Latest / AI for Founders with Ryan Estes / Enterprise AI for $20 a Month

Transcript

Enterprise AI for $20 a Month · AI for Founders with Ryan Estes

• 0:00
  Most AI search is just Google with a concussion.
• 0:02
  It looks confident, it talks fast and still hallucinates like
• 0:06
  it's paid per lie. Founders are betting their
• 0:08
  entire companies on tools that can't tell the difference
• 0:11
  between a fact and a vibe. And the bigger your document
• 0:13
  pile gets, the worse it gets. You wait 20 seconds, open 6 more
• 0:17
  tabs, and forget why you asked the question in the 1st place.
• 0:20
  That's not productivity, that's digital quicksand.
• 0:23
  Now, if your company's brain is built on untrusted memory, how
• 0:26
  many smart decisions are you making?
• 0:28
  They're actually just beautifully formatted mistakes.
• 0:31
  Welcome to AI for founders. I'm Ryan Estes.
• 0:34
  Today you'll learn how everything search is getting
• 0:36
  faster, cheaper, and way more accurate by shrinking massive
• 0:39
  knowledge bases and killing hallucinations.
• 0:42
  This is for founders, CTOS, and operators drowning in SOP,
• 0:46
  strategy docs, customer notes, researcher sensor data, and
• 0:50
  wondering why AI still feels unreliable.
• 0:53
  And here's the wild part. Imagine taking a giant library
• 0:56
  of data, reducing it up to 99 point. 5.
• 0:59
  Percent while making searches up to four time faster and
• 1:02
  improving accuracy to the 99th percentile.
• 1:05
  Think 260 gigabytes turn into 1.3.
• 1:09
  That means small businesses can finally afford the kind of
• 1:12
  enterprise rag that used to cost a fortune.
• 1:14
  And it opens a new founder play. Take a weird, valuable data set
• 1:18
  nobody else has, compress it, make it searchable, and then
• 1:21
  sell predictions and insights as a micro SAS.
• 1:24
  If you want more of this, sign up for the newsletter at AI for
• 1:27
  founders.co. It's like a gym for your founder
• 1:29
  brain minus the mirror selfies 3 minutes a week that can save you
• 1:33
  3 months of building the wrong thing.
• 1:35
  And if this hit, leave a review on the podcast with one sentence
• 1:38
  that would either make your Co founder laugh or your competitor
• 1:41
  nervous. And quick break for our.
• 1:43
  Sponsorkickcaster.com. This is the premier podcast
• 1:46
  booking agency for founders who want customers to find them, not
• 1:50
  the other way around. You show up with your story,
• 1:52
  Kickcaster gets you booked on the right shows, and those
• 1:55
  episodes turn into a compounding content engine across clips,
• 1:58
  posts and search. Go to kickcaster.com to get you
• 2:01
  booked. That's Kid Caster by MO Burst
• 2:11
  Ankit, welcome to the show. Thank you so much for joining
• 2:13
  me. Absolutely.
• 2:14
  Thank you, Ryan. It's my pleasure. 100% really
• 2:17
  excited to talk about your product and talk about your
• 2:19
  experience building it. We're talking about Morphos dot
• 2:23
  AII believe the product is called Key.
• 2:25
  Is that right, Ki? Correct.
• 2:27
  Correct. Key stands for Energy in
• 2:29
  Japanese. It is a search enterprise
• 2:32
  application that we build for your average consumer and for
• 2:35
  your small business. So it is a a rag application if
• 2:39
  you will that utilizes LLM technologies with our
• 2:42
  proprietary vectorization technology to be able to
• 2:45
  outperform some of the the bigger companies and and just
• 2:48
  raw LLMS when it comes to data inference.
• 2:51
  Incredible. So let's unpack that a little
• 2:53
  bit above the fold. It says all your work in one
• 2:56
  place introducing the everything search.
• 2:59
  So give us a couple use cases and and couple integrations that
• 3:02
  kind of paints the picture a little bit.
• 3:04
  Yeah, absolutely. So if, if you've ever used
• 3:07
  notebook LM or chat TPT or Gemini, there's a, there's a,
• 3:10
  you know, pretty good chance that you're uploading documents
• 3:13
  or some type of information for you to be able to reference that
• 3:15
  later on, right? You're seeing a lot of companies
• 3:18
  are uploading SOP documents, strategy documents into these
• 3:21
  large language models to be able to reference, you know, as part
• 3:25
  of their automation stacks or just to make it easier for
• 3:27
  people to access that information.
• 3:30
  And so, you know, this technology has been around for,
• 3:32
  for quite a bit with enterprise rag.
• 3:34
  The only problem with it is that it's extremely expensive and
• 3:37
  highly inefficient. And so it's priced out all of
• 3:40
  the small businesses or even individual consumers that would
• 3:42
  like to access the technology. And so when you see tools like
• 3:46
  Notebook LM come out, it's very reminiscent to what it is that
• 3:48
  we do. We're just focusing on the
• 3:50
  performance. Amazing.
• 3:52
  Talk a little bit about the performance.
• 3:54
  Yeah, absolutely. So you know, I can, I can talk a
• 3:57
  little bit about some of our case studies.
• 3:58
  And so, you know, our, our big claim is that we're able to
• 4:01
  reduce the size of a vector database by up to 99.5%.
• 4:06
  And so, you know, if you have experience working with RAG
• 4:09
  pipelines or vectorization as a whole, you can understand just
• 4:12
  how big that that number is in terms of reduction.
• 4:15
  And so, you know, we were able to put in 260 gigabytes of the
• 4:19
  Project Gutenberg or sorry, the, the Gutenberg library, I should
• 4:22
  say, as part of our Project Gutenberg case study, 260
• 4:25
  gigabytes got reduced down to 1.3.
• 4:28
  And so you know, that's over 15,000,000 vectors getting
• 4:31
  reduced down to 76,000. And So what did that mean?
• 4:35
  It means you have the 99.5% decrease in your storage and
• 4:38
  therefore your, your storage cost.
• 4:40
  You got four time faster queries and you get a, don't call me on
• 4:46
  this just below 60%. I believe it's 5859% improved
• 4:50
  search quality up until the 99th percentile.
• 4:53
  And so that is extremely, extremely powerful because a big
• 4:58
  issue with these LM technologies that they hallucinate a lot and
• 5:01
  they take a long time to process queries when you're asking it to
• 5:05
  search a very large database, right.
• 5:08
  If I give it, you know, 500 gigabytes worth of documents,
• 5:10
  which is not uncommon in in medium sized or even large scale
• 5:13
  enterprise for specific things or specific departments, you
• 5:17
  know, a query can take 1520 thirty seconds to process and
• 5:20
  the information is still, you know, inaccurate, right,
• 5:23
  Couldn't hallucinate it. So that was the big issue that
• 5:25
  we wanted to solve. And we've been able to do that
• 5:27
  and make it affordable for, you know, small businesses.
• 5:29
  Your average user for the cost of Chachi PD Pro at $20.00 a
• 5:32
  month. Incredible.
• 5:34
  Yeah. If you're, if you're waiting
• 5:35
  2030 seconds, you're going to have six more tabs open.
• 5:38
  You're going to be looking at a YouTube video, what you were
• 5:40
  doing. Exactly, exactly, exactly.
• 5:43
  So that, you know, that's a big bottleneck because now we're
• 5:45
  looking at, you know, AI adoption is absolutely
• 5:49
  skyrocketing. Everybody wants to know how to
• 5:50
  utilize it. And there's a McKinsey report
• 5:53
  that came out and I forget the number now, but you know,
• 5:55
  there's a lot of distrust when it comes to information
• 5:58
  retrieval by these LLMS, right? People don't always trust the
• 6:01
  information that they're getting.
• 6:02
  And, you know, people are still dealing with this on the on the
• 6:04
  day-to-day. I was using, you know, Google
• 6:06
  Gemini earlier and I was throwing in some trick ball
• 6:08
  questions and it hallucinated and gave me wrong information.
• 6:11
  That's why LLM technologies have not been able to integrate very
• 6:14
  well into specific industries like medical or, you know,
• 6:18
  militatech applications, because in these environments, you can't
• 6:21
  afford to be wrong even once. And so when you have
• 6:23
  hallucination rates like that and you're dealing with
• 6:25
  lossiness, that's just something that can't be done.
• 6:28
  And so that's why LMS and, you know, these AI companies are
• 6:30
  struggling to, to break into it. And you know, we, we came in and
• 6:33
  and we solved that problem. Love it.
• 6:36
  And kind of this, this is kind of my favorite instance in, in
• 6:40
  creating business models right now as you're taking
• 6:44
  unstructured data sets, let's just say paper people have
• 6:48
  houses of paper and data that, you know, in the past, maybe
• 6:51
  it's cost prohibitive to like vectorize that, turn them into
• 6:54
  PDFs, put bring them in filter bright the software around it,
• 6:57
  boom, boom, boom. But now you have LLMS that can
• 7:00
  do that heavy lifting for you. So it sounds like, you know, if
• 7:03
  I've got a tough shed full of like paper and I spend the time
• 7:07
  to vectorize it, what I want to do is bring it into key and I
• 7:11
  that way it's going to like, it's going to shrink the data,
• 7:13
  get my response way up. And now I have kind of a a data
• 7:17
  set, a knowledge base that I can reference exactly faster and
• 7:21
  easier. Correct exactly.
• 7:23
  And so, you know, speaking on on that, the way that the current
• 7:28
  that current right pipeline has worked is that every time you
• 7:29
  want to upload new data and you want to add more vectors into a
• 7:32
  vector database, you have to re vectorize the entire data set.
