HubSpot Co-Founder Introduces The Future Of AI Agents | Fiverr Marketplace For AI Agents

00:49:13
https://www.youtube.com/watch?v=WV9gT1oox-U

Resumo

TLDRThe discussion centers around agent.ai, a new AI platform offering a suite of free tools and a focus on AI agent creation. Daresh, the creator, highlights the platform's capabilities in deploying AI agents to achieve complex tasks by leveraging multiple AI tools. The conversation covers how AI can transform processes in marketing, such as automating CRM lead routing and creating targeted LinkedIn content. Agent.ai is designed to democratize software creation, enabling non-coders to participate in AI innovation through low-code solutions. By integrating different AI models, users can optimize the performance of their AI agents, providing benefits such as increased productivity, task automation, and potential monetization of expertise. As AI improves, its application expands beyond traditional frameworks, allowing businesses to enhance operations, efficiency, and creativity. The platform reflects a broader trend towards making AI tools accessible and functional for diverse uses, promising advancements in individual productivity and broader market transformations.

Conclusões

  • 🤖 agent.ai is a platform offering free AI tools, focused on customizable AI agents.
  • 🚀 AI agents use multi-step reasoning and multiple AI tools to achieve goals.
  • 💡 AI agents can greatly benefit marketers by automating content creation and CRM tasks.
  • 📊 agent.ai allows non-coders to create AI agents with a low-code approach.
  • 🎯 The platform integrates various AI models, enhancing task efficiency.
  • 💼 AI can transform businesses by optimizing processes and increasing productivity.
  • 🧩 AI agents offer potential for monetizing expertise and improving market efficiency.
  • 📈 AI tools help in strategic social media content creation, impacting engagement.
  • 🔧 agent.ai’s modular framework supports adaptive AI development.
  • 🌟 The potential of AI is in unlocking creativity and expanding problem-solving capabilities.

Linha do tempo

  • 00:00:00 - 00:05:00

    The discussion begins with a brand new platform being introduced by Daresh, offering free AI tools. The marketing community is encouraged to persuade Kieran to release his AI agent. The conversation then transitions to the concept of AI agents and how they differ from traditional tools by automating multistep tasks.

  • 00:05:00 - 00:10:00

    Daresh explains what an AI agent is and emphasizes their potential to automate complex processes traditionally handled by human interaction. A key advantage is their ability to integrate multiple AI models and tools to achieve goals, compared to a simple chat interface.

  • 00:10:00 - 00:15:00

    The conversation continues with examples of how AI agents can streamline processes, like lead routing in CRM systems, by performing tasks autonomously once set up by a human. Daresh outlines a project he used this approach for, facilitating content creation for LinkedIn based on video transcripts.

  • 00:15:00 - 00:20:00

    Various styles of content and writing approaches are discussed, such as educational posts and spicy takes, showing how AI can generate and iterate content, allowing users to select different styles and formats suited for specific platforms like LinkedIn.

  • 00:20:00 - 00:25:00

    Darresh shares the success of using AI-generated content for a LinkedIn post that outperformed usual engagement. The idea of using AI for content curation and analysis is seen as a valuable tool for enhancing creativity and productivity efficiently.

  • 00:25:00 - 00:30:00

    Discussion on future development includes the potential for agents to post directly to platforms and track performance, providing feedback on engagement levels and which styles work best. Daresh also hints at plans for further integration of tools.

  • 00:30:00 - 00:35:00

    There is a detailed discussion on the possibility of having a professional network for AI agents, allowing users to discover, rate, and possibly pay for agents that perform well. This network could democratize access to domain expertise and usage.

  • 00:35:00 - 00:40:00

    Kieran expresses hesitation about sharing his AI agents due to the potential of them making his creative processes widely accessible, questioning how to maintain uniqueness or competitive advantage in AI-enhanced task automation.

  • 00:40:00 - 00:49:13

    Finally, the discussion touches on the broader implications of AI in the workspace, with possibilities for composing different agents for specific tasks. Real-world examples of AI enhancing workplace productivity and creativity are highlighted.

Mostrar mais

Mapa mental

Vídeo de perguntas e respostas

  • What is agent.ai?

    agent.ai is a platform offering free AI tools and focuses on AI agents to accomplish multi-step tasks.

  • How do AI agents differ from traditional AI models?

    AI agents use a combination of AI tools to achieve complex, multi-step goals, unlike traditional models which tend to operate within single-step, conversational frameworks.

  • What are some applications of AI agents in marketing?

    AI agents can automate tasks like lead routing in CRM systems and generating social media content that fits specific styles and engagement strategies.

  • What role does agent.ai play in building AI agents?

    agent.ai provides a platform where users can create, test, and share AI agents integrated with various AI tools to accomplish specific tasks.

  • Can AI agents be used by non-programmers?

    Yes, the platform is designed to help non-programmers create AI agents using an approachable, low-code environment.

  • How does AI help in content creation for platforms like LinkedIn?

    AI can assist by analyzing content characteristics, generating creative templates, and automating the creation of engaging post drafts tailored to specific platforms like LinkedIn.

  • What is the potential commercial impact of AI agents?

    AI agents can transform business by automating and optimizing processes, enhancing productivity, and allowing individuals to monetize their expertise efficiently.

  • How do different AI models work on agent.ai?

    agent.ai integrates multiple AI models, allowing users to select the most suitable one for each stage of their task, enhancing flexibility and effectiveness.

  • What measure is there for new users of agent.ai?

    New users of agent.ai receive 100 free credits, and they can obtain additional credits using promotional codes.

  • What are some additional features or capabilities of agent.ai?

    The platform allows integration of various AI models, offers extensive toolkits for building personalized AI workflows, and supports collaborative AI development and experimentation.

