimize. So for us, we're trying to build this tool that can help you do this whole process end to end. But I think the most exciting part is the content piece and how we are able to get that correct most of the time. You know, I'm not saying that, you know, perfectly yet. Obviously there's a long way to go, but I think we're doing a little bit better.
- [07:15]
Kip
You said something important there, which is like, we had to build some of our own models. We built computer systems offshore. We're not just a wrapper like you're not just a interface wrapped around one of the models. Can you just talk a little bit about some of the hard things you had to build to get the product to be really good at the content you're going to show us?
- [07:34]
Darius Lam
Yeah, and you know, I think this is something that we could definitely talk about more at the end of the episode too. And you know, I'm excited to hear what you guys think too, where the technology is going. Right. Obviously we know things are moving really quickly. So, like, what is the future of content marketing, especially in the face of this new technology? But for us at nex, you know, earlier in the year we showed the world this new model that we created called Icon 2. It's an 8 billion parameter diffusion model, among the largest, largest in the space. It was able to outperform models like Nvidia's Sauna, Apple's Airflow, Deepseek's Janus Pro on the Gen Eval benchmark. We haven't really released it yet fully for public use due to some issues around, you know, generalized performance, but we're definitely continuing to work on next generation versions of the model. I think the biggest challenge, right, when you talk about like marketing and cpg, you get a lot of folks who are really great and knowledgeable about traditional marketing, whether that be like tv, whether that be social media, meta ads, Instagram ads, but sort of lack a lot of the technical sophistication to really build upon that. And I think what we are really trying to do is to fuse CPG expertise with novel innovations in AI so that we can bring these brands the best technology in the world and hopefully let them focus more on building great products, selling those products to their customers, and less time, you know, fiddling with Adobe Express and fiddling with like workflow tools to create content at scale and things like that. Basically let us do that work for you. Cool.
- [09:24]
Kip
You want to show us some of our audience some of the examples and then maybe we can riff on the ability of AI to do content.
- [09:31]
Darius Lam
Yeah, absolutely. All right, so maybe I can tell this story as like, okay, what can existing tools do at the start? Right? So as an example, we basically took this bottle here, which is a soap bottle, and we put it through a standard workflow that you might see and read about online, right? Which is like, how do we use ChatGPT to create this kind of ad content at scale? So this is an example of a result that we're able to get. We got, you know, many more like it. Here's like the other one that we got here, and you might think, at a high level, find your balance, okay, decent tagline. But when you actually look at the product itself, it's like so far away from the actual brand's product that it becomes very difficult to run this kind of thing.
- [10:20]
Kieran
Right.
- [10:20]
Darius Lam
Like, even something as simple as, like, the logo is wrong. Or here it says that this bottle is 141 ounces, which is like 8 pounds, right? So this is like 10 times, you know, larger it's saying, than it is. This is particularly a problem for cpg, right, because you can't falsely advertise your product. And it's not just that this is like, bad for, you know, the brand owners, it's actually really bad for their customers. It has negative brand value. So, like, this is pretty much the level of quality that you're gonna get if you are trying to sort of do this out of the box today. So we had put together, you know, a little workflow that allows us to get stuff that is a little bit higher in quality. So, you know, this is an example of an image that we were able to put together. I don't know if you can see this pop up here, but this is like 6,000 pixels by 11,000 pixels. So we are, you know, you could literally print this out and frame this a wall if you wanted to. We're able to get the logo correct, the name correct, of course, and even all the way down to like the individual small pieces of text here at the bottom. Like, it. This here says like plus probiotics or something like that. Right. And, you know, this level of fidelity is what brand owners expect of traditional photography. So why shouldn't, you know, AI tools be able to get there too? Here's an example of something that actually didn't work out. This could be interesting for us to continue to talk about here, right? Which is we had put this together for a client and they had come back to us and said, wait a minute, Darius, take a look at this set of reeds here in the center, where this background read is merging in with the foreground read. Wait a minute. From a consumer angle, I would have looked at this and said, okay, this is basically indistinguishable from a real photograph. Same. But that detail matters, right? And being able to get that, you know, with 80%, 90% accuracy is, I think, going to be the difference between folks who are, you know, just trying out this kind of content on like a novelty basis versus true usage of the photographs for things like social content. Or ads. So, you know, I think that this is something that is going to be the future. The quality of the content has to go up. The accuracy of the content has to go up. One last thing I'll add here is it's not just about the quality. It's about whether or not the image itself resonates with the brand. So in this image here, you can see that, like there are some sea urchins on the left and there are some fish on the right. Okay. That is obviously relevant to soy sauce, a brand selling soy sauce. Right. And if you were to just kind of do this out of the box, you know, these generators aren't able to do that. Right. And so that goes into the ideation phase. Are the background colors matching your brand? Colors? Right. Like, does the content align with what your brand is actually selling and to your customers and all of that. And then obviously, how do we produce this at scale?