• 7:35
  And so these companies are dealing with downtime when it
• 7:38
  comes to updating their their vector databases on the back end
• 7:41
  on a nightly basis, right? So that's a dedicated dev team
• 7:44
  purely focused on that with our core technology, which is called
• 7:48
  green vectors, ingestion is seamless.
• 7:50
  So that means your vector database has no downtime.
• 7:53
  And so we realized that this is so important and we wanted to
• 7:56
  build a consumer app that we're currently in beta for, which is
• 7:58
  key. But we wanted to really give
• 8:01
  this to a lot of the AI companies, vector database
• 8:04
  companies and ragged as a service companies.
• 8:06
  And so that's where we actually built our enterprise platform
• 8:09
  called Cortana. And so that is an API and SDK
• 8:13
  combination. It's a hybrid model.
• 8:14
  So, you know, waiting and fasting is done on your end.
• 8:16
  Your proprietary data does not need to leave the, you know,
• 8:19
  company premise or premises, I should say.
• 8:21
  Nothing gets sent to us. And you're able to use our API
• 8:24
  tax as our core algorithm and you just plug and play it right
• 8:26
  into your rag pipeline. So even if you have an AI
• 8:28
  application of your own that you know you're utilizing LMS and a
• 8:32
  couple other tools to deliver some type of product or user
• 8:35
  experience, you can seamlessly plug in our solution and you're
• 8:38
  going to get instant performance benefits.
• 8:41
  This episode is brought to you by a dangerous idea.
• 8:44
  It's called reading, specifically the AI for Founders
• 8:47
  newsletter. 47,000 Founders strong now, and it's the good
• 8:51
  stuff. No top ten AI tools that will
• 8:53
  change your life written by someone who doesn't ship
• 8:55
  product. It's just real breakdowns of
• 8:58
  what's actually working in AI, who's raised, who failed, who
• 9:02
  pivoted, and which founders quietly built $1,000,000 in ARR
• 9:05
  while everyone else was arguing on.
• 9:07
  X. It's written for people building
• 9:10
  with AI. Every week you get tactical
• 9:12
  frameworks, unfair advantages, breakdowns of real companies,
• 9:15
  and the occasional existential question about whether your
• 9:18
  startup is intelligent or just well, branded.
• 9:20
  It's free. It's sharp.
• 9:21
  It might mildly offend your current workflow.
• 9:24
  Go to aiforfounders.co and subscribe before your competitor
• 9:27
  does, because the only thing worse than being replaced by AI
• 9:31
  is being replaced by a founder who reads.
• 9:33
  That's the newsletter at AI for founders.co, now back.
• 9:36
  To the show. Amazing.
• 9:38
  Cool. And then the interface itself is
• 9:39
  chat based, Is that correct? Correct, correct.
• 9:42
  Yeah. The interface for key is chat
• 9:44
  based, very similar to, you know, a ChatGPT or Google Gemini
• 9:47
  looks very similar to that. We are on beta and you know, we
• 9:50
  wanted just to get it to a beta stage to be able to kind of demo
• 9:54
  the product. And we've been focusing
• 9:56
  primarily on the enterprise solution because that's where a
• 9:59
  lot of our customers are, you know, really wanting it to be
• 10:02
  integrated into. You do have applications like
• 10:04
  Notebook LM and that they have a wide range of features.
• 10:08
  They're extremely feature rich. We wanted to focus on the, you
• 10:11
  know, the infrastructure layer, so to speak, that technology is
• 10:14
  now built on top of. And so we want to become the
• 10:16
  status quo, I should say, when it comes to RAG and you know,
• 10:20
  RAG applications and now even going forward, other
• 10:22
  vectorization applications as well.
• 10:25
  Let's talk about examples of some of those to paint the
• 10:27
  picture. So this is going to be kind of
• 10:29
  the infrastructure for RAG application.
• 10:30
  So what what does that mean exactly?
• 10:32
  Is that basically so that you have kind of a siloed data
• 10:36
  sector knowledge base that you can you can enter interweave
• 10:39
  into other kind of sectors of the business?
• 10:42
  Yeah. So right now we're actually
• 10:44
  working with a militatech agency that primarily focuses on drone
• 10:48
  technology. And so, you know, you see this
• 10:51
  across the board, the drones, aerial defense systems,
• 10:54
  practically any hardware systems that that's deployed in, in any
• 10:57
  type of environment, it's used a lot of utilizing, I should say,
• 11:01
  a lot of sensors. Sensor data itself is, is very
• 11:04
  heavy and it's, it's very loaded, you know, especially if
• 11:07
  you're talking about like Lidar sensors, thermal imaging
• 11:09
  processing, you name it, it's a lot of information.
• 11:12
  And So what sensors currently do in a lot of these applications
• 11:15
  that it's constantly streaming information.
• 11:17
  And, you know, if I wanted to, let's just say, put an NVIDIA
• 11:20
  Jetson on a drone and run a machine learning model on the
• 11:23
  drone itself, you know, you're constantly streaming
• 11:25
  information. The, and the, the GPU is getting
• 11:27
  utilized constantly. And there's a lot of
• 11:29
  computational, there's a lot of computational cost that goes
• 11:32
  with that. And So what we're able to do,
• 11:35
  and you know, we're, we're just going to be publishing a white
• 11:37
  paper about this very, very soon.
• 11:38
  And we're testing this is by utilizing our technology, we're
• 11:41
  able to reduce the noise, right? That's the big claim of
• 11:43
  greenbacks. We're able to reduce up to 99.5%
• 11:47
  of the noise. And so when you're talking about
• 11:48
  sensor data, if you're able to strip that down, you know, with
• 11:52
  all the noise, you know, your images, the colours, borders,
• 11:55
  you name it. Just to be able to get the
• 11:56
  information that you want. Like for example, you know, the,
• 11:59
  the door was opened, a human crossed the road, there's, we
• 12:03
  spotted a car over here. This is the, you know, another
• 12:06
  car in front of us so we don't crash.
• 12:08
  These are the type of applications you want to really
• 12:09
  focusing on by bringing a vector database and being able to bring
• 12:12
  this vector edition talking technology to the edge.
• 12:18
  Cool. So it's, it's really like, you
• 12:21
  know. Because it's a chat.
• 12:22
  You're giving it an agent persona and with the data you're
• 12:26
  giving it personality and skills.
• 12:28
  So much like you when you you Neo taps into the matrix and
• 12:32
  he's like, oh, I know Kung Fu. Like that's basically what
• 12:35
  you're doing with your organization is you're like, OK,
• 12:37
  we just took all this information on Kung Fu we just
• 12:39
  uploaded in. So now we know Kung.
• 12:41
  Fu exactly. And so, you know, it's there's a
• 12:44
  wide range of applications that green vectors can be embedded
• 12:49
  into. And so, you know, when it comes
• 12:51
  to the the child section, that is a rag based application,
• 12:54
  right, very similar to ChatGPT Gemini right now, that's a rag
• 12:58
  based application for data inference.
• 13:00
  What we're also exploring, like I mentioned with the drone
• 13:02
  technology, is it doesn't have to be an SLM or a large language
• 13:05
  model. It can be any type of it can be
• 13:07
  a computer vision model. It can be any type of machine
• 13:09
  learning model that relies on sensor inputs to calculate
• 13:12
  something and then spit out an output that we want to be able
• 13:15
  to impact by being able to reduce the noise on the inputs
• 13:18
  to make the model run more efficiently, to make it run on
• 13:21
  cheaper and not as bulky hardware and be able to run it
• 13:25
  faster. Cool.
• 13:27
  Let's take it into science fiction terrain.
• 13:29
  Is that cool? Absolutely.
• 13:33
  So you're talking about Lidar. Let's say we we scanned the
• 13:35
  entire Giza Peninsula with LIDAR and we can find certain, let's
• 13:42
  say underground chambers. We can unlock different things.
• 13:45
  We see where pyramid footprints, we see new stuff, you know the
• 13:49
  application. Then as we bring it into, we
• 13:51
  bring that in to information into Key, which makes it
• 13:55
  lightweight, which gives it a personality, which gives it
• 13:57
  skills. We can upload, you know, all of
• 14:00
  the history of Egyptology. And then we could chat with Key
• 14:04
  to basically give us maybe some assumptions of what the culture
• 14:08
  looked like, how old it was, this kind of thing.
• 14:10
  Because basically it's learning all the information we're
• 14:13
  providing it. Yeah, absolutely, you can
• 14:15
  definitely do something like that.
• 14:16
  And so, you know, if you're talking about lighter, lighter
• 14:18
  sensors, if if you've worked with them, it's a lot of
• 14:20
  information, you know, especially if you're deploying
• 14:24
  it on something like an autonomous system, a drone,
• 14:26
  something to be able to capture actual imagery at at scale.
• 14:30
  You know, if you're doing it with a skyscraper or like a
• 14:32
  pyramid or whatever the case is, you definitely need to put it on
• 14:34
  a drone unless you plan on climbing all the way up and
• 14:36
  doing it yourself. And So what typically happens is
• 14:39
  when you, when you store Lidar sensor inputs, you have to
• 14:43
  process it, which means you have to bring a drone back, you have
• 14:45
  to bring it back to a command center, whatever the case is,
• 14:47
  process that data locally on a, on a pretty expensive machine to
• 14:50
  be able to process as our goal. And what we're planning on
• 14:52
  prototyping is to be able to remove that need where you can
• 14:56
  actually process information on device via and you know, NVIDIA
• 14:59
  jets and Oren or something like that.