Ver mais resumos de vídeos

Obtenha acesso instantâneo a resumos gratuitos de vídeos do YouTube com tecnologia de IA!
Legendas
en
Rolagem automática:
  • 00:00:00
    well first of all we're sharing a brand
  • 00:00:02
    new platform that daresh is working on
  • 00:00:04
    that you're going to want to check out
  • 00:00:04
    that has lots of free AI tools daresh is
  • 00:00:07
    giving the marketing instagra Community
  • 00:00:09
    extra credits so more usage of that tool
  • 00:00:12
    and I would love it if you all could
  • 00:00:14
    harass Kieran in the comments to release
  • 00:00:17
    the tool he's got an amazing AI agent
  • 00:00:19
    that he is being selfish and not wanting
  • 00:00:21
    to share you need to hit him up in the
  • 00:00:24
    comments and get him to share this tool
  • 00:00:27
    because it is sick that and some
  • 00:00:29
    learnings on the future of AI all from
  • 00:00:31
    daresh all on Today
  • 00:00:33
    [Music]
  • 00:00:35
    Show dares thanks for so much for coming
  • 00:00:37
    on the show I've been teasing the show
  • 00:00:39
    uh over the last couple of weeks on
  • 00:00:41
    LinkedIn I am an avid user of agent that
  • 00:00:44
    AI so we are excited to have you on
  • 00:00:46
    excited to be on this is a long time
  • 00:00:48
    coming I know I feel like I feel like
  • 00:00:50
    we've been talking to D M about coming
  • 00:00:51
    on the Pod for like a year we finally
  • 00:00:53
    have like but I want something like
  • 00:00:55
    worthy of your time folks it's like okay
  • 00:00:57
    I had to go like build some real stuff
  • 00:00:59
    I'm a user of real stuff and what we
  • 00:01:01
    might want to do is just jump into like
  • 00:01:03
    what do we even mean by an agent because
  • 00:01:05
    you were kind enough to give me some
  • 00:01:06
    access I've been like hammering away on
  • 00:01:09
    Claude for a long time to improve my
  • 00:01:11
    writing process I hit a new goal on
  • 00:01:13
    LinkedIn over the last week some of it
  • 00:01:15
    is trolling but like up to like over a
  • 00:01:17
    million impressions a week and actually
  • 00:01:20
    one of the things that's really working
  • 00:01:22
    is AI assistance to help me write and
  • 00:01:25
    your agent unlocked a ton of creativity
  • 00:01:28
    so I want to like get into to show
  • 00:01:30
    people you know what what are these
  • 00:01:32
    agents and how do you actually use them
  • 00:01:34
    let me do this dares before I get
  • 00:01:36
    specifically into this agent that we can
  • 00:01:37
    kind of go back and forth on this do you
  • 00:01:39
    want to just maybe give a little summary
  • 00:01:42
    of like what even is an AI agent so you
  • 00:01:44
    can kind of set the scene of like what
  • 00:01:46
    is an AI agent and why are you so
  • 00:01:49
    excited to build this platform sure um
  • 00:01:51
    so let's we'll we'll take a step back so
  • 00:01:54
    there are varying definitions of what an
  • 00:01:56
    AI agent actually is and I think people
  • 00:01:59
    sometimes tend to over complicated and
  • 00:02:01
    so there's a spectrum of capabilities
  • 00:02:02
    you have very very simple agents and you
  • 00:02:04
    can have sophisticated agents that do
  • 00:02:06
    kind of multi-agent things and reasoning
  • 00:02:08
    but at the bare minimum I think of an AI
  • 00:02:10
    agent as a piece of software that uses
  • 00:02:13
    AI to accomplish a multistep goal a goal
  • 00:02:17
    that requires multiple steps to get
  • 00:02:18
    through so if you think about it uh the
  • 00:02:20
    best comparison is that when you're
  • 00:02:21
    using uh something like a chat GPT a
  • 00:02:23
    conversational tool you're going back
  • 00:02:25
    and forth you're saying write me a blog
  • 00:02:27
    post or do this and tweak that but it's
  • 00:02:28
    a very kind of synchronous back and
  • 00:02:30
    forth model whereas with an agent you're
  • 00:02:32
    giving it a higher order goal and it may
  • 00:02:34
    need to invoke llms it may need to
  • 00:02:36
    invoke multiple llms um and use other AI
  • 00:02:38
    tools and some classic tools and bring
  • 00:02:40
    it all together to pull together
  • 00:02:42
    whatever goal you're trying to
  • 00:02:43
    accomplish so that's the difference
  • 00:02:44
    between an agent and a in a regular kind
  • 00:02:46
    of conversational ux and the reason I
  • 00:02:49
    think it's interesting especially for
  • 00:02:50
    the kind of marketers and the audience
  • 00:02:52
    the way I think about it is that back in
  • 00:02:54
    the day uh like way back in the day you
  • 00:02:56
    know we used to do kind of what is now
  • 00:02:58
    called marketing automation sort of by
  • 00:03:00
    hand right we had developers in the back
  • 00:03:02
    room working in b2c companies saying oh
  • 00:03:04
    when this happens I want to send this
  • 00:03:05
    email out and track this or whatever and
  • 00:03:07
    then Along Came the marketing automation
  • 00:03:09
    category that says hey we're going to
  • 00:03:10
    have non developers using like a
  • 00:03:13
    workflow Builder HubSpot has one there
  • 00:03:14
    are others out there basically automate
  • 00:03:17
    some number of these things right and so
  • 00:03:18
    it's a script a multi-step thing that
  • 00:03:20
    could do that so A simple way to think
  • 00:03:22
    about it is one class of Agents can be
  • 00:03:25
    these next Generation automations so I
  • 00:03:27
    know what I need to do but now that we
  • 00:03:29
    have access to die the kinds of little
  • 00:03:31
    steps that you can do are actually
  • 00:03:33
    mind-bogglingly powerful right things
  • 00:03:35
    you could never do before now can be
  • 00:03:36
    part of that automation script so that's
  • 00:03:38
    one way to think about it darash
  • 00:03:40
    building on that you know if somebody's
  • 00:03:42
    watching the show today and they're like
  • 00:03:43
    oh okay when should I think about using
  • 00:03:45
    kind of a traditional chat or
  • 00:03:47
    conversational interface versus an agent
  • 00:03:50
    like how should they think about that so
  • 00:03:52
    that they don't you know go down the
  • 00:03:54
    wrong path and waste a bunch of time or
  • 00:03:55
    get frustrated everyone should stay
  • 00:03:57
    practical uh the time to use it agent is
  • 00:04:00
    when you have something where you need a
  • 00:04:02
    human of the loop but there are lots of
  • 00:04:04
    steps that can be done without human
  • 00:04:06
    intervention so maybe the human kicks it
  • 00:04:07
    off maybe the human reviews the work
  • 00:04:09
    before it goes out so I'll give you an
  • 00:04:10
    example for instance uh you could in
  • 00:04:12
    theory say I want to build a uh I not
  • 00:04:15
    build I want to do like lead routing in
  • 00:04:18
    my CRM using chat gbt like I'm a small
  • 00:04:20
    business I'm going to take the customer
  • 00:04:22
    data I'm going to paste it into chat GPT
  • 00:04:24
    or Claude and then it's going to tell me
  • 00:04:25
    things whatever it's like you could do
  • 00:04:27
    that that's sort of a manual process but
  • 00:04:28
    every interaction is sort of human base
  • 00:04:30
    there so you're not really automating a
  • 00:04:32
    lot other than the kind intermediate
  • 00:04:33
    steps but you could also Imagine hey I
  • 00:04:35
    want to build an agent that says bring
  • 00:04:36
    me the most recent company that was
  • 00:04:38
    added to HubSpot CRM let's say I want to
  • 00:04:40
    go some do some research on that based
  • 00:04:42
    on what's there I'm going to invoke
  • 00:04:43
    perplexity to answer some questions that
  • 00:04:44
    I have then I'm going to like write out
  • 00:04:46
    my rules as far as how I want my lead
  • 00:04:47
    routing to happen if it's the leads from
  • 00:04:49
    Europe I wanted to do this if it's in
  • 00:04:51
    the tech business I wanted to do that
  • 00:04:52
    and R it and all that is doable with
  • 00:04:55
    today's technology with what I think is
  • 00:04:57
    like a high degree of accuracy and
  • 00:04:58
    precision right because you're reducing
  • 00:05:00
    the numbers of degrees of freedom at
  • 00:05:02
    each step of the process like yep AI is
  • 00:05:04
    good at that it can do that and you're
  • 00:05:05
    just sort of chaining it all together
  • 00:05:07
    and Auto or at least semi-automating it