- [13:39]
Kieran
Real quick, Darius is giving you his ultimate ad framework with over 20 prompts.
- [13:44]
Kip
So you can create the best ads.
- [13:45]
Kieran
At each stage in the ad creation process. If you want it, scan the QR code or click the link in the description. Now let's get back to the show.
- [13:54]
Kip
So you had a framework because I want to get into the. The actual ideation part. So I think you had a framework.
- [13:59]
Darius Lam
Yeah, it's ideate, create, distribute, optimize. Yeah, right. That's like that four step cycle that we go through.
- [14:08]
Kip
And so if you think about AI, even for what you're doing, like is the photo the hard part or is the idea the hard part? I think the idea is the hard part. And a lot of examples, AI's ideation and creativity is not better than a skilled human. Maybe talk to us about like the ideation part. Like you talked about it being applicable to the brand. Like in that picture there, is that really a creative ad? Does that stand out in a sea of noise? Does it stop people scrolling through ads? I think that's actually a really hard. Like you've built a model. Your pictures look amazing. I would love to go there. I was talking to Kip and just like print out some art and put it on my wall. Like rip off great artists. It's that level good, right? I can literally go and create art with your model and print it out and put on my wall. I probably would do that. But the ideation part is just so hard with AI because what it's basically looking at is it's training set and then coming up with like the average. There are ways we can Talk about that. You can get more in the prompt. You can definitely create a prompt that allows it to be much more creative and thoughtful. But, like, what are your thoughts on ED as a ideation creative partner?
- [15:13]
Darius Lam
I agree with you, Kieran, entirely. AI has a long way to go to replace human creativity when it comes to the idea. What we have actually done is there's a real person, a creative expert who's going in and creating these prompts and these images. So this isn't something that is like, you literally put a prompt in a box and it just pops out, right? Like, there's real ideation that goes behind here. A little bit about that process, right. Often what we do is we start with a mood board. So we go and we look on like Pinterest or Instagram or other ad spy tools, right? To go and, like, figure out, okay, this is what competitors in the space are building. This is what a good piece of, like, product content should look like. We take that and then we use that to, like, ideate on prompts. So we never do one image at a time. That one image that you see might be the result of 50 plus actual generations. To be able to, like, go through different variations, then you have to actually pick the one that you like and then edit it, right? So this is really getting down to like, you know, why is it so hard to do this out of the box today? And why there's still a role for, I think, you know, the human element. It's that if you were to try to just do one shot chatgpt, you're going to get something that's like, okay, but not quite there. You need to have that mood boarding. You need that ideation, you need to test the variations. You need to find one that works. You need to edit that image. You know, you need to tune the details, all that to be able to get this to work as one final result that you can then post or run as an ad.