• 15:02
  And then being able to upload that information into a tool
• 15:04
  like key and giving it more reference material and then
• 15:08
  asking it questions in natural language to be able to get
• 15:11
  better and better insights. That's definitely something that
• 15:13
  you can you can do to be able to speed up the process as well as
• 15:16
  to access information you normally wouldn't be able to
• 15:17
  access. Amazing and consumer application
• 15:21
  might be this is kind of your this is your second memory you
• 15:25
  know so it's pulling from all of available metrics.
• 15:30
  It could pull from the audio from your phone, it could pull
• 15:33
  biofeedback from your watch. It could pull, it could pull
• 15:37
  environmental, all that stuff in real time and basically provide
• 15:40
  you with like a photographic memory of.
• 15:43
  All exactly. Yeah.
• 15:45
  We're completely agnostic, completely agnostic for for all
• 15:48
  data inputs. You know, the way that works is
• 15:50
  you know, all data types can be vectorized, right?
• 15:53
  They're all numerically represented in in some way.
• 15:55
  And so we, we're completely agnostic with our core, you
• 15:59
  know, core technology. And so that's something we want
• 16:01
  to do in being, you know, rag applications were just the, the
• 16:04
  first step. It's the, the first thing that
• 16:05
  we decided to do because it was the the simplest and the easiest
• 16:09
  to, to build the beta off of, as well as to demo the, the
• 16:12
  technology that's core and its efficiency.
• 16:13
  You know, you can do a side by side of key and ChatGPT and just
• 16:17
  look at the accuracy as well as the speed in comparison.
• 16:20
  You'll, you'll be mine, you know, completely blown away.
• 16:23
  And now with the, you know, 2026, our big, big goal was now
• 16:27
  expanding that and want to see how do we apply this to, you
• 16:30
  know, non language models? How do we apply it to edge
• 16:33
  technology? And you know, that's something
• 16:35
  that we're building right now is going to be a handheld device.
• 16:38
  We're planning on calling it the Keyboy, which is, you know, a
• 16:41
  rag application on a edge device that can be built on top of a
• 16:45
  Raspberry Pi 5. And it's completely handheld.
• 16:48
  It's completely off the grid, decentralized, and you're going
• 16:51
  to be able to reference, you know, vast amounts of data with
• 16:55
  extreme precise accuracy and speed.
• 16:58
  And the fact that it's off the grid means it's kind of it's not
• 17:01
  you can't tamper with it or what's the benefits there?
• 17:04
  Benefit there is that you don't need a Wi-Fi connection to make
• 17:06
  it work right. It can be completely
• 17:09
  disconnected. You don't need server side
• 17:11
  technology to be able to run this model in conjunction with a
• 17:15
  vector database, right? That currently isn't possible.
• 17:18
  Nobody's even thinking about putting a vector database on an
• 17:20
  edge device. It's just something that's we
• 17:23
  thought about it, it hasn't worked, so we're not doing
• 17:25
  anymore. And so now you're seeing the
• 17:27
  need for data centers, you're seeing the need for expensive
• 17:30
  hardware. That's why, you know, the top
• 17:31
  companies like NVIDIA are pouring billions and billions
• 17:33
  into R&D to build better and better hardware to be able to
• 17:37
  process all this information. So by being able to reduce the
• 17:39
  noise, we're able to reduce the, the hardware that's required to
• 17:44
  be able to run some of these models and to do, you know,
• 17:47
  inference search or whatever, whatever you're trying to
• 17:49
  utilize it for. And we're able to empower every
• 17:52
  single person that has an iPhone or an Android to be able to do
• 17:55
  this completely locally. So that way you know, you're not
• 17:57
  paying for those expensive, you know, API costs and you're not
• 18:00
  requiring a, a good strong Internet connection if you're,
• 18:03
  you know, far away from, you know, a city or you don't have
• 18:06
  cell service. Again, if we're in Egypt looking
• 18:09
  for, you know, pyramids, then you don't, you're not going to
• 18:12
  have that. You don't want to carry around
• 18:13
  exactly, You know what I mean? Exactly, exactly.
• 18:16
  So, you know, an example, an example I gave is that, you
• 18:19
  know, if you're a doctor, like a, if you're a dermatologist or
• 18:22
  what not, you can have a pretty efficient computer vision model
• 18:27
  that's, you know, doesn't require that many, I would say,
• 18:29
  parameters that you can store on a, on a, on a, on an edge device
• 18:33
  or some sort. It might have to be a little bit
• 18:35
  bigger and then give it a vector database for for reference and
• 18:38
  be able to interact with it. So you can actually diagnose
• 18:41
  people on the go with this portable device, right?
• 18:44
  You don't need to utilize such expensive technology.
• 18:47
  People don't need to go to hospitals.
• 18:48
  You can go to them. It's very empowering when you
• 18:51
  start looking at being able to put that data on people's phones
• 18:54
  and be able to utilize that where everything can be done
• 18:56
  locally. Especially in a world now where
• 18:57
  decentralization is becoming the trend, people want to be able to
• 19:00
  tap out of the grid because we've seen when, when these
• 19:03
  legacy systems like, you know, AWS or cloud fair go down and
• 19:07
  then all the connected systems, you know, get shut down with it,
• 19:11
  right? You know, this happened not too
• 19:12
  long ago, a couple months ago, right?
• 19:14
  AWS was down for, you know, half a day a day, whatever the case
• 19:17
  is. And you're seeing essential
• 19:18
  services go down with it as well.
• 19:19
  So that's, you know, that's a, that's a big problem when we
• 19:22
  start relying on these large corporations to provide this
• 19:26
  infrastructure for us. We have the hardware
• 19:29
  functionality to be able to run this locally, so why not?
• 19:32
  Yeah. Well, and also it kind of path
• 19:34
  provides a path back to ownership where exactly we've
• 19:37
  been renting everything with SAS forever, which has its benefits
• 19:40
  for sure, but also has downsides.
• 19:42
  You know, do you really, do you really own your Tesla or does
• 19:45
  Elon own all of them? Correct, Correct.
• 19:48
  Yeah, yeah, yeah, exactly. So you know, a lot of people
• 19:51
  have been really aligning with the idea of owning their data,
• 19:56
  right? Not wanting to give it to a
• 19:57
  third party provider to store it.
• 19:59
  They want to own their data, they want to have it on the
• 20:01
  device. And you, you know, any
• 20:03
  application ends up running faster when you just think about
• 20:05
  it, if it's running locally on your machine, opposed to needing
• 20:08
  to make an API call and then all the server side infrastructure
• 20:11
  that's required it, it's faster and it performs better.
• 20:13
  So it's like, why don't we do it?
• 20:14
  We get data sovereignty and independence and we get better
• 20:17
  performance and cheaper costs. And ultimately, I mean, you
• 20:20
  know, flash forward 12 months, 2436, we're going to have little
• 20:24
  robots, you know, and we're, I, I think that's going to the
• 20:28
  privacy of robots considering like how intimate people are
• 20:31
  getting with LLMS already once The thing is in your lap
• 20:35
  stroking your hair, I'm assuming like we're going to be a lot
• 20:39
  more honest with those. You know.
• 20:42
  Yeah. Absolutely, Yeah, yeah.
• 20:43
  So that that's, you know, funny you brought robotics, autonomous
• 20:46
  systems and robotics is something that we're really
• 20:48
  looking into this year that we want to integrate with.
• 20:51
  We plan on being able to integrate with, you know, the
• 20:54
  entire NVIDIA lineup and, you know, putting our core
• 20:56
  technology directly on it. So you know, a lot of the
• 20:59
  autonomous systems, for example, they're they're using, you know,
• 21:01
  the The Jets and Thor I believe is it is for like humanoid
• 21:04
  robotics and whatnot. So being able to put our
• 21:06
  technology on that, we would have a massive impact in that
• 21:08
  space as well. Yeah.
• 21:10
  What's the special sauce? What?
• 21:12
  What gives you guys an advantage here?
• 21:14
  This, this feels like very natural.
• 21:18
  Like it's like, oh, yeah, absolutely.
• 21:20
  I want faster speeds. I want smaller size data.
• 21:24
  Like why are you guys doing it the best?
• 21:27
  You know, that's that's a very difficult question to answer
• 21:29
  just because we are patent pending still.
• 21:32
  What I can say is that we have reimagined the the architecture
• 21:37
  at its core and the way that these legacy vectorization
• 21:42
  methods work, we a lot of people mistake us for compression
• 21:46
  company. We're not compression.
• 21:47
  And so you know, the an example I can give you, or I'd say to
• 21:51
  best describe that is, you know, a vector is a multi dimensional
• 21:55
  variable and we're able to take a variable that's multi
• 21:58
  dimensional with with many, many, many dimensions.
• 22:00
  You know, you can it's up to like 1000 plus and reduce it
• 22:03
  down to 1. So if I want to describe a water
• 22:06
  bottle to you, I don't, you don't need to know what color it
• 22:09
  is. You don't need to know the exact
• 22:10
  shape of it. You don't need to know the brand
• 22:12
  of it. You don't need to know the
• 22:13
  density of the material that's used.
• 22:15
  You just need to know very limited information for you to
• 22:17
  go to discern that this is a water bottle.
• 22:19
  And so, you know, our technology kind of revolves around that
• 22:22
  thought and that, that approach rather than trying to ingest all
• 22:25
  the information. And so, you know, when people
• 22:28
  ask me like the, the similar question, how does it work?
• 22:30
  I, I, I tell them, I said close your eyes and can you still hear
• 22:34
  me and understand what it is I'm saying having closed your eyes,
• 22:36
  say yes, of course I can hear it.