  • 00:05:09
    uh so that's where you want kind of
  • 00:05:11
    agents to get built where you sort of
  • 00:05:13
    know what to do manually and you could
  • 00:05:15
    do it manually but you want to automate
  • 00:05:17
    some of those steps uh instead of trying
  • 00:05:18
    to do this kind of oneoff shot with chat
  • 00:05:21
    GPT we have a perfect example this is
  • 00:05:23
    actually sets up the the kind of agent
  • 00:05:25
    that I got to build on Dar meses you
  • 00:05:27
    platform really well so one of the
  • 00:05:29
    things that I've been obsessed about
  • 00:05:31
    with AI in general is when can AI take
  • 00:05:34
    some things and build you a template to
  • 00:05:37
    replicate that thing because whenever it
  • 00:05:39
    can do that my mind starts to melt right
  • 00:05:41
    because I I was like hey if I can take
  • 00:05:43
    like olav's ads and
  • 00:05:47
    templatized for the AI assistant to be
  • 00:05:49
    able to replicate that then everything
  • 00:05:51
    can be templatized like every singular
  • 00:05:54
    thing and that you can actually go do
  • 00:05:56
    can be templatized in some way and
  • 00:05:58
    specifically for Ryden what what I
  • 00:05:59
    wanted to do was start to figure out if
  • 00:06:01
    I could create processes to take good
  • 00:06:04
    and replicate it for platforms like for
  • 00:06:06
    social platforms and so the thing uh I
  • 00:06:09
    was kind of talking to you about dares
  • 00:06:10
    when we got into this conversation was I
  • 00:06:12
    was trying to teach the llm to be able
  • 00:06:15
    to figure out like what is a good short
  • 00:06:19
    form piece of content right like how do
  • 00:06:21
    you create a good short form piece of
  • 00:06:23
    content and then how do you create it
  • 00:06:25
    specifically for that platform and then
  • 00:06:26
    apply a writing style and I we were
  • 00:06:29
    going back forth into your point I was
  • 00:06:30
    trying to do this all within like a
  • 00:06:31
    singular prompt and actually that was a
  • 00:06:34
    bunch of different steps that you would
  • 00:06:35
    need llm to do different things for and
  • 00:06:37
    so then I built you give me access to
  • 00:06:39
    this and what I loved about this was
  • 00:06:41
    just how quickly you can stack rank
  • 00:06:43
    things on top of each other like
  • 00:06:44
    basically add in additional AI assistant
  • 00:06:47
    to do one part of the the puzzle and so
  • 00:06:49
    what this does is basically this is
  • 00:06:51
    specifically for LinkedIn and so it can
  • 00:06:53
    take a a YouTube URL and it can
  • 00:06:56
    basically create it just gets the
  • 00:06:58
    transcript this here is like what kind
  • 00:07:00
    of LinkedIn post do you want to create
  • 00:07:01
    so it when I really look at LinkedIn I
  • 00:07:04
    had kind of honed in on three things
  • 00:07:06
    that really worked on LinkedIn and I
  • 00:07:07
    think you have to have some ability and
  • 00:07:10
    some experience to know this right it's
  • 00:07:13
    not it's not easy to know this unless
  • 00:07:14
    you know how to create content for that
  • 00:07:15
    platform but for me it was like if you
  • 00:07:17
    could create an educational post which
  • 00:07:19
    is like a singular lesson where that
  • 00:07:20
    person learns something and in the
  • 00:07:21
    background we'll get into the background
  • 00:07:22
    but that's I taught the llm what these
  • 00:07:25
    things are a spicy take which is
  • 00:07:27
    unfortunately unfortunately like how the
  • 00:07:29
    internet works today which is just like
  • 00:07:31
    say spicy things get a bunch of like
  • 00:07:33
    controversial opinions and a head not a
  • 00:07:35
    Hadnot the way I taught the the LM aeot
  • 00:07:38
    is is like everyone agrees to this but
  • 00:07:40
    no one's really been able to articulate
  • 00:07:41
    it in a really Punchy way and so
  • 00:07:43
    everyone like oh like you've said the
  • 00:07:45
    thing I wanted to say for a long time
  • 00:07:47
    and so that's the kind of posts that
  • 00:07:49
    work in short fir content then these are
  • 00:07:50
    the different post Styles so then that's
  • 00:07:53
    another part of that's another AI
  • 00:07:55
    request that actually says hey these are
  • 00:07:57
    the kind of formats that work really
  • 00:07:59
    well on LinkedIn everyone were everyone
  • 00:08:01
    will know the first one right the
  • 00:08:02
    Personal Achievement story that has been
  • 00:08:05
    done to death on LinkedIn so then you
  • 00:08:06
    can kind of pick one of these I picked I
  • 00:08:08
    picked like the spicy take I picked the
  • 00:08:11
    hot industry take this is done from a
  • 00:08:13
    great podcast we had with Alex liberman
  • 00:08:15
    where we actually did say some spicy
  • 00:08:16
    things and then you have different
  • 00:08:18
    writing styles so this is basically
  • 00:08:20
    created from different creators that I
  • 00:08:23
    think fit these Styles and then me
  • 00:08:24
    editing the style guide in the
  • 00:08:26
    background and again teaching the LM how
  • 00:08:27
    to do this and so you can have different
  • 00:08:29
    style you different writing styles I
  • 00:08:31
    picked a direct to No Nonsense because
  • 00:08:32
    anyone who's worked with me knows that's
  • 00:08:33
    kind of who who I am better or
  • 00:08:37
    worse and then it creates an incredible
  • 00:08:41
    first draft and I do want to emphasize
  • 00:08:43
    first draft I still think the skill set
  • 00:08:45
    and uh there's some great uh quotes
  • 00:08:47
    around like the power of writing really
  • 00:08:49
    is in the editing and so you still have
  • 00:08:51
    to bring the Nuance to it in terms of
  • 00:08:53
    how you edit it you have to bring the
  • 00:08:54
    magic your personality in terms of how
  • 00:08:56
    you edit but this is a really good start
  • 00:08:59
    right it creates a spicy take that you
  • 00:09:02
    should stop obsessing over your personal
  • 00:09:03
    brand on social media and basically just
  • 00:09:05
    ship content right because the more you
  • 00:09:07
    ship the more you learn talks about
  • 00:09:09
    consistency and volume has a quote from
  • 00:09:11
    Alex which is really good like actually
  • 00:09:13
    pulled a quote from the YouTube the
  • 00:09:15
    YouTube podcast has a nice thing here no
  • 00:09:17
    one gives a [ __ ] about your ego knows
  • 00:09:19
    how to create Punchy lines for LinkedIn
  • 00:09:22
    and basically the only thing I need to
  • 00:09:23
    edit it is I I need to edit this side of
  • 00:09:25
    it which is like it always ends with
  • 00:09:27
    like how do you get engagement LinkedIn
  • 00:09:28
    and I don't think that works more so if
  • 00:09:31
    I was like if I'm doing LinkedIn and I
  • 00:09:34
    don't know what to do I can basically
  • 00:09:36
    take any YouTube video that I think is
  • 00:09:37
    cool and basically create first drafts
  • 00:09:39
    and any and I could just continue to
  • 00:09:41
    like click the button and change I could
  • 00:09:42
    say okay now I want a head nod now I
  • 00:09:45
    want a industry Insight Revelation and
  • 00:09:47
    now I want a different writing style and
  • 00:09:48
    it would create a first draft and I
  • 00:09:50
    guess like to me this is one of the
  • 00:09:54
    powers of Agents plus AI that it can
  • 00:09:57
    truly unlock your creativity and you can
  • 00:09:59
    just iterate so so much faster yeah yeah
  • 00:10:03
    that it's and this is an amazing agent
  • 00:10:05
    by the way uh I I used it myself
  • 00:10:06
    confession I used it for a LinkedIn post
  • 00:10:09
    I think last week I got 500 likes on
  • 00:10:11
    that LinkedIn post and so it was good
  • 00:10:14
    right right in that particular case just
  • 00:10:15
    to prove a point not the kids don't try
  • 00:10:18
    this at home I did not change a single
  • 00:10:20
    character from the output of your agent
  • 00:10:22
    take that content I'm just going to post
  • 00:10:25
    it just to see what happens and and
  • 00:10:27
    darbes what what's your average like if
  • 00:10:29
    what would the average post get if that
  • 00:10:30
    got 500 my average
  • 00:10:32
    on probably between 200 and 250 give or
  • 00:10:36
    take um so above average it's above
  • 00:10:38