- [16:57]
Kieran
What I think is interesting, I think there are a couple things interesting about what we're talking about. One, throughout the history of marketing, one of the core advantages have been, do you have a really good creative process run by very good creatives to get a great output? And what we're essentially saying is like, like that is just as if not more important than ever in a world of AI, there are a lot of people who maybe would have thought that that would have gotten less important. And what we're actually saying is that's more important. And the reason that's more important is because that's how you get like really deep creative density where the lighting is perfect, the subtle brand attributes are perfect. How it resonates with the target audience just comes through really clearly. You can't do that without a rigorous process and really good creative. I think what we're talking about is what AI allows us to do is to do that way faster and way cheaper. Because it's not like that Reed picture you showed a few minutes ago for folks watching on YouTube, for example. Right. Like before you would have had to set up a professional photo shoot. It would have been a month of planning. You would have hired a really great director and photographer and you would have just done this for, you know, a full day of taking pictures and then went through and reviewed and edited the best ones. Now you can do that without all those logistics, with all that lead time. And that's really where the opportunity lies, I think. Is that what you're saying?
- [18:16]
Darius Lam
A hundred percent. And you know, it's significantly more than just the money cost. Obviously that's a big deal, right? Like we're talking about going from 500 to $1,000 a photo down to dollars per photo, which is like just a massive collapse in cost. But I think the flip side of what you're saying, right, is the timing and the opportunity cost improvement here. We don't see that because the cost goes down, you know, that this like market is gonna shrink. Just the opposite, because the cost is going down. More brands than ever are going to be able to create higher quality content than ever at greater volume. You know, one thing that I'm sure you guys have seen is this proliferation of brands that run multiple social accounts. I think Duolingo does this super well with like Duo France and Duo Italy and stuff like that. Right? And we think that's the future, right? How can mid sized CPG brands run a dozen TikTok and Instagram accounts where each piece of content needs to come out, you know, three times a week, be targeted to that audience, be high quality on brand. That would have just been impossible, you know, a year ago. And we think that we're getting into a future where that's going to be really possible for these brands and everybody is going to be able to do it and give them better access to their audiences.
- [19:37]
Kieran
Question on that for both you, Darius, and for Kieran is like, is it going to be possible in a year or two to succeed in marketing, especially things like social, without just having 10 times more content than a brand is producing today? Like, can we talk about just the Cost of content coming down, the volume is likely going to go up. Is that a good thing? Is that a bad thing? Are brands going to get on that like big hamster wheel of just tons and tons of content?
- [20:06]
Kip
Isn't it only a good thing if you have the targeting options to get way more specific? Because a brand increasing volume, if you're having to promote that through the same channel, you don't have really good targeting options to like segment users into smaller audience so you can craft content specifically for them. Then is it not just like instead of producing three posts, I produce six posts? But does that really matter? I think it's actually the targeting that really matters. You can create bespoke content now for smaller groups of people. I don't know what you think, Darius. That's kind of how I think about it.
- [20:39]
Darius Lam
Targeting obviously matters a lot. Let me see if I could provide a bit more of a out there perspective on this, which is everything you said. Kieran. I agree with insofar as the media of today stays the same. But as I'm sure you've seen with like Google's GENIE and you know, new technology that's out there, what if the future of media is like virtual reality where everything is like generated or augmented reality where you're able to put pieces of content in front of people, you know, in widely different scenarios. In that case, the amount of content that a person actually consumes on a daily basis skyrockets and we are not yet there. Right when we went from like think about it, TV to mobile, right? And that's like from like me watching, you know, Cartoon network on the TV to now like scrolling TikTok, right? Like the density of content on TikTok has increased versus TV 10x and I think that we're not saturated. Future media could increase that density another 10x. In that case require technology like AI to be able to even keep up. Right. And that could be a very interesting future, you know, not far from now.
- [21:55]
Kip
Like from working within this space and technology. How far off do you think the models are to be a great creative partner in that? The litmus test for me is you would have a creative idea from AI that is better than what the human could generate. You know, we're recording this on the Thursday the 7th of August, where GPT5 is maybe coming out in 20 minutes.
- [22:22]
Kieran
It's not. Maybe it's coming out in 20 minutes.
- [22:24]
Kip
Yeah, it's coming in 20 minutes. And so from my experience it's not as far off as we think. If you know how to prompt. I think there's ways to prompt that you can actually force the model to be much more creative than the average person would be able to do via like standard prompts. But I'm curious, like you have a human in the loop now because they're needed to do a lot of the creativity. When do you think you do not potentially need that human in the loop?