• 22:38
  I said perfect. Your brain is also having an
• 22:40
  easier time processing audible information because now you've
• 22:43
  gotten rid of 1 sensory input. So that's a a very similar, I
• 22:47
  would say, approach that we have when it comes to the actual
• 22:51
  mathematics of of vectorization technology.
• 22:54
  Interesting. OK, cool.
• 22:58
  And you've priced it very affordably, I mean 20 bucks a
• 23:00
  month. So you are kind of getting into
• 23:02
  a consumer place where you're finding small businesses.
• 23:06
  What we're seeing now is there's a lot of examples of these
• 23:09
  almost micro businesses and maybe this is a case study and
• 23:13
  you can play with it a little bit.
• 23:14
  You know, let's say my mom was documenting snow conditions and
• 23:19
  weather conditions for her mountain town for the last 40 or
• 23:23
  50 years, and she might have the most complete set of that data,
• 23:28
  you know, so it, it sounds like with with key, basically she can
• 23:33
  upload all of those handwritten newsletters, vectorize it,
• 23:37
  structure the data, bring it in. And now she can have a
• 23:40
  conversation with, with her, her data and then maybe make
• 23:44
  predictions there for forward. So if in the past she was
• 23:47
  contributing to the weather page for the local newsletter, now
• 23:51
  she can maybe create a, a, a SAS software where it's like, hey,
• 23:55
  based on 50 years of historical data, here's what we can expect
• 23:58
  to see. And, you know, maybe then the
• 24:00
  ski resorts and things like that become customers of hers.
• 24:03
  Exactly. Absolutely.
• 24:04
  And so, you know, at at some point, we want to, we have an
• 24:08
  enterprise solution, but we do want to be able to often offer
• 24:11
  an open API and to open sources to a certain degree so people
• 24:15
  can build applications on top of it outside of, you know, large
• 24:18
  scale enterprise. But absolutely, you're, you're
• 24:20
  absolutely correct on that. We wanted to make it affordable.
• 24:23
  We wanted to be like the cost or Trader Joe's of, you know, AI
• 24:28
  search and inference. And so it was either you paid a
• 24:31
  premium fee, 50,000 up and you still didn't get great
• 24:35
  performance, but it was the best on the market, or you just
• 24:38
  weren't able to afford it. And you have to utilize LLMS
• 24:40
  like ChatGPT, which you're going to deal with all the issues that
• 24:42
  we discussed earlier. We wanted to be able to provide
• 24:45
  the best solution at the most affordable price.
• 24:48
  And the only reason why we have to do it is because of our
• 24:51
  vector technology and our reduction in vector storage.
• 24:54
  Our costs are so ridiculously low that we have to translate
• 24:57
  that to the customer. And so now for the same price of
• 25:00
  ChatGPT, you can get, you know, ChatGPT with us or Gemini with
• 25:04
  us, plus the vectorization technology, which outperforms
• 25:07
  all of them. Incredible.
• 25:10
  As far as the enterprise solution, what what categories
• 25:13
  or verticals of businesses have big juicy data sets that you
• 25:17
  know would benefit from this on day one that aren't really
• 25:20
  employing it right now? Very good question.
• 25:26
  I would say all of them, but you know, if you know, all data is
• 25:30
  at this point is is really, really bloated.
• 25:32
  And then when you're talking about artificial intelligence,
• 25:34
  machine learning application, it's all extremely bloated.
• 25:36
  That's why we've had to build better and better models.
• 25:38
  That's why we've had to build better and better hardware to
• 25:41
  make up for the fact that the underlying data is so bloated.
• 25:43
  It's like I ask you to find needle in a haystack, but the
• 25:46
  haystack is the size of a house, right?
• 25:48
  It's very difficult to do so it's going to take you a lot of
• 25:50
  time. And so that's been a big
• 25:52
  problem. But I would say in terms of
• 25:54
  applications where it's not being utilized the most, I
• 25:56
  would, you know, I really got to go with with edge devices and
• 26:01
  robotics and, and, and Militech, it's not being utilized there at
• 26:05
  all. A lot of, you know, Militech
• 26:07
  systems are utilizing very old archaic systems for satellites,
• 26:11
  drones, any type of autonomous vehicle, you name it.
• 26:15
  Augmented reality, virtual reality headsets like Andrew
• 26:18
  just came out with Eagle Eye. A lot of it is, you know,
• 26:20
  utilizing old technology. And so I feel like if we were to
• 26:23
  apply Green Vector into that, there'd be massive, massive
• 26:26
  benefits because now it's able to run completely locally on the
• 26:30
  edge, right? So on the device itself, if
• 26:33
  you're talking about, you know, your average Gemini user,
• 26:36
  they're going to notice a difference in terms of speed and
• 26:38
  accuracy. But it probably won't be that
• 26:40
  transformational for you as an average user of it opposed to a
• 26:44
  large scale company that is utilizing this in their
• 26:46
  application itself or for whatever purpose.
• 26:49
  Or, you know, if you're a drone that wants to be able to fly
• 26:52
  very precise routes and you need instant, instant feedback from
• 26:58
  sensors, vectorize that information, calculate something
• 27:01
  from that data, and then execute some type of instruction set,
• 27:04
  there'd be a bigger, I would say a bigger transformational
• 27:07
  process on on that end. Yeah, educate me a little bit on
• 27:11
  the workflow, like why is the vectorization an important point
• 27:14
  or important step? Yeah.
• 27:17
  So, you know, when you're uploading any type of data and
• 27:20
  you want to be able to give it, you know, any type of, let's
• 27:25
  just say, an LLM model, access to that data, the LLM turns it
• 27:29
  into numbers, right? Computers operate in the
• 27:32
  language of numbers. It's something that we as humans
• 27:34
  don't naturally always do. But mathematics is the language
• 27:37
  and of itself, you can describe anything with numbers.
• 27:40
  And so vectors is the way that we're able to do it.
• 27:42
  So I can describe, you know, a water bottle like I mentioned,
• 27:45
  or a house, or, you know, a room with a set of numbers, right?
• 27:50
  I can describe, you know, a room by its dimensions, right?
• 27:53
  Length, width, height, volume, what is the color on the walls?
• 27:57
  Dementia, like all these things. And so you have to translate it
• 28:00
  into numbers, which therefore is a vector for a model to be able
• 28:03
  to utilize, calculate some information off of and then to
• 28:06
  be able to utilize that information for either executing
• 28:09
  instruction sets or retrieving that information, spinning it
• 28:11
  back to the user. And so that's why vectorization
• 28:13
  such an important technology, it allows us to be able to give
• 28:17
  anything that we want to a model and translate it into a language
• 28:22
  that the model understands and the habits spit back back out to
• 28:25
  us in a language that we understand.
• 28:27
  And so that is what RAG is. That's what vectorization
• 28:30
  primarily is for. Nice.
• 28:32
  So you're the CTO and I'm curious about your workflows and
• 28:37
  how you're leveraging AI to basically amplify and augment
• 28:41
  what you're working on and how you see that interplaying with
• 28:44
  kind of humans and maybe where you're differentiating what
• 28:46
  tasks are are get being performed by AI agents and
• 28:49
  managed by certain team members. Yep, you said so, you know,
• 28:53
  we're, we're a small startup. And so when you're startup, you
• 28:55
  try to be as resourceful as, as possible.
• 28:57
  And the one of the beautiful things about being a startup is
• 28:59
  that because you're small, you can move very fast.
• 29:01
  So testing new ideas, new approaches, we can do it on a
• 29:04
  day-to-day basis to see what works best.
• 29:06
  We're running split tests all the time with our approaches
• 29:08
  whenever we're working on something.
• 29:11
  I utilize AI on a daily basis for every single one of my tasks
• 29:15
  from my emails to, you know, planning projects,
• 29:19
  documentation, programming, you name it.
• 29:22
  And so we've been seeing a lot of success, especially when you
• 29:26
  know, you're exploring new avenues of where can this
• 29:28
  technology be applied to by utilizing tools like Gemini Deep
• 29:32
  research, right? Being able to sift through
• 29:35
  mounds and mounds of data from research documents, textbooks,
• 29:39
  textbooks, you name it, and being able to compile all that
• 29:42
  information into one document to quickly explain is this even
• 29:45
  possible? What's the benefit of this?
• 29:47
  You know, what is the importance of that?
• 29:48
  That is something that we're doing on a day-to-day basis just
• 29:51
  because now we're expanding our application set and where we
• 29:55
  want to be able to integrate this technology that is probably
• 29:58
  has has had the biggest impact on on my day-to-day workflow.
• 30:02
  I can say when it comes to programming that, you know,
• 30:04
  Claude code had been an absolute game changer for us.
• 30:08
  That allows us to ship out updates extremely fast.
• 30:10
  We've used tools like lovable when it comes to, you know,
• 30:12
  building out the website or building out presentation
• 30:14
  portals in a, in a, in a fun way.
• 30:17
  So, you know, you're not just sifting through a slide deck
• 30:19
  like like everybody else, you're, I'm able to dynamically
• 30:21
  prevent this technology to you in a very simple, easy way that
• 30:24
  you can access no matter where you are just by clicking on the
• 30:26
  link. And so, you know, it's, it's
• 30:29
  been a crazy process and it's a lot of split testing, but
• 30:34
  overall it has had a massive, massive impact.
• 30:37
  And the more tools we use for a variety of different use cases,
• 30:39
  the more we start realizing how important green vectors is and
• 30:42
  where it can be applied into. And so that has been a major
• 30:46
  reason why we even decided to dive into mil tech applications
• 30:50
  or drone applications or satellites and whatnot.