    average probably top deile as far as my
  • 00:10:40
    post yeah yeah and so I think the part
  • 00:10:43
    of it and this is something we should
  • 00:10:44
    talk about is that it's it's those
  • 00:10:46
    pieces coming together it's like oh I
  • 00:10:47
    know I you know I watch YouTube videos
  • 00:10:49
    all the time U as do as do you folks um
  • 00:10:52
    it's like oh I watch that YouTube video
  • 00:10:54
    and then something's kind of stuck in my
  • 00:10:55
    head it's like oh that piece was good
  • 00:10:57
    let's just say it's nice to be able to
  • 00:10:58
    take that YouTube video for it to
  • 00:11:00
    generate the transcri all these things I
  • 00:11:01
    can do manually right it's like then I
  • 00:11:02
    have to kind of think through it's like
  • 00:11:04
    okay what was that piece and for it to
  • 00:11:05
    extract those kind of spicy takes or had
  • 00:11:07
    nodding moments or whatever uh and then
  • 00:11:09
    give me a list of like oh here the ones
  • 00:11:10
    you could use pick one and then write
  • 00:11:12
    the LinkedIn post is just assembling all
  • 00:11:15
    those pieces together is I think what
  • 00:11:17
    what makes it useful so the thing I'm
  • 00:11:18
    working on K this will make you happy uh
  • 00:11:21
    the next two steps over the next we'll
  • 00:11:23
    say it should be available within the
  • 00:11:24
    next week or two is you'll be able to
  • 00:11:26
    post it right within uh within the agent
  • 00:11:28
    right like oh I'm going to review
  • 00:11:30
    aop or whatever but then here's where it
  • 00:11:32
    starts to get just more magical since
  • 00:11:34
    it's being posted from LinkedIn the
  • 00:11:36
    agent will know about it and then he can
  • 00:11:37
    track it on your behalf and give you a
  • 00:11:40
    report 24 hours later and say hey you
  • 00:11:42
    got this many lights or whatever here
  • 00:11:44
    here's how that compares to other ones
  • 00:11:45
    you've done and do spicy takes do better
  • 00:11:47
    than head nod takes or not like you know
  • 00:11:49
    we have a thesis so it can actually do
  • 00:11:51
    the monitoring for you not just the kind
  • 00:11:53
    of editing and curation and posting and
  • 00:11:55
    do the the post activity tracking if you
  • 00:11:57
    will hey everyone quick stat for for you
  • 00:11:59
    AI usage has increased 53% since last
  • 00:12:03
    year and you know how I know that cuz I
  • 00:12:05
    just read our brand new 2024 AI Trends
  • 00:12:08
    in marketing report this report is
  • 00:12:11
    packed with original data that data not
  • 00:12:13
    only highlights the latest in AI Trends
  • 00:12:16
    but also shows how businesses today are
  • 00:12:18
    using AI to grow their business you'll
  • 00:12:20
    learn the top five AI tools for
  • 00:12:23
    marketers exactly how teams are
  • 00:12:25
    investing in Ai and where they're seeing
  • 00:12:28
    the return on their investment
  • 00:12:29
    and you get a datab backed answer to the
  • 00:12:32
    big question is AI helping or hurting my
  • 00:12:36
    marketing career I invite you to steal
  • 00:12:38
    and take all of our hours of research
  • 00:12:41
    for free grab your copy from the link in
  • 00:12:43
    the description below now back to
  • 00:12:45
    Today's Show one confession here I'm
  • 00:12:47
    actually interested to hear how you
  • 00:12:48
    think about this throughout my entire
  • 00:12:50
    career I've wrote like I like you know
  • 00:12:52
    wrote put post anything I learn I just
  • 00:12:54
    put out there because you know that's
  • 00:12:55
    how you grow an audience I would say
  • 00:12:57
    this is the first time I don't want to
  • 00:12:59
    share this right like I feel like and
  • 00:13:02
    and I and I wonder about that because
  • 00:13:04
    there's two things that when I built
  • 00:13:06
    this was a real aha first of all it
  • 00:13:09
    makes the better people better because I
  • 00:13:11
    have the ability to like really
  • 00:13:13
    understand how to put something like
  • 00:13:14
    this together because I understand how
  • 00:13:16
    to create good content for LinkedIn but
  • 00:13:18
    then I'm
  • 00:13:19
    like if I just give it away everyone can
  • 00:13:22
    kind of do it right and and it's going
  • 00:13:23
    to be given away because I'm not the
  • 00:13:24
    only one who can build something like
  • 00:13:25
    this people are going to be building
  • 00:13:26
    these tools and giving it away and I'm
  • 00:13:28
    like and I'm like like okay well if
  • 00:13:29
    everyone has this I was trying to figure
  • 00:13:31
    out like where is my leverage then yeah
  • 00:13:34
    what do I do differently Shing yeah like
  • 00:13:37
    I'm I'm really crushing it right now but
  • 00:13:39
    I don't know how long that that lasts
  • 00:13:40
    for yeah okay so I you crushing it was
  • 00:13:45
    just so funny to me um okay so here's
  • 00:13:49
    couple thoughts on that one is you you
  • 00:13:50
    get to control so if you build an agent
  • 00:13:52
    on agent. there's nothing that says you
  • 00:13:55
    have to share that with anyone and you
  • 00:13:57
    can just use it for yourself right you
  • 00:13:58
    can just use it as personal productivity
  • 00:14:00
    uh self- automation tool Next Step Up is
  • 00:14:03
    you can say hey I want this available
  • 00:14:05
    but only to these six people right you
  • 00:14:07
    can say I'm only going to share the link
  • 00:14:08
    with his maybe the people on your team
  • 00:14:10
    or something like that because you want
  • 00:14:10
    to give them a productivity boost that
  • 00:14:12
    are on your team that's fine too like no
  • 00:14:14
    judgment here if you've come across to
  • 00:14:15
    something uh you know super amazing uh
  • 00:14:18
    the other the other thing to recall like
  • 00:14:20
    right now the way people share kind of
  • 00:14:22
    AI uh GPT kind of tidbits are sharing
  • 00:14:24
    props where you sort of have to give up
  • 00:14:26
    the source code for the entire thing in
  • 00:14:28
    order for people to be able to use it
  • 00:14:29
    because that's the mechanism by which
  • 00:14:31
    you share something in this case like
  • 00:14:33
    you can share the agent but you're not
  • 00:14:35
    sharing the underlying TRS fure out what
  • 00:14:37
    a spicy take is or figure out what the
  • 00:14:38
    post styles are how it's actually doing
  • 00:14:40
    the writing or you're picking the
  • 00:14:42
    genetic code of the writing style all of
  • 00:14:43
    that is still kind of cure and Secret
  • 00:14:45
    Sauce U and you can put it out there you
  • 00:14:47
    could have multiple levels of agents to
  • 00:14:48
    say hey I'm going to give this one away
  • 00:14:50
    for free this one I'm going to give to
  • 00:14:52
    close friends and family whatever it's
  • 00:14:53
    just a more powerful version of this
  • 00:14:55
    kind of thing that I'm putting out on
  • 00:14:56
    the market right now so it's unclear as
  • 00:14:58
    to what the
  • 00:14:59
    uh yeah I I have two quick thoughts um
  • 00:15:02
    number one I just want to say to
  • 00:15:04
    everyone that I teased this episode to
  • 00:15:05
    say I was going to give away this agent
  • 00:15:07
    and you've convinced me I don't I didn't
  • 00:15:08
    mean to this that I'm not going to give
  • 00:15:09
    it away I didn't think about just giving
  • 00:15:11
    it for myself you're you're not putting
  • 00:15:12
    the link I'm going to give no no not yet
  • 00:15:16
    I'm not going to give it away yet I
  • 00:15:17
    don't know what I'm going to do I
  • 00:15:18
    actually have to really think about this
  • 00:15:19
    it's the first time I have I'm just like
  • 00:15:21
    why would I give this away I do like
  • 00:15:22
    building like build a private team for
  • 00:15:24
    yourself and then actually give some of
  • 00:15:26
    it away but wouldn't wouldn't the thing
  • 00:15:28
    you would want to do are you gon to add
  • 00:15:29
    payment links because I'm a capitalist
  • 00:15:32
    at the end of the day so maybe I will
  • 00:15:33
    give away if you pay
  • 00:15:35
    me so that's that's the vision so if we
  • 00:15:37
    we'll take a step back uh once again so
  • 00:15:40
    agent. the way it's positioned and the
  • 00:15:43
    way I've kind of dreamed the big dream
  • 00:15:44
    is that it's a professional Network for