- [22:48]
Darius Lam
You know I would have a hard time telling you a date. Right. But I think it's coming sooner rather than later. Yeah, I think the key solution is what they call in context learning. So obviously you know the problem with these models is that they tend to regress to the mean. So if you just use it out of the box, they give you very mean looking outputs. So the solution traditionally has been to then fine tune a model based off of your use case.
- [23:14]
Kip
Right.
- [23:14]
Darius Lam
I think that for a lot of creative tasks that set of outputs is actually quite narrow and can be determined if conditioned on sufficient past information. For example, if you are a brand that's been running for three years, you have three years worth of posts already, including data on posts that have done well versus not well. So the future is probably going to be you can put all of that past knowledge into the model and let the model then learn in an in context fashion from that data to be able to generate new content that is similar to your content in your style.
- [23:57]
Kip
But that's a problem I have with AI as it works today. And that's like one of the things I was getting at with creative assets. I think there's a couple of problems with creative assets versus other types of AI. Creative assets are so hard to distinguish why something works and why it doesn't work.
- [24:11]
Darius Lam
Yes.
- [24:12]
Kip
It's like I could put out a post and spend a minute on it and it gets like hundreds and hundreds of likes and then I put out something I really craft and spend time on. It gets like three likes. Right. It's kind of hard to like figure out why something works all of the time, especially in like visuals and creative tasks. Whereas like in coding you built the app or you didn't build the app right. Logic, science or these things are somewhat binary that you can kind of show the model the right or the wrong. Whereas in like a lot of these creative assets it's, it's a lot of like interpretation and taste. Whereas to your point. So the way you can get around that is like well I can train the model on the good and the bad. I can say we have three years worth of data, all of These creative assets and these ones work well and these ones did not work well. But the limitation of AI is it's just going to give you a bunch of things that are similar to the things that work well. Whereas what a human can do is think of something completely different. When does AI make that leap? Right. Like because of this success I make a completely tangential like leap somewhere else versus just like here's more of the same things but quite similar to what you've done in the past.
- [25:15]
Darius Lam
So we are working on this problem and I'd actually appreciate your guys thoughts on this to see whether you think this is the right approach. There is this technique called reinforcement learning, right. They use it to train models to be good at coding. Because coding is an easily verifiable domain where you can have the model generate code, run the code and see whether it's right or wrong.
- [25:38]
Kip
Exactly.
- [25:39]
Darius Lam
So what is the verifiable domain for creative? I would argue it is CPM conversion rate. Right? Like that is a measurable, you know, data point that is already being collected today that provides very high signal and that is non stationary, right. Which is that it adapts to the world around you. So the real holy grail here is imagine a model that is capable, that is basically now hands free, right? That you know, you might not even know why it's generating this piece of content, but maybe one day it generates like YETI content. Why? Because it finds that YETI content gets the lowest cost per mile. Because it's really hyped up on TikTok, you know, that week and then the next week there's another trend that's coming out and it just adapts.
- [26:34]
Kip
Right.
- [26:35]
Darius Lam
Today that would be a person sitting there scrolling their feeds, maybe having some monitoring tools to figure out what's hot. You know, we've worked on pieces of content for our clients that involve like some trendy things like you know, the glass ASMR on TikTok, the Yeti trend, you know, all kinds of these things. They get a lot of views, a lot of engagement, but the future is a system that can do that on its own. And the only way to do that is to have some kind of verifiable domain for creative assets. And we think that there is already a potential verifier out there. It's a global verifier in the form of, you know, the market.
- [27:14]
Kieran
Yeah, a couple things on that. I think if you use CPM or conversion rates to like do that, you're going to get what a pretty standard deviation from the mean, right. Like those conversion rates and those costs are normally pretty packed in, regardless if it's B2B, CPG, whatever. They're only packed in in a given market. Pretty close and tight together. And so you'll get like a standard deviation. That's great. You won't get the Sydney Sweeney American Eagle ad.
- [27:42]
Kip
This is what I'm getting at.