• 30:53
  Well, you know, and now with all these tools, which is great
• 30:57
  because you know, it, it allows us to 10X and beyond kind of our
• 31:00
  productivity. But you don't have a lot of
• 31:02
  redundancy because you have a small team.
• 31:05
  So which puts your responsibility, it's very high,
• 31:10
  you know, So how do you kind of hedge against kind of redundancy
• 31:14
  and like, you know, if and kid gets the flu, does everything
• 31:17
  stop? Yeah.
• 31:18
  So very good question. That's something that you have
• 31:21
  to deal with whenever you're a small organization of any type,
• 31:24
  right? There's there, there's more
• 31:26
  important points of failure. And So what we've done is, you
• 31:30
  know, one, we're extremely, extremely lucky to have an
• 31:32
  amazing team of highly, highly intelligent people.
• 31:36
  We have some really great board members that advise us if you
• 31:40
  know, a certain member is out of Commission due to sickness or
• 31:42
  whatever the case is. We do have redundancies in place
• 31:45
  and we do have, I would say contingency methods in place.
• 31:47
  It's probably a better word for it in terms of the sops of what
• 31:51
  needs to be done, what can be paused and what can be kind of
• 31:53
  shift and pivot forward onto and to work on that in the meantime.
• 31:57
  So that way we're, we're still staying productive.
• 31:59
  That has been a, a massive, massive, I would say
• 32:04
  contributing factor to our productivity is that we always
• 32:06
  have contingencies. Is that if this doesn't work out
• 32:09
  for so and so reason, or if this fails for so and so reason, what
• 32:12
  are we gonna be doing in the meantime?
• 32:13
  Is there another project we can work on?
• 32:14
  Can we prototype this? Can we ship out that feature
• 32:16
  instead before, you know, this person gets back on board,
• 32:19
  whatever the case is. And so that allows you to move
• 32:22
  fast, break fast, pause fast, pivot fast.
• 32:25
  You know that that's like silicon's Silicon Valley's, you
• 32:28
  know, go to saying, hey, everything is about speed.
• 32:32
  And so especially when you're a start up, you don't have much to
• 32:34
  lose. And that's what makes it, you
• 32:36
  know, extremely important to understand is that if you just
• 32:38
  keep working fast, at some point, you know, you're going to
• 32:41
  figure out what clicks, you're going to figure what works.
• 32:45
  And you know, that's something that we have over these large
• 32:48
  scale enterprises, which have become very bureaucratically
• 32:49
  designed. And it takes a lot of a lot of
• 32:52
  time, I should say, for information to travel from the
• 32:54
  bottom to the top. We don't have that issue.
• 32:56
  And so, you know, that's, that's a big contributing factor to our
• 32:58
  productivity. I'd love how you think about or
• 33:01
  I'd love to know how you think about product 2 as far as like
• 33:04
  time to value, because ultimately, you know, if you're
• 33:06
  delivering a subscription product, you know, people want
• 33:10
  to get in there and see results quick.
• 33:12
  I think lovable is a great example of winning really fast.
• 33:16
  You know, where like you, you start playing with it, you're
• 33:19
  like, wow, this is really cool. You know, how do you build in
• 33:22
  those wow moments and like what's what's the 1st 5 to 10
• 33:25
  minutes of like using key look like?
• 33:28
  Yeah, so I think, you know, the, the biggest thing is 1 we make
• 33:31
  it easy to upload your documents.
• 33:32
  We, we learn from the best, right?
• 33:34
  So we're, we're taking a lot of inspiration from Chachi Petit,
• 33:38
  Google Gemini in terms of the user interface to make sure that
• 33:40
  it's sticky, that people actually enjoy utilizing this
• 33:43
  platform. The biggest benefit I would say
• 33:45
  is that getting onboarded, uploading data and we're going
• 33:48
  to be connecting, you know, data, you know, your, your
• 33:50
  accounts. So data automatically pulled in.
• 33:53
  You don't have to manually upload anything, but we make it
• 33:55
  extremely easy. And I think the biggest wow
• 33:56
  factor is going to be the speed, right?
• 33:58
  When you're able to ask an LLMA question regarding 1000 page
• 34:02
  textbook and you get an answer in 3-4 seconds and you don't
• 34:06
  need to worry about it being inaccurate.
• 34:08
  That I think is a big wow factor, right?
• 34:10
  For someone who's doing research, if you're a student or
• 34:13
  a teacher or if you're a scientist or a mathematician,
• 34:16
  being able to get answers that quick with that level of
• 34:20
  accuracy is, is absolutely ridiculous.
• 34:22
  It is just hasn't been seen before.
• 34:24
  And you can pay for speed. You know, companies have been
• 34:27
  doing it with, with compression technology and by being able to
• 34:30
  deploy more and more hardware and, and you know, horizontally
• 34:33
  and vertically scaling their applications.
• 34:35
  But you can't pay for for speeding alongside accuracy,
• 34:39
  right? It's a very difficult thing to
• 34:40
  do. You can't just buy more GPU's
• 34:42
  and get better accuracy. It's not going to work that way.
• 34:44
  And so that's a big thing with us is that you get best of both
• 34:47
  worlds. And so, you know, we live in a
• 34:48
  world now where people want information instantly, right?
• 34:51
  And unfortunately, you know, with, with LLM's like this,
• 34:54
  there's a trade off, you know, you can get it almost instantly,
• 34:56
  but is it always accurate? And then what is the, the time
• 34:59
  like to now go and, you know, verify that information by
• 35:03
  actually going on Google, looking at some Reddit forum to
• 35:06
  make sure that it's actually correct or whatever the case is.
• 35:08
  It completely ruins and completely gets rid of the, the,
• 35:11
  the need to actually utilize the LLM technology for information
• 35:14
  retrieval because now you have to second guess every single
• 35:17
  thing. And so that's something I think
• 35:19
  is going to be a big wow factor when people, you know, first hop
• 35:21
  on the platform and start using it.
• 35:23
  Yeah, and that then it starts including to industries that
• 35:27
  can't have hallucinations. You know, if you're dealing with
• 35:30
  medical records or you're talking about military
• 35:32
  applications, it's like 99.999 for Infinity.
• 35:36
  I know we can't get 100%, but like it really can't with some
• 35:40
  stuff, especially if you're doing research and like maybe
• 35:43
  you're, you're, you're professional, like reputation is
• 35:48
  on the line and you, you, you quote some, some hallucination
• 35:52
  like that could be devastating. Absolutely.
• 35:54
  So you know, we were thinking about this as well as like we
• 35:58
  want, we would love to be able to work for the hospital or
• 36:00
  veterinary hospital of of some sort and be able to connect
• 36:03
  their EMR systems directly into key.
• 36:05
  And so now every single, you know, especially with, with
• 36:08
  veterinary, because that doesn't require HIPAA compliance, but
• 36:11
  all of your patient records now are accessible via, you know,
• 36:15
  key itself. So if you're a doctor and you're
• 36:17
  on the 5th floor and you know, you're just on the go, you're,
• 36:21
  you're working with a lot of patients, doctors, you know,
• 36:22
  this is very typical of, of, you know, hospital environments.
• 36:25
  Everything is fast. Go, go, go, go, go.
• 36:27
  You don't have the time to do a lot of manual information
• 36:31
  retrieval or if you do, it takes up a lot of time, right?
• 36:33
  And then cross referencing that information with your with
• 36:36
  textbooks or journals to find out, you know, what is this
• 36:39
  specific disease? What is the, you know, the go to
• 36:42
  form of, you know, therapy or or treatment for, for the specific
• 36:45
  thing. And it that takes up time and it
• 36:47
  overall it adds up hours in in the week that you could have
• 36:50
  been more productive. Now we would be able to enable a
• 36:53
  doctor to be able to go on their phone, type in a patient name.
• 36:56
  It gives everything that you need to know regarding that
• 36:59
  patient and then cross reference that with a 2000 page textbook
• 37:02
  of some sort of repository of data that talks about specific
• 37:05
  diseases that, you know, correlate with, with whatever
• 37:07
  this patient has. And so that technology doesn't
• 37:09
  exist right now, right? And doctors still forced to do
• 37:12
  things the the old way, but with our technology we now enable
• 37:15
  that. Yeah.
• 37:16
  I mean, with medical, there's so many applications and so note
• 37:19
  many opportunities for new data sets, you know, because we're
• 37:23
  just kind of breaching the surface of the kind of
• 37:26
  biofeedback that we can gather, you know, involuntarily through
• 37:30
  our bodies, whatever, or, you know, what people actually want
• 37:34
  to upload. And in particular, that's going
• 37:36
  to be stuff that you want to hold pretty close to the chest
• 37:39
  and make sure it's private. You know, unfortunately, you
• 37:42
  know, in the United States, I'm sure the insurance companies
• 37:46
  would love to take a look and be like, oh, well, Ryan, you know,
• 37:50
  half the time you didn't walk 10,000 steps.
• 37:52
  So I don't think we're gonna, we're gonna cover your, your
• 37:55
  ankle surgery because actually, you know, there's the, the data
• 37:59
  privacy now more than ever, just because it seems like there's
• 38:02
  really kind of nefarious intent. And I hate to say that, but I
• 38:06
  think the the fact that you guys are making it the possibility
• 38:10
  for for private is is amazing. How big's the team.
• 38:16
  Team is just, we're 10 members and we have 4 board members as
• 38:21
  well, but core team is 10, yeah. Incredible How?
• 38:24
  How did you finance? So everything is bootstrapped,
• 38:28
  so self funded. This was a you know, I, I came
• 38:30
  on in October of 2025, but it had already been a A2 year
• 38:34
  project prior to me coming on. And so, you know, my, my biggest
• 38:38
  contribution in my job has been what can we apply this
• 38:42
  technology to? How do we do it?