  • 00:15:46
    agents that's the tagline that's what it
  • 00:15:49
    is and so then the immediate question
  • 00:15:50
    would be like why do agents need a
  • 00:15:52
    professional Network um and the answer
  • 00:15:54
    is that with a fullness of time what's
  • 00:15:56
    going to happen I think where the way
  • 00:15:58
    these things are going to shap pop is
  • 00:15:59
    you can think of a an agent as a digital
  • 00:16:01
    coworker an intern you might hire right
  • 00:16:02
    and they're going to be good at discreet
  • 00:16:04
    kinds of tasks like like the one we just
  • 00:16:06
    talked about it's like oh I want you to
  • 00:16:07
    do this kind of LinkedIn post on on my
  • 00:16:09
    behalf uh but in the future we're going
  • 00:16:11
    to have hundreds thousands of these
  • 00:16:13
    agents out there and so how do you find
  • 00:16:15
    them how do you know what their
  • 00:16:16
    experiences are what are the ratings and
  • 00:16:17
    reviews all the things that's like right
  • 00:16:19
    now we sort of figure out when we're
  • 00:16:21
    hiring uh kind of human Freelancers we
  • 00:16:22
    go to Fiverr we go to upwork and they
  • 00:16:24
    have reading review there's a discovery
  • 00:16:25
    mechanism there's all that I want to
  • 00:16:27
    recreate that for the world of Agents so
  • 00:16:30
    the the way you kind of when you use
  • 00:16:32
    agent. there there's an agent like a
  • 00:16:35
    directory every agent has a Professional
  • 00:16:37
    Profile agents can actually do social
  • 00:16:39
    posts where they do like updates to the
  • 00:16:41
    the software or something like that so
  • 00:16:42
    they can kind of communicate with their
  • 00:16:43
    audience and someday agents are going to
  • 00:16:46
    have a salary that says hey this agent
  • 00:16:49
    costs $2 a month 99 a month whatever the
  • 00:16:52
    you know whatever the price tag happens
  • 00:16:54
    to be and so what you will do is you
  • 00:16:56
    will basically say I want to assemble my
  • 00:16:58
    digital team and my digital theme
  • 00:17:00
    consists of an agent that that's
  • 00:17:01
    kieran's agent that does this really
  • 00:17:03
    wicked cool YouTube to LinkedIn post
  • 00:17:05
    thing then I'm going to do this other
  • 00:17:06
    thing from D Mas this other thing from
  • 00:17:07
    someone else and pull it all together
  • 00:17:09
    and that's my digital team and the the
  • 00:17:11
    really I'm sorry I get overly excited
  • 00:17:13
    about this stuff the really really
  • 00:17:15
    exciting thing about these agents is
  • 00:17:17
    that they can actually use each other
  • 00:17:20
    all right so let's say I'll give you an
  • 00:17:21
    example this is a very very concrete
  • 00:17:23
    example so I am very excited about all
  • 00:17:25
    the image generation tools most of them
  • 00:17:27
    have been very kind of consumer Centric
  • 00:17:29
    uh creating stock photography images and
  • 00:17:31
    super high quality right they're really
  • 00:17:32
    really good but if you're using them in
  • 00:17:33
    a business context often what you need
  • 00:17:36
    is you need them in a certain style okay
  • 00:17:38
    you can do that you can put that in the
  • 00:17:39
    prompt you need them with a certain
  • 00:17:40
    color palette often right because I want
  • 00:17:42
    to create this consistency for let's say
  • 00:17:44
    my a series of block close or whatever
  • 00:17:46
    that I'm doing and so I built an agent
  • 00:17:48
    uh this weekend called a color palette
  • 00:17:51
    extractor very very simple agent all you
  • 00:17:53
    do is you give it the homepage or any
  • 00:17:55
    web page it will take a screenshot of
  • 00:17:57
    that web page it'll figure out which
  • 00:17:59
    colors are used to give you a palette
  • 00:18:01
    that is proportional based on what
  • 00:18:02
    percentage of the real estate that color
  • 00:18:04
    took so the bigger S that color swatch
  • 00:18:06
    right and it'll give it to you in Json
  • 00:18:08
    or it'll give it to you U as a color
  • 00:18:09
    palette you can take that kind of data
  • 00:18:11
    of the color palette that that agent
  • 00:18:13
    produced and then pass it to another
  • 00:18:15
    agent to say hey I'm going to uh
  • 00:18:16
    generate images with flux. one or uh
  • 00:18:19
    idiogram 2.0 we just launched last week
  • 00:18:22
    and here's a color palette I want you to
  • 00:18:23
    use that came from another agent right
  • 00:18:25
    so that's the composing two agents that
  • 00:18:27
    didn't know anything about each other
  • 00:18:28
    right until the fact that you glue them
  • 00:18:30
    together in a in a creative way it's
  • 00:18:32
    like oh the color palette now feeds into
  • 00:18:34
    the the image generator the image
  • 00:18:35
    generator that generates image May then
  • 00:18:38
    kind of go into a LinkedIn post
  • 00:18:39
    generator that says oh in addition to
  • 00:18:41
    this thing whatever I want a super cool
  • 00:18:43
    thing that takes the headline and uses
  • 00:18:45
    the ideogram 2.0 text capabilities that
  • 00:18:47
    are really brilliant and create this
  • 00:18:49
    really really nice LinkedIn post that
  • 00:18:51
    has an actual image that's topical
  • 00:18:52
    that's in my color palette in my scheme
  • 00:18:54
    in my style pull all of that together so
  • 00:18:56
    what's exciting is this this
  • 00:18:58
    composability
  • 00:18:59
    of Agents but then so back to answering
  • 00:19:01
    your original
  • 00:19:02
    question the hope is to make this kind
  • 00:19:04
    of very inefficient Market efficient
  • 00:19:06
    right so I want to create incentives for
  • 00:19:07
    developers and people like you to say I
  • 00:19:10
    can build this agent I can put it out
  • 00:19:11
    there I'm a company of one but it's
  • 00:19:13
    useful I've got some domain expertise I
  • 00:19:15
    know how agent. works and I've got these
  • 00:19:16
    five or 10 agents out there that people
  • 00:19:18
    can hire just like you would uh put your
  • 00:19:20
    single on the Fiverr thing that says oh
  • 00:19:22
    I can do these little tasks you're going
  • 00:19:23
    to build agents that can use little
  • 00:19:25
    tasks and let people hire them and then
  • 00:19:27
    rate and review and track experience and
  • 00:19:29
    all that so it really does make your
  • 00:19:31
    domain expertise much more profitable
  • 00:19:35
    because they're so much easier to create
  • 00:19:36
    so so there's two things I want I want
  • 00:19:37
    to follow up on there one just resetting
  • 00:19:39
    for everybody that really agents. a or
  • 00:19:42
    any agent network is going to be like a
  • 00:19:45
    Fiverr but for AI and AI agents more
  • 00:19:49
    instead of humans and it'll be kind of
  • 00:19:50
    the automation agent equivalent the
  • 00:19:53
    second thing kieren I think you're just
  • 00:19:55
    wrong if you just listened to everything
  • 00:19:57
    that Dar mesh said wouldn't you want to
  • 00:19:59
    be the dominant early content agent that
  • 00:20:02
    all the other agents started to adopt so
  • 00:20:05
    that as it becomes monetized your
  • 00:20:07
    amazing content agent that you're not
  • 00:20:09
    sharing with everybody would be like one
  • 00:20:11
    of the default standards and you could
  • 00:20:13
    make more and more money long term
  • 00:20:16
    versus the indirect benefit of just
  • 00:20:18
    getting better LinkedIn oppressions no
  • 00:20:20
    but how do I make money if I give it
  • 00:20:22
    away and don't and like give it away for
  • 00:20:24
    free and everyone can just because one
  • 00:20:26
    of the things I just want to actually
  • 00:20:27
    just before you answer that do want to
  • 00:20:29
    quickly show something because it will
  • 00:20:30
    it it it is important for the users to
  • 00:20:33
    see how these are built and then I can
  • 00:20:35
    answer this because the thing the thing
  • 00:20:37
    I want to the thing dares said which is
  • 00:20:39
    like really important to answer that
  • 00:20:41
    question is and I had thought about is
  • 00:20:44
    in the back end this is it's it's kind
  • 00:20:47
    of like codin right if you're actually
  • 00:20:50
    the thing I would actually uh be
  • 00:20:52
    interested in you're take Dar mes is
  • 00:20:53
    you're an incredible coder like you've