- [27:43]
Kieran
And this is what Kieran's pushing on.
- [27:44]
Kip
That's my point.
- [27:45]
Kieran
Right. You know, it occurs to me in this conversation we're having. I think it's helpful. It's like one of the negative things we've talked about with artificial intelligence over the last two years. And, Kieran, I think you have said it about a hundred times on the show is hallucinations and the reliability of AI to get consistent output. The reality is, when it comes to creative work, we want good hallucinations.
- [28:05]
Kip
Yes, we want hallucinations.
- [28:06]
Kieran
We want really far outlying things that are smart and based on human behavior. And so I guess the push I would make if I were dedicating some amount of my time to solve this problem, I don't think I would solve for metrics. I would try to solve for the cycle of human behavior, like the nostalgia loops, the things that are gonna come back in. Because all these things happen now. People are running to the opposite of these things. Those are the things that makes marketing magic, not the metrics. Which sounds funny as somebody who is obsessed with metrics his whole life, but I've learned that the other path is the better path. I don't know what you guys think.
- [28:46]
Darius Lam
Yeah, you know, I guess Omnicom and Publicis are still going to be in business in five years. Right. You know, they're the ones who can manage the Sydney Sweeney type of engagements. Obviously, there's a big part of that, which is that it's, like, culturally relevant. So there's a whole nother discussion to be had around how do you, like, put likenesses and people in these images? Yeah, right. You know, we have tried to avoid putting people in our content so far. We think that there's a lot of opportunity here around, like, you know, how do we make sure that every person in the image is, like, based on a real person and has their sign off, you know, otherwise you get a very dystopian world very quickly where it's like these random AI generated people who have no connection to reality, you know, trying to sell you something. Right. So I think that you're exactly correct, Kip, where there has to be some tie in to, like, like culture. And that is probably going to be done through a system that can assist the human, but not replaced entirely by the AI.
- [29:50]
Kip
Yeah, I think maybe to tie this episode up, it comes all the way back to where we started, which is originality. And AI, for the most part, is taking a bunch of these things and predicting what the next thing should be because of that data. Whereas I think what humans can do is do the unpredictable, you know, and I think the best. And I'm not saying the average company does not need to do this to, like your point. The average company actually does need you to plug in a bunch of data and then for their audience, like, figure out good advertisements that will work. But I think at some point, I wonder when AI can do the unpredictable, like create an ad that stands out because it's so different from what exists. It basically starts in your timeline. Right. You have a successful ad and then you have maybe two years of variations of that ad that's like, wow, that thing worked. And everyone tries to create different variations of it for their work. And then someone comes out with something that creates a whole new timeline of ads because that ad is so original and unpredictable from what came in the past. And that to me is like true originality and creativity. And I think until it's not reliant on the existing training set and can actually, like, think about things that are unpredictable and potentially net new, that creativity and originality part is going to be.
- [31:02]
Darius Lam
Hard to, like, figure out very much. So. The counterpoint I would say here, though, is that the vast majority of content is not that kind of highly original content.
- [31:13]
Kip
Right.
- [31:14]
Darius Lam
Maybe 5%. The top 5%.
- [31:16]
Kip
I agree.
- [31:17]
Darius Lam
And so I do think that there's going to be this split where on the one side, the highly creative trail breaking content, we absolutely have a place for the poets and plays, so to speak. And this technology will also be used very heavily in the future.
- [31:33]
Kip
Oh, no, I totally agree. I think originality is a very small part of what exists in the content space for a reason. Because it's really hard to be original. And I think you can do a lot of good and drive a lot of results by just doing better.
- [31:48]
Kieran
I think that's totally true.
- [31:49]
Kip
Like better versions of what exists and variations of what exists. Yeah. Awesome. This was a really good, really cool show. Great show, Darius. I think for people who want to check out your company and brand, do you want to just give them where to go and follow you? Follow the company?
- [32:05]
Darius Lam
Yeah, I'm on LinkedIn. You can just look me up. Darius Lamb. On LinkedIn. We're also@nexcpg.com Cool. Sa.