• 38:44
  How do we get, you know, beta test set up?
• 38:46
  How do we actually integrate and go to market with this
• 38:48
  technology, not just for a rag application, but for other
• 38:51
  applications that we talked about before.
• 38:52
  That is majority, I would say the say 80 to 90% of my my work
• 38:56
  scope. You could do anything.
• 39:00
  What what, what stood out about this when they came?
• 39:03
  So when they came with the idea, what was it that that caught
• 39:06
  your attention? Yeah.
• 39:07
  So, you know, when it comes to technology like this, when I was
• 39:13
  kind of pitched on the whole idea and what the, the the
• 39:15
  proprietary technology is able to do with up to 99.5% reduction
• 39:19
  in, in vector storage. That is a number that you think
• 39:23
  of. And you, you know, you instantly
• 39:24
  start thinking, no way, this is, this is impossible, right?
• 39:27
  No one has been able to do it. The big companies that have
• 39:29
  tried and failed, we have applications and infrastructure
• 39:34
  that's built on top of this core belief that it can't be done.
• 39:36
  And you're telling me that it can't.
• 39:38
  And so, you know, when I first heard about I, you know, just
• 39:41
  like everybody else, I, I talked to about the company, they all
• 39:44
  wanna, you know, call me and say that's impossible.
• 39:46
  And then I get to show them a demo.
• 39:48
  They're like, wow, right. And so that's the same thing
• 39:50
  that got me as well. And I got very excited thinking
• 39:52
  about that how can we utilize this technology not just for RAG
• 39:56
  and we would be able to massively benefit RAG in of
• 39:59
  itself. That's the primary use case for
• 40:01
  green vectors as of right now. But what else can we can we do
• 40:05
  in industry, right? And so for me, it was almost
• 40:07
  like the invention of the LED, right?
• 40:09
  The LED light bulb, you know, once the LED was invented, now
• 40:13
  it's being utilized everywhere. It's utilized for your street
• 40:15
  lamps your, your light bulbs, your.
• 40:18
  Like light bulbs in your home, your cars, you name it.
• 40:20
  And so now it has become the new status quo that everything is
• 40:22
  built on top of. And so that was something that
• 40:24
  got me really, really exciting. And, and, you know, you can
• 40:26
  imagine how many possibilities there are where this technology
• 40:29
  can be implemented and how fun it would be to be a part of that
• 40:32
  process to build things that the world has been seen as it got.
• 40:35
  Foundational level. Absolutely.
• 40:38
  Yeah, that's super cool. When you're thinking about go to
• 40:40
  market, how do you find your first customers?
• 40:43
  So, you know, we, we're very blessed, I would say to have a
• 40:47
  extremely, extremely good board as well as team in terms of some
• 40:52
  of our connections and you know, who we're able to have
• 40:54
  discussions with. And so that has been massively
• 40:57
  beneficial. You know, we've been able to
• 40:58
  talk to some pretty large corporations from the Gecko.
• 41:01
  They've been kind of helping us and, and steering, steering us
• 41:04
  in the right direction. You know, the, the, the best way
• 41:08
  that I find is that when you're able to make the claims that we
• 41:12
  are able to make companies listen, right?
• 41:14
  You know, the executive VP directors, the CE OS, they
• 41:16
  listen because this is something that's completely brand new.
• 41:19
  So it's it's so, you know, enticing that they're almost
• 41:22
  forced to turn their heads and be like, wait, who's that?
• 41:24
  And you know, what, what did that company, what did they do?
• 41:26
  What did they just say? And so that has been, AAI would
• 41:30
  say, a really good and, and strong variable when it came to
• 41:35
  being able to have meetings and sit downs with these companies.
• 41:38
  And then, you know, one person knows another who knows another,
• 41:40
  and it becomes word of mouth and word starts spreading along
• 41:43
  these, you know, within these networks and circles that from
• 41:46
  there we've been able to get a lot of organic growth.
• 41:48
  Yeah. When you go into those
• 41:49
  conversations, are you treating them as like sales calls or as
• 41:52
  you, are you kind of giving them kind of some authority as design
• 41:56
  partners? Like how do you initiate that as
• 41:58
  a start up? Yeah.
• 42:00
  So it kind of depends on the company we're speaking with.
• 42:03
  You know, if it is majority of the company, I should say, are
• 42:06
  utilizing or wanting to utilize artificial intelligence on, you
• 42:11
  know, edge devices or specific use cases.
• 42:13
  And from there, a lot of these conversations really start out
• 42:16
  as a conversation of how can we work together, right?
• 42:18
  Does our technology overlap with yours?
• 42:20
  Can we impact you guys in a positive manner?
• 42:24
  Can it run on your, your hardware stack?
• 42:26
  What would be the theoretical before and after if we did?
• 42:29
  And from there we're then able to start working out a deal
• 42:32
  because once the numbers become apparent, now it becomes, you
• 42:35
  know, very simple, yes or no decision if we want to do it,
• 42:38
  you know, a pilot or beta to test this out.
• 42:41
  And so a lot of the conversations are very organic,
• 42:43
  not really salesy. When you're able to tell people
• 42:45
  about technology, a lot of times they start thinking themselves,
• 42:48
  where can I apply this in my company?
• 42:50
  And if this is beneficial And you know, so far we've we've had
• 42:53
  a very high, high success rate with that approach.
• 42:56
  Love it. All right, We're ready for the
• 42:57
  quantum round. Are you ready?
• 42:59
  Absolutely. Is that a piano behind you?
• 43:02
  Yes, it is. Do you play?
• 43:04
  I used to, not anymore. I haven't had the time for, for
• 43:06
  a while now, but I, I hope to, to, you know, pick it up again.
• 43:10
  It's, it's, you know, as a programmer, we're, we're so used
• 43:12
  to typing and, and doing a lot of keyboard work.
• 43:15
  I decided to know why, why not shift over my talents
• 43:18
  back-to-back to the instrument? And so I'm hoping to get back
• 43:20
  into it soon once I can, you know, open up my, my bandwidth a
• 43:24
  little bit. Well, love it.
• 43:27
  Everyone listening to this podcast loves business ideas and
• 43:29
  shiny objects. What's a?
• 43:31
  What's a business idea you have right now you feel like could go
• 43:35
  zero to 1? Let's call it 1,000,000 annual
• 43:38
  run rate in 12 months. It's a very good question and
• 43:46
  it's very difficult to answer because especially with
• 43:50
  artificial intelligence now, you don't know what idea is going to
• 43:52
  be good today and terrible tomorrow or what's going to be
• 43:55
  good today and replace tomorrow. And so I would say if we're
• 44:00
  looking at artificial intelligence in general, there's
• 44:02
  five primary layers. And so you have energy, you have
• 44:04
  chips, you have infrastructure and data, you have models, then
• 44:07
  you have your applications. You're seeing a lot of people go
• 44:09
  into the application space or the model space, right?
• 44:12
  They wanna build a new model that's better than the
• 44:14
  competition or they wanna build an application that has a better
• 44:16
  user experience. I would say those are the 1st
• 44:18
  2:00 to get, you know, beating or for you it's it's very easy
• 44:22
  to out compete another company if you have something unique.
• 44:25
  And so I would focus on the the the three foundational layers,
• 44:28
  which is energy chips and infrastructure chips is very
• 44:32
  difficult to get into because it's very costly.
• 44:34
  Hardware is hard, right? That's why you have companies
• 44:36
  like NVIDIA and energy is also very difficult because, you
• 44:41
  know, unless you come up with a a new renewable energy source or
• 44:44
  some type of method to make energy more efficient, that can
• 44:47
  be difficult, You know, if if you don't have the background
• 44:49
  for it, it requires a lot of money.
• 44:51
  And so I would focus on infrastructure problems very
• 44:53
  similar to us. There's a lot of when you start
• 44:55
  looking in technologies and how they work today, you know,
• 44:58
  within AI and outside of AI, you start looking at the system and
• 45:01
  you realize just how ridiculously old they are that
• 45:05
  this technology built on top of, but it's just become the norm.
• 45:08
  And so that's why we do it. And I think there's a lot of
• 45:10
  niches in the infrastructure space that if you're a little
• 45:14
  creative with the help of, you know, LLMS and being able to
• 45:17
  educate yourself on this topic, you can come with some pretty
• 45:19
  impressive approaches to innovate on, on the wheel that's
• 45:23
  already been invented and just figure out how to make it a
• 45:24
  little bit better. And if you're able to make
• 45:26
  infrastructure a little bit better, the impact once you get
• 45:29
  down to the application layer becomes massive, right?
• 45:32
  When you're looking at from macro view, a 1% change in
• 45:36
  efficiency can mean a massive, massive benefit when you're
• 45:39
  applying it to millions of users or consumers. 100% Do you have a
• 45:44
  epic failure that you like to talk about?
• 45:47
  Too many, too many. So you know what what, which one
• 45:51
  should I start with? I think that's a prerequisite if
• 45:53
  you want to be an entrepreneur and if you want to work in a
• 45:55
  start up, right? It's like how many times are you
• 45:57
  willing to fail before you finally hit the one?
• 46:00
  And it's funny now because I can't say for certain that this
• 46:03
  company is also going to be absolutely you can never say
• 46:05
  that when you're a start up, right.
• 46:08
  You know, I, I founded a, a automation and data analytics
• 46:12
  platform called Realtor Copilot. And so I wanted to, you know, I
• 46:17
  have my real estate license. I got licensed when I was 18.