  • 00:20:55
    built companies I triy I uh graduated
  • 00:20:58
    University with with a college computer
  • 00:20:59
    science degree and I was a terrible
  • 00:21:01
    coder it's why I ended up in marketing
  • 00:21:02
    no offense to marketers I'm a marketer I
  • 00:21:04
    love being a marketer but like I tried
  • 00:21:05
    to be a coder building with llms is the
  • 00:21:08
    first time I thought I felt like a
  • 00:21:10
    competent coder so I was sitting here
  • 00:21:11
    like the way I'd always envisioned my
  • 00:21:13
    life like late at night uh building all
  • 00:21:16
    of these kind of like different little
  • 00:21:18
    what I think is pieces of codes but what
  • 00:21:19
    they really are is like different um you
  • 00:21:22
    know AI assistants or a part of the AI
  • 00:21:24
    agent and do you when so when I'm
  • 00:21:27
    building this this is basically the the
  • 00:21:29
    different requests stack ranked on each
  • 00:21:30
    other so this actually makes up the
  • 00:21:32
    agent that we shown right so it has like
  • 00:21:33
    the different calls that make up the
  • 00:21:35
    agent that we that we had shown but from
  • 00:21:37
    what you're seeing today and how you've
  • 00:21:39
    thought about building that agent the
  • 00:21:40
    color pallet agent you had thought of do
  • 00:21:43
    you believe that the average like do you
  • 00:21:44
    think we're going towards that direction
  • 00:21:46
    that the average consum or the average
  • 00:21:48
    person like me can actually code and
  • 00:21:51
    what do you think about that for all of
  • 00:21:52
    the people who have kind of learned that
  • 00:21:54
    as a as a learned skill and or have true
  • 00:21:57
    expertise in that yeah it's a great
  • 00:21:59
    question so um part of what excites me
  • 00:22:01
    about Ai and then agents particularly is
  • 00:22:04
    that it's going to democra so you you
  • 00:22:06
    may not be a coder in the classic sense
  • 00:22:08
    Karen but you are a builder right right
  • 00:22:10
    and the issue with coding right now it's
  • 00:22:12
    a little bit like you have these really
  • 00:22:13
    good thoughts in your head let's say and
  • 00:22:14
    you're like traveling to a foreign
  • 00:22:15
    country so it's not like you can't think
  • 00:22:17
    thoughts and don't have Frameworks
  • 00:22:18
    whatever you just don't express yourself
  • 00:22:20
    in the language that they happen to know
  • 00:22:22
    and that's the issue that most humans
  • 00:22:23
    have with coding it's like it's not that
  • 00:22:25
    you don't understand the problem it's
  • 00:22:26
    that even that you couldn't even write
  • 00:22:27
    out the recipe to solve Sol that problem
  • 00:22:29
    if you were teaching it to another human
  • 00:22:30
    or whatever we do this all the time when
  • 00:22:32
    we kind of bring people onto the team we
  • 00:22:33
    kind of show them here's how you do this
  • 00:22:35
    what agents are going to unlock is the
  • 00:22:37
    ability because of AI and it's kind of
  • 00:22:39
    natural language uh capabilities as well
  • 00:22:41
    is it's going to allow people to express
  • 00:22:44
    themselves and solve problems in not a
  • 00:22:46
    coding language but a human language or
  • 00:22:48
    at least a human approachable language
  • 00:22:50
    and this is a the kind of low code no
  • 00:22:52
    code thing that we've been talking about
  • 00:22:53
    for years but now you can actually do
  • 00:22:55
    real things like for instance the the
  • 00:22:57
    color palette example that I gave you
  • 00:22:58
    all built right inside the agent Builder
  • 00:23:00
    right I did not write any python code to
  • 00:23:02
    do that and the reason I could do that
  • 00:23:04
    is quad you 3.5 uh Sonic lets you
  • 00:23:06
    actually generate really really good
  • 00:23:08
    code so all I had to do is describe the
  • 00:23:10
    thing that I wanted so the first pass at
  • 00:23:11
    it was go look at this website and give
  • 00:23:13
    me a list of all the colors okay it can
  • 00:23:16
    do that and then I like well you know it
  • 00:23:18
    would be more useful if I could get like
  • 00:23:20
    a sense for which are the so that it's
  • 00:23:21
    like prioritize the list with the most
  • 00:23:23
    common color first so I know that orange
  • 00:23:25
    is the most popular color on the upot
  • 00:23:27
    website okay okay that's useful but I
  • 00:23:29
    wonder if we could actually like create
  • 00:23:31
    a color swatch where each Square in the
  • 00:23:32
    color swatch is proportional to the
  • 00:23:34
    percentage of real estate that color
  • 00:23:35
    takes all of that completely no coding
  • 00:23:38
    and literally in three iterations uh you
  • 00:23:40
    know with CLA it's like okay now I have
  • 00:23:42
    something useful it's like okay well now
  • 00:23:43
    I have it in a color swatch give it to
  • 00:23:45
    me in Json because now I can reuse it in
  • 00:23:47
    other agents that I kind of want to
  • 00:23:48
    compose on so the answer is I think now
  • 00:23:51
    for the first time in reality and you've
  • 00:23:53
    seen this like this agent exists and I
  • 00:23:55
    know I'm going to talk you into sharing
  • 00:23:58
    it just like is we're going to put
  • 00:23:59
    pressure on you I'm a hard P I'm not
  • 00:24:01
    going to take anything from what you
  • 00:24:02
    have in there this
  • 00:24:04
    is and I'm going to kick your
  • 00:24:07
    butt yeah this is actually the point I
  • 00:24:10
    was trying G going to make to Kip
  • 00:24:11
    actually is the reason I brought that up
  • 00:24:13
    and the reason I asked that question is
  • 00:24:14
    because I actually realized what what
  • 00:24:16
    you had said Dar mas and Kip your
  • 00:24:18
    question is as long as I don't it's like
  • 00:24:19
    do I open source this part or not yeah
  • 00:24:22
    this because this is where the magic is
  • 00:24:23
    it's like the tool is the output but
  • 00:24:25
    this is where the magic is and I guess
  • 00:24:27
    the thing is like
  • 00:24:28
    like in the world we're going to live in
  • 00:24:31
    it still feels that you can represent
  • 00:24:33
    your domain expertise in this kind of
  • 00:24:35
    code here which is the AI natural
  • 00:24:37
    language kind of code and you would want
  • 00:24:39
    to be thoughtful about open source in
  • 00:24:41
    that because like there's just going to
  • 00:24:42
    be core demain expertise that you have
  • 00:24:45
    that you do want to keep to yourself in
  • 00:24:48
    some ways um and I think that that's the
  • 00:24:51
    part yeah I I that's the part I think is
  • 00:24:53
    like the where the value is yeah that's
  • 00:24:55
    and that's not where I'm going to push
  • 00:24:56
    you I'm going to push you for instance
  • 00:24:57
    you know has launched uh you know lotss
  • 00:25:00
    of free tools I've buil a lot of free
  • 00:25:01
    tools and we did AI grader um a couple
  • 00:25:04
    weeks ago that you guys talked about on
  • 00:25:05
    the Pod so think about it that way right
  • 00:25:08
    say I'm going to launch a free tool what
  • 00:25:09
    do I get for putting free tools out
  • 00:25:11
    there agent. is a professional Network
  • 00:25:13
    so Kieran is going to have a profile
  • 00:25:14
    your agent's going to have a profile so
  • 00:25:16
    imagine your profile gets a bunch of
  • 00:25:18
    followers and a bunch of users that's an
  • 00:25:20
    audience that you're now building you're
  • 00:25:22
    building this kind of thought leadership
  • 00:25:24
    and do leadership um in terms of you've
  • 00:25:26
    actually done something that's useful
  • 00:25:27
    but doesn't mean you have to we're not
  • 00:25:28
    giving away the code for AI grader we're
  • 00:25:30
    not even telling you what's going into
  • 00:25:31
    the algorithm figure it out we're just
  • 00:25:33
    making something that other people can
  • 00:25:34
    use right that's the idea um so and it's
  • 00:25:38
    going to vary from agent to agent you
  • 00:25:39
    might have some things that's like okay
  • 00:25:41
    well I just want to use this for myself
  • 00:25:42
    some things that you say I want to share
  • 00:25:43
    with the world but I'm not going to
  • 00:25:44
    release the code that actually makes it
  • 00:25:46
    happen and you might actually put some
  • 00:25:48
    example apps out there or out there
  • 00:25:50
    here's a minimal version of my original
  • 00:25:51
    one that doesn't actually use the prompt
  • 00:25:53
    that I use but just so you have a sense
  • 00:25:54
    for how these things work and then you
  • 00:25:56
    can use that as your Baseline versus the
  • 00:25:58
    I think I think that is actually the
  • 00:26:00
    answer to our mesh is that you will have
  • 00:26:01
    two versions of this agent Kieran one
  • 00:26:04
    that you use and one that you let
  • 00:26:05
    everybody use to get adoption get
  • 00:26:07
    reviews get maybe make some money and
  • 00:26:11
    the difference behind them is
  • 00:26:12
    essentially the code which in this case
  • 00:26:14
    is natural language English in your case
  • 00:26:16
    prompts right and you've got very
  • 00:26:18
    specific prompts as part of that agent
  • 00:26:20
    and you're going to make you're going to
  • 00:26:22
    keep making those better and better and
  • 00:26:23
    the question is how much do you pass
  • 00:26:25
    those improvements on to everybody
  • 00:26:27
    versus just like keeping them for the
  • 00:26:29
    core one that you're using right yeah by
  • 00:26:31
    the way I'm I'm a warm-hearted
  • 00:26:33
    red-blooded capitalist curing so and my
  • 00:26:37
    my passion is really around making
  • 00:26:39
    inefficient markets efficient right so
  • 00:26:42
    not to get uh too economically I'm not
  • 00:26:44
    an economics person but um this is It's
  • 00:26:47
    a good diversion because I think it kind
  • 00:26:49
    of identifies opportunities for people
  • 00:26:50
    um so the definition of an e efficient
  • 00:26:53
    market uh in economic theory is that
  • 00:26:55
    when all possible transaction that could
  • 00:26:57
    occur in a Market actually do occur
  • 00:27:00
    that's the definition of an efficient
  • 00:27:01
    market uh most markets are not efficient
  • 00:27:04
    right um and there are multiple reasons
  • 00:27:06
    why transactions that should be
  • 00:27:08
    occurring don't occur one is buyer and
  • 00:27:10
    seller don't even know about each other
  • 00:27:11
    it's a discovery problem buyer and
  • 00:27:13
    seller the buyer doesn't trust the
  • 00:27:14
    seller there's no there's no trust
  • 00:27:15
    mechanism and so the transaction doesn't
  • 00:27:17
    happen even if they know about each
  • 00:27:18
    other uh the buyer and seller have no
  • 00:27:20
    way to establish a fair price that's
  • 00:27:22
    another reason uh transactions that
  • 00:27:23
    should be occurring because you can't
  • 00:27:24
    find a clearing price with the
  • 00:27:26
    transaction should occur so if you look
  • 00:27:27
    at some of the most successful
  • 00:27:28
    businesses we've had in the last few
  • 00:27:30
    decades at varying levels they've taken
  • 00:27:32
    inefficient markets and made them
  • 00:27:34
    efficient classic example is eBay right
  • 00:27:36
    made an efficient market for all these
  • 00:27:37
    niche market Goods like I've got Pez
  • 00:27:39
    dispensers or whatever I'm going to
  • 00:27:40
    Collectibles I'm going to sell them on
  • 00:27:42
    eBay that makes sense Google in this
  • 00:27:44
    early manifestation was a efficient
  • 00:27:46
    market model that says I have got this
  • 00:27:48
    niche market content over here and these
  • 00:27:50
    14 people on the planet that want to
  • 00:27:52
    read that content and Google made that
  • 00:27:54
    market more efficient by connecting and
  • 00:27:56
    making that Discovery possible uh what
  • 00:27:58
    asian. is about is taking folks like you
  • 00:28:01
    that have domain expertise that can
  • 00:28:02
    solve concrete problems and then there's
  • 00:28:05
    a list of people that are looking for
  • 00:28:07
    those problems to be solved and helping
  • 00:28:09
    connect the dots right yeah in order for
  • 00:28:11
    the market to truly be efficient though
  • 00:28:13
    you have to be able to have enough
  • 00:28:14
    incentive you Karen to say here's what
  • 00:28:16
    I'm getting out of making this
  • 00:28:18
    investment of my time energy or whatever
  • 00:28:19
    and maybe uh like you do with L the
  • 00:28:21
    reason you post a LinkedIn is you're
  • 00:28:23
    building authority on LinkedIn right you
  • 00:28:24
    could say well why would I post any of
  • 00:28:26
    this really award when I why would I
  • 00:28:27
    generate million impressions for
  • 00:28:29
    LinkedIn when I can just build up my own
  • 00:28:31
    website and and draw things there well
  • 00:28:32
    there's a side benefit to that because
  • 00:28:34
    you can use that audience for other
  • 00:28:35
    things same thing here right um but
  • 00:28:37
    maybe sometimes you need like a
  • 00:28:39
    financial intensive to be able to draw
  • 00:28:41
    the developers in and the builders in
  • 00:28:42
    let them set their price or maybe we
  • 00:28:44
    help them figure out the price uh here's
  • 00:28:45
    what the Market's willing to pay and
  • 00:28:47
    figure out all those details for you I
  • 00:28:49
    think the thing that you know caus me to
  • 00:28:51
    pause a little bit is like when I was on
  • 00:28:53
    paternity leave this year like a true
  • 00:28:55
    dork instead of you know taking my P to
  • 00:28:58
    leave and just getting offline I
  • 00:29:00
    released an AI course which is like the
  • 00:29:03
    dumbest thing and I apologize to my
  • 00:29:05
    partner uh my daughter but it was late
  • 00:29:08
    at night it was late at night when
  • 00:29:10
    everyone was in bed it was like not
  • 00:29:12
    disruptive to that time and um I was
  • 00:29:13
    again it was like it was like this
  • 00:29:16
    Timeless marketing course where you
  • 00:29:17
    could easily use these templates to
  • 00:29:19
    create Timeless marketing content and I
  • 00:29:21
    started getting all of these
  • 00:29:22
    notifications on LinkedIn of like hey uh
  • 00:29:25
    here's my everyone using those templates
  • 00:29:27
    to do LinkedIn post and I was like oh my
  • 00:29:29
    God like if I continue to do this it's
  • 00:29:30
    just this whole Army of people kind of
  • 00:29:33
    it perfectly kind of it doesn't
  • 00:29:34
    perfectly replicate my style because of
  • 00:29:36
    like editing and stuff but like it gets
  • 00:29:38
    you a first draft of my style so there's
  • 00:29:41
    that part is like it adds a little bit
  • 00:29:43
    of hesitancy and there's a fall one from
  • 00:29:45
    that that I think you're the best person
  • 00:29:46
    to ask this question to is you've built
  • 00:29:49
    one of the most successful SAS companies
  • 00:29:50
    of all time you've built many companies
  • 00:29:52
    and HubSpot if you when it started was a
  • 00:29:55
    point solution like I think HubSpot was
  • 00:29:57
    like a blogging tool
  • 00:29:58
    if you start Hobs if you tried to start
  • 00:30:01
    that company today someone could have
  • 00:30:04
    could potentially just have built an
  • 00:30:05
    agent right because it's like so much
  • 00:30:07
    easier to build this kind of software do
  • 00:30:10
    you have what are your thoughts about
  • 00:30:12
    the you know you're deep in the agent
  • 00:30:13
    space you know a lot about the software
  • 00:30:15
    industry what are your thoughts about
  • 00:30:17
    the repercussions for the software
  • 00:30:18
    industry of the ability to kind of build
  • 00:30:20
    these agents so over yeah I've been in
  • 00:30:22
    software commercial software now for 30
  • 00:30:24
    years right so I've been that this for a
  • 00:30:25
    long time um and every time software
  • 00:30:29
    gets easier to build which we've had
  • 00:30:31
    happen right and AI is just the most
  • 00:30:33
    recent manifestation of it uh you know
  • 00:30:35
    having seen like I've been on like
  • 00:30:37
    character mode terminals working on
  • 00:30:38
    mainframes back in my kind of early
  • 00:30:40
    years every time that has happened it's
  • 00:30:42
    actually increased both the value of
  • 00:30:44
    software engineers and the value of
  • 00:30:46
    software companies and the reason that
  • 00:30:48
    happens is that yes it lowers the bar so
  • 00:30:50
    more people can get in but it also
  • 00:30:52
    raises the bar I mean well increases the
  • 00:30:55
    market in terms of the kinds of problems