• 46:20
  And so, you know, I'm able to see a lot of Realtors in the
• 46:22
  space. And you know, a lot of them lack
• 46:24
  the tools that I would say level the playing field between a
• 46:28
  agent that is just starting out with no knowledge and you know,
• 46:32
  no money, no team comparatively to an agent that's been in the
• 46:34
  industry for 10 plus years that has a team to help them with
• 46:37
  marketing, follow up sales, you name it.
• 46:39
  And so I wanted to build a platform that levelled that
• 46:42
  playing field by introducing AI automation.
• 46:45
  And this was, you know, year, let's say two years ago now.
• 46:49
  And so I worked an entire year on this project and, you know,
• 46:53
  we're able to scale it up and build it and I'm ready for
• 46:56
  launch. And I'm based just outside of
• 46:58
  Toronto. And so if you know anything
• 46:59
  about, you know, you know, Canadian real estate specific
• 47:01
  within the GTA is that it completely crashed.
• 47:04
  The market is, is in a, in a pretty bad spot.
• 47:07
  And you're seeing a record number of agents no longer
• 47:09
  renewing their licenses. They're not doing any deals.
• 47:12
  There's a crazy statistic. I think it's like 90% of deals
• 47:14
  are by 2% of agents, something like that.
• 47:17
  Don't quote me on that number, but it's something ridiculous
• 47:19
  like that. And so you're seeing a massive
• 47:21
  inflation in terms of people that actually have licenses, but
• 47:24
  not anyone is doing any deal flow.
• 47:26
  And so from there, the market completely just pull the rug
• 47:31
  from out from under me. And I'm thinking that I can try
• 47:34
  to double down and try to scrape users to make this profitable.
• 47:37
  Or would it be better to kind of chop this up, right?
• 47:40
  Chop up this car, if you will, take out the engine, take out
• 47:42
  the chassis, take out the wheels and repurpose it for something
• 47:45
  else. And then I would say a lot of
• 47:47
  that experience the the process and the automation technologies
• 47:52
  I was able to utilize in my next venture, which was Skylar Bay at
• 47:54
  consulting, where we built custom AI automations and custom
• 47:58
  software for, for clients. And so that was a big, big help.
• 48:02
  And that's what I was been. You know what I was doing before
• 48:04
  I joined. More for more folks.
• 48:06
  Cool. Yeah, I was talking to Siggy.
• 48:09
  He's the founder of Overtune, and he said being a founder is
• 48:13
  like, you know, when Homer Simpson was a successful boxer,
• 48:18
  just because he would stand there and take shots until the
• 48:21
  guy got so tired that he could just push him over.
• 48:24
  Correct, correct. You know, people ask me, you
• 48:27
  know, what is the number one trait that you'd say for an
• 48:29
  entrepreneur? And I wouldn't even say it's not
• 48:32
  intelligence. It's not, you know, your
• 48:36
  capability, your competency, It's none of these things.
• 48:37
  I would say the number one thing is your pain tolerance, right?
• 48:41
  It's easy to work and to work on a project when you're excited
• 48:43
  about it. It's easy to work on it when you
• 48:45
  know you're going to win. It's very difficult to do that
• 48:50
  when you don't know when you're going to win, when you can't see
• 48:52
  the finish line and when you're just getting hit in the face day
• 48:55
  after day after day after day. And you're almost, you're almost
• 48:59
  mad. You've almost got a little crazy
• 49:00
  to it. And so, you know, a lot of
• 49:02
  people, you know, you look at Einstein, you look at any of
• 49:06
  these founders that have, you know, been successful in, in
• 49:09
  industry, all of them say something very similar.
• 49:11
  And you know, one thing I tell people is that you just got to
• 49:14
  show up every single day, right? If you, if you throw, as you
• 49:17
  know, 100 darts on the board does, you know, a very, very
• 49:22
  high chance you're going to hit bullseye at least once, right?
• 49:25
  It's just a matter of times you're willing to step up to the
• 49:27
  plate, show up every day, act like a professional and one day
• 49:30
  you might hit, right. So pain tolerance and endurance
• 49:34
  I would say are the the number one traits that any founder
• 49:36
  should should look to employ if they don't have it already.
• 49:40
  Sounds like such fun. Yeah, absolutely right.
• 49:43
  It's, you know, it, it looks glamorous on the outside.
• 49:46
  And then you start doing it. You realize I don't want to do
• 49:48
  this, right? But it's still like the problem.
• 49:51
  Yeah, it's too late. When the, when the idea is so
• 49:54
  enticing and it's so exciting, right?
• 49:56
  You don't have a choice, right? You, you say I could go and get
• 49:58
  a job and I can take a, a simpler route, you know, if, if,
• 50:02
  if you know if you will and go do that.
• 50:04
  But when you know you have this amazing idea staring at you
• 50:07
  right in the face, and then the idea or the thought of what
• 50:10
  could be starts creeping in, that's when you're like, OK, I
• 50:13
  can't do this anymore. And you're just forced to do it.
• 50:15
  And sometimes it can be a toxic, toxic relationship.
• 50:18
  It's like, I love her, but she doesn't love me back until she
• 50:21
  does. And then it works.
• 50:23
  Yeah, it's the idea of Type 2 fun.
• 50:25
  You know, it's completely miserable when you're doing it,
• 50:28
  but in retrospect you're like, wow, that was awesome.
• 50:31
  Absolutely, absolutely. Learning to embrace the suck,
• 50:34
  That's something that, you know, I think a lot of Navy SEAL guys
• 50:37
  say, I think David Goggin says that a lot.
• 50:39
  It's like learning to embrace the suck and enjoying the bad
• 50:42
  times. It's almost a form of brainwash.
• 50:46
  But when you're able to do that, you know you end up having fun
• 50:49
  even on your worst days, right? Just because you're focusing on
• 50:53
  the journey rather than the destination.
• 50:55
  And so just being able to play the game, win or lose is is a
• 50:58
  win. And so every single day becomes
• 51:00
  a good day, despite the fact it might be a 12 hour, 15 hour work
• 51:03
  day, you're off, no sleep, you've been coding for hours,
• 51:06
  doing something for hours. It still becomes an enjoyable
• 51:08
  process. And so I think if people rewired
• 51:10
  what they, you know, thought of as bad or considering a bad day
• 51:14
  or that, you know, this wasn't the best of days or this was
• 51:16
  exhausting and start reframing their mindset towards, you know,
• 51:19
  I got to play the game today. It was fun, right?
• 51:22
  And so I think that's that's a big, big mental shift that if
• 51:25
  people start deploying, it actually becomes easier to have
• 51:27
  more endurance because now you're looking at it
• 51:28
  differently. It's no longer a negative.
• 51:30
  It's always a positive love. It if you were entering school
• 51:35
  college right now, what would you study?
• 51:43
  I would stay away from programs that are very, very technical in
• 51:50
  terms of teaching you a very limited skill set to perform one
• 51:53
  or two or three tasks, right. So you see this with, you know,
• 51:58
  programs like, you know, HR programs, sometimes marketing
• 52:01
  programs, a lot of which are in the business scope, even
• 52:03
  engineering as well. Some there's some universities
• 52:05
  and programs that are very limited scope.
• 52:07
  They teach you a few tasks and they they tell you to become a
• 52:09
  specialist in this. I think because of artificial
• 52:12
  intelligence, it's catching up faster than anybody can, you
• 52:15
  know, hope to compete with you should go to university if you
• 52:18
  do plan on going and get yourself in a program that
• 52:21
  forces you to think so. This is why I like engineering
• 52:24
  so much, because what you learn in university is not nearly as
• 52:28
  important as to you learning how to learn and you learning how to
• 52:31
  break up a big problem into small chunks and then and being
• 52:34
  able to tackle 1 chunk in time. That is a much more valuable
• 52:37
  skill set than you know, remembering what the specific
• 52:40
  textbook said, right? I've studied computer science.
• 52:43
  I don't remember 90% of the stuff that was taught in my, in
• 52:46
  my lectures. In fact, I barely went to class.
• 52:48
  And so, you know, but learning how to learn is such a valuable
• 52:52
  skill set that when you combine that now with LLM technology and
• 52:55
  then with products like key and, and I'm sure there's going to be
• 52:57
  dozens and dozens, hundreds of more products that help you
• 53:00
  learn at such a fast rate. You become very, very difficult
• 53:04
  to compete against because you know, you're able to bring the
• 53:06
  macro while the AI is able to bring the micro.
• 53:09
  And so, you know, I can't hope to get better than machine
• 53:12
  learning at, you know, remembering A textbook.
• 53:14
  It's just impossible. I'm never going to be as good as
• 53:16
  it. But how do you apply that
• 53:17
  information from that textbook? How do you think big picture?
• 53:20
  How do you connect dots that typically should not be
• 53:22
  connected to come up with unique ideas and solutions and
• 53:25
  approaches? That's where the big skill set
• 53:27
  is. And if you learn how to do that,
• 53:28
  you know you're guaranteed job no matter what you study.
• 53:32
  Is there a product you'd recommend, whether it's ATV show
• 53:36
  or a movie or food or what's what's something you're really
• 53:40
  excited about that people should know about?
• 53:45
  I'm very excited about a company called Clicks and they have a
• 53:49
  new small phone coming out and I forget what it's called.
• 53:52
  I think it's called the connector, Clicks Connector.
• 53:54
  I'm very excited about that because I think right now with
• 53:58
  smartphone technology, we have so much in our hand and it's
• 54:01
  becoming, it's coming to a point where now it's too stimulating,
• 54:04
  right? And you know, you're looking at
• 54:05
  crazy statistics for like 80% of tweets are AI generated or bots.