  • 00:30:57
    that are now solvable that weren't
  • 00:30:58
    solvable before um so when you make
  • 00:31:00
    something easier it's like oh we're yes
  • 00:31:01
    we're going to have more developers
  • 00:31:02
    there's going to be more competition so
  • 00:31:04
    if I were starting HubSpot today I
  • 00:31:05
    obviously wouldn't do exactly what
  • 00:31:07
    HubSpot did 18 years ago but I would say
  • 00:31:10
    oh well like agent De AI was not
  • 00:31:13
    buildable uh even three years ago or
  • 00:31:15
    four years ago so I would do more things
  • 00:31:17
    like that uh I would say okay well like
  • 00:31:19
    even HubSpot you know has become a
  • 00:31:21
    platform now third parties can build on
  • 00:31:22
    our apis and build apps that integrate
  • 00:31:23
    with it like I would say oh well maybe I
  • 00:31:27
    get to the kind of point of doing a
  • 00:31:28
    platform even sooner because I
  • 00:31:30
    understand the mechanics of kind of
  • 00:31:31
    platform building those kinds of things
  • 00:31:32
    so my personal take is it's less so
  • 00:31:35
    being an engineer like um is really less
  • 00:31:38
    about the coding that's just like saying
  • 00:31:39
    I can put words together into sentences
  • 00:31:41
    in a certain language it's really about
  • 00:31:42
    thinking and it's less about coding and
  • 00:31:44
    that's what AI is bringing to the table
  • 00:31:46
    now is that if you can like think
  • 00:31:48
    coherent analytical thoughts and know a
  • 00:31:51
    particular customer do uh customer
  • 00:31:53
    domain and know the use cases you're
  • 00:31:55
    going to be able to solve those problems
  • 00:31:56
    and I think that's that's a great
  • 00:31:58
    um yeah so I I I think I couldn't agree
  • 00:32:00
    with you more on that dares I I still
  • 00:32:03
    think the scarce asset in the world is
  • 00:32:05
    not the technology Kieran I think it's
  • 00:32:07
    people who are obsessed about are
  • 00:32:09
    obsessed about problems and can make
  • 00:32:11
    rational like steps to solve those
  • 00:32:14
    problems versus kind of getting caught
  • 00:32:16
    up in the activity And the emotions and
  • 00:32:18
    everything and that we're we're still in
  • 00:32:20
    very short supply of just smart rational
  • 00:32:22
    humans applying logic and leveraging
  • 00:32:25
    technology and I'm hoping that AI is
  • 00:32:28
    just the next unlock for the software
  • 00:32:31
    Market regardless of what kind of
  • 00:32:33
    software you're building right and the
  • 00:32:35
    other one other thing I'll say just at
  • 00:32:36
    the macro level in terms of U kind of AI
  • 00:32:39
    and kind of label Force overall is that
  • 00:32:41
    most of the arguments against it that
  • 00:32:43
    says oh this is going to kind of disrupt
  • 00:32:44
    the label and there's going to be uh you
  • 00:32:46
    know some disruption for sure but um it
  • 00:32:49
    assumes kind of this uh fixed Pi zero
  • 00:32:51
    some game that says oh there's a finite
  • 00:32:53
    amount of work that needs to happen in
  • 00:32:54
    any organization the more work that AI
  • 00:32:56
    does the less work we're going to uh
  • 00:32:58
    need humans to do well that assumes this
  • 00:33:00
    fixed pie so let me give you an example
  • 00:33:02
    we just talked about engineers so we had
  • 00:33:03
    HubSpot have rolled out kind of AI tools
  • 00:33:06
    across 100% of the the product team now
  • 00:33:08
    over the course of the last uh of the
  • 00:33:09
    last year uh and our estimate is we're
  • 00:33:11
    getting somewhere around 15 to 20% of
  • 00:33:14
    measurable improvements uh to an
  • 00:33:16
    engineer's productivity all right so
  • 00:33:18
    does that mean that we're now all of a
  • 00:33:20
    sudden say oh we're going to let 20% of
  • 00:33:21
    our Engineers go because we can
  • 00:33:23
    accomplish the same work with 20% fewer
  • 00:33:24
    people no the answer is absolutely not
  • 00:33:27
    we have an infinite supply of like
  • 00:33:29
    product Vision product roadmap stuff or
  • 00:33:31
    whatever and actually it makes me able
  • 00:33:34
    to rationalize even more investment in
  • 00:33:36
    R&D and and hiring Engineers because now
  • 00:33:38
    they're more productive so the return on
  • 00:33:40
    each engineer went up and so now the the
  • 00:33:43
    bar is even lower as far as what I have
  • 00:33:45
    to do to rationalize H same thing in
  • 00:33:47
    sales if you have enough leaves and now
  • 00:33:49
    your sales paper all of a sudden 15 20
  • 00:33:51
    25% more productive let's let's hire
  • 00:33:54
    more sales people now right because
  • 00:33:55
    before here's how much you could
  • 00:33:57
    actually afford for a salesperson
  • 00:33:59
    because this is the return you got if
  • 00:34:00
    the return you get is higher or the cost
  • 00:34:02
    you have to spend is lower let's let's
  • 00:34:04
    bring on more not less I think it
  • 00:34:05
    creates abundance uh creates a bigger
  • 00:34:07
    pie in most cases not in all cases uh
  • 00:34:10
    but if you look at it through the right
  • 00:34:11
    lens even for something like customer
  • 00:34:13
    support right which is like oh we have
  • 00:34:15
    to manage 500 tickets and we're going to
  • 00:34:17
    kind of Auto solve them with AI which is
  • 00:34:19
    certainly going to happen over time but
  • 00:34:21
    what that means is those people that we
  • 00:34:22
    were using for lower uh complexity
  • 00:34:24
    ticket res resolution can now be
  • 00:34:27
    redirected to customer success to
  • 00:34:28
    revenue driving activities to kind of
  • 00:34:30
    higher order customer relationship
  • 00:34:32
    building activities versus these kind of
  • 00:34:34
    mundane kind of ticket resolution thing
  • 00:34:36
    that uh that software can do equally
  • 00:34:38
    well so yeah yeah and I definitely like
  • 00:34:41
    I do think the the vision is part of
  • 00:34:43
    that would you painted which is in the
  • 00:34:45
    future you'll have like teams and AI
  • 00:34:47
    agents that just part of that team and
  • 00:34:49
    you'll have a Works space for them and
  • 00:34:50
    they'll be able to do a certain tasks
  • 00:34:52
    that is what smart people are you know
  • 00:34:54
    the the early adopters or already doing
  • 00:34:57
    like we we talked to Ethan who came from
  • 00:34:58
    Asana he he was a bdr and he basically
  • 00:35:02
    used AI to Auto s automate out of a role
  • 00:35:05
    and then he basically was asked to do
  • 00:35:07
    that for the entire sales team like
  • 00:35:08
    apply the same sort of logic and when
  • 00:35:10
    you actually see what he did he just
  • 00:35:12
    created like assistant single assistants
  • 00:35:14
    to do single things with a guided prompt
  • 00:35:16
    to like solve individual tasks and it
  • 00:35:19
    does speak to the fact that if you're a
  • 00:35:21
    knowledge worker today there is a new
  • 00:35:23
    skill set to learn because that is not a
  • 00:35:26
    thing that comes naturally to a lot of
  • 00:35:27
    people I still see a lot of people
  • 00:35:29
    saying like oh well AI is hype but it
  • 00:35:31
    it's because that person is not a good
  • 00:35:33
    like AI manager right they're not able
  • 00:35:35
    to um figure out how to prompt the AI
  • 00:35:38
    correctly they're not able to figure out
  • 00:35:40
    how to set the right context of what
  • 00:35:42
    good good looks like they're not able to
  • 00:35:44
    on board the AI because the on boarding
  • 00:35:47
    of the AI assistant is really important
  • 00:35:48
    like you have to kind of set context of
  • 00:35:50
    what you want done and I don't know if
  • 00:35:51
    you have like tips for people on just H
  • 00:35:54
    how can they shift their thinking to be
  • 00:35:56
    much more how do I actually integrate AI
  • 00:35:59
    into my work yeah so I'll say a couple
  • 00:36:01
    things one is anytime a new technology
  • 00:36:05
    uh comes along there's this uh period of
  • 00:36:07
    time where there's this massive
  • 00:36:09
    Arbitrage so when if you were one of the
  • 00:36:11
    early ones that got on the internet if
  • 00:36:12
    you one of the early ones that figure
  • 00:36:13
    out out anything like pick a marketing
  • 00:36:15
    channel pick a technology doesn't really
  • 00:36:