• 54:09
  And so you have people that are spending ridiculous amount of
• 54:11
  time on these applications and they're not getting anything out
• 54:13
  of it. And so being able to narrow your
• 54:15
  scope of focus, but still being able to access these technology
• 54:19
  to help you learn and help you interact with the world in the
• 54:21
  way that they are doing now, I think that's very, very
• 54:23
  impressive and very, very useful.
• 54:26
  I would say a big productivity hack.
• 54:28
  A lot of people ask me, how do you stay focused for so long?
• 54:30
  Just get rid of all the distractions, put yourself in a
• 54:32
  room with nothing but you and your computer, block out all
• 54:35
  social media applications and just see how fast you get things
• 54:38
  done right. And so a company like Clicks is
• 54:41
  focusing that productivity space by going back and, you know,
• 54:43
  kind of, but back in, in, in time to what phones used to be,
• 54:46
  which was an application, a device purely for communication.
• 54:50
  It wasn't for posting selfies. It wasn't for, you know, surfing
• 54:53
  the web or doing all these other things.
• 54:55
  And I, you know, that's something that they're focusing
• 54:57
  on. I think a lot of people can
• 54:58
  benefit from that, especially when you're looking at the
• 55:00
  average attention span going down, you're looking at a lot of
• 55:03
  issues in the education system. Kids aren't able to remember
• 55:05
  information as much. They're not, you know, they're
• 55:08
  losing critical thinking skills. A lot of that I do think has to
• 55:12
  do with social media and just reducing people's attention
• 55:15
  span. Netflix is now planning, I
• 55:17
  believe, to adopt A vertical scroll type of function similar
• 55:20
  to Instagram or Facebook to make it more interactive for, for the
• 55:23
  younger generation. And it's like we should go back
• 55:25
  to how do you stay engaged when, you know, watching a 2 1/2 hour
• 55:29
  movie like we did back in the day?
• 55:31
  How do you get engaged when you're reading a book for three
• 55:33
  hours? How do you stay focused for
• 55:34
  longer periods of time and kind of opening up your, your, your
• 55:38
  mind to be able to absorb all these ideas for, for prolonged
• 55:40
  periods of time. I think that's a massive
• 55:42
  superpower in the in the modern day, technology has made it
• 55:45
  extremely easy to access information and it's the biggest
• 55:48
  repository for information and education.
• 55:50
  But they made it so difficult to utilize because of the attention
• 55:54
  issues and the applications that they offer us.
• 55:57
  Yeah, no, I this is a good one man.
• 55:58
  I appreciate you saying this. It's clicks dot tech if people
• 56:01
  want to check it out. It said designed for doing, not
• 56:04
  doom scrolling. And it looks like a BlackBerry,
• 56:07
  you know, so bring back the keyboard.
• 56:10
  I I've kind of had this, this instinct for a while now.
• 56:13
  You know, it's like it just TikTok and Instagram and YouTube
• 56:17
  just kind of lost its rostrum. It's like I don't like it.
• 56:20
  So if I'm on my phone and I have the doom scrolling time, I'll
• 56:22
  just play Clash, you know? Exactly.
• 56:26
  Truly, it's just like, it's just too much, you know?
• 56:29
  So I think an idea like this is fantastic because ultimately
• 56:34
  even, you know, even for me in particular, you know, if I have
• 56:39
  the option to just go like drown 90 seconds in in TikTok videos,
• 56:43
  I'll do it, you know? And it's so easy to do.
• 56:47
  It's so easy to do. They, they, they're paying data
• 56:49
  scientists and engineers millions of dollars to figure
• 56:52
  out how to keep you on the platform for as long as
• 56:54
  possible. And you're now seeing, did you
• 56:57
  know, we're getting closer and closer, I think I believe it's
• 56:59
  called dead Internet theory, which is like a theory that, you
• 57:01
  know, majority of the content on the Internet is going to be
• 57:04
  computer generated, AI generated.
• 57:06
  And so now it's like, it's not even organic material coming
• 57:09
  from actual people. Like if you go to your average,
• 57:11
  you know, Twitter post or Twitter thread, you just look at
• 57:14
  the comments, a lot of it is bots.
• 57:16
  And so you know, that can be used for a variety of use cases.
• 57:18
  When you're talking about propaganda or pushing ideologies
• 57:22
  or one person operating 2030 accounts, the information that
• 57:25
  you're getting is not really, I would say accurate to what
• 57:27
  people are actually thinking, what they actually do want to
• 57:29
  discuss. And so it, it becomes very
• 57:31
  dangerous when you allow that to happen, you can sway public
• 57:34
  opinion very easily. This has been proven time and
• 57:37
  time again. And so, you know, going back and
• 57:40
  making life simpler, I think is going to be the next, you know,
• 57:43
  superhero trait that people start deploying.
• 57:46
  You're already seeing, seeing people wanting to go
• 57:47
  minimalistic. You're seeing people wanting to
• 57:49
  live off the grid doing things the unconventional way because
• 57:53
  now they're sick and tired of, you know, the constant
• 57:55
  notifications on the phone, the constant, you know, doom
• 57:58
  scrolling, the constant this, the constant dopamine firing off
• 58:00
  in your brain. Then by the end of the day, you
• 58:01
  feel exhausted. Your sleep isn't good, you feel
• 58:04
  lethargic all the time. And you just realize, if I got
• 58:06
  off this thing, I'm so much more focused.
• 58:08
  I feel so much more alert. I feel healthier, I'm moving
• 58:12
  better, I'm thinking better. And so, you know, now adopting
• 58:15
  technology like that, I think that's going to be a big, big
• 58:18
  trend. You're already starting to see
• 58:19
  it. And I'm excited to see what, you
• 58:20
  know, companies like Clicks and anyone else do.
• 58:23
  And I'm definitely going to be playing around with some of that
• 58:25
  stuff because, you know, it's just so valuable when you can
• 58:28
  control your attention like that.
• 58:30
  Percent You know when you have space in your life there's room
• 58:35
  for poetry and beauty otherwise, absolutely it's that you're
• 58:39
  missing a richness you know cool we.
• 58:41
  We we've forgotten the greats, right?
• 58:43
  If I, if I speak about Dostoevsky, you know, people
• 58:45
  don't know what I'm talking about.
• 58:46
  They think I'm 100 years old. And it's like this was the norm
• 58:49
  when we were growing up. As you know, we had cartoons.
• 58:51
  Yes, we did, but a lot of our source of fun and learning, you
• 58:56
  know, the way that we learned imagination, creativity was via
• 58:58
  books, right? We learned it by reading the
• 58:59
  Harry Potter's and, you know, whatever the case is and it, it
• 59:03
  has so many inherent benefits and there's so many studies that
• 59:06
  you know, can probably prove this as well, that when you
• 59:09
  teach people and especially children how to think rather
• 59:12
  than, you know, specific facts or memorization or flooding them
• 59:16
  with information, you end up getting much stronger critical
• 59:19
  thinkers. And that applies across the
• 59:22
  board to anything and everything, whatever it is that
• 59:25
  you want to be doing in life or industry or whatever the case
• 59:27
  is. When you when you learn how to
• 59:29
  think at at large and you learn to be creative and to connect
• 59:32
  ideas, you know you're, you're a superhuman in of itself,
• 59:35
  especially in this this modern day and age.
• 59:39
  Incredible Ankit. This has been so much fun, I
• 59:41
  really appreciate it. The website is Morphos dot AI.
• 59:45
  That's MORPHOS dot AI. What's the best place for people
• 59:50
  to connect with you personally online?
• 59:53
  LinkedIn would definitely be the best.
• 59:54
  LinkedIn and Kadensa, you know my, my first and last name.
• 59:58
  Feel free to, you know, follow me, shoot me a message.
• 1:00:02
  I love speaking with founders, tech enthusiasts.
• 1:00:05
  Anyone involved in the tech space?
• 1:00:07
  I'm excited about technology and wants to discuss.
• 1:00:09
  Love to have a conversation with you.
• 1:00:11
  That'd be the the the best place to reach me.
• 1:00:13
  Amazing. Thanks for joining me.
• 1:00:15
  Awesome. Appreciate it, Ryan.
• 1:00:16
  Thank you very much. This episode is brought to you
• 1:00:21
  by MO Burst, the AI first growth engine trusted by iconic
• 1:00:24
  startups like Calm, Robin Hood, Reddit, Uber and Google.
• 1:00:30
  Moburst is a digital, full service, mobile first marketing
• 1:00:33
  agency that helps companies to scale and become category
• 1:00:36
  leaders. Whether you're finding product
• 1:00:38
  market fit or ready to accelerate growth, Moburst
• 1:00:41
  delivers creative, data-driven solutions that help with scale
• 1:00:45
  and efficiency and impact. At Moburst, we're not just
• 1:00:49
  guessing what works, we're predicting it.
• 1:00:51
  As an AI first agency, Moburst uses cutting edge tools to
• 1:00:55
  analyze. User behavior adapt creative in
• 1:00:58
  real time and. Cover hidden growth
• 1:01:01
  opportunities. The result?
• 1:01:02
  Marketing that's faster, smarter, and built to convert
• 1:01:06
  from early stage startups to global brands.
• 1:01:08
  We've helped hundreds of companies overcome complex
• 1:01:11
  challenges, unlock new audiences, and optimize every
• 1:01:15
  metric that matters. At Moberst, we believe that
• 1:01:18
  creative thinking, advanced technology, and data are the
• 1:01:21
  core drivers of success at scale.
• 1:01:23
  If you're serious about scaling your business with precision and
• 1:01:26
  purpose, it's time to talk to Moberst.
• 1:01:29
  Visit moberst.com and let's build something category
• 1:01:32
  defining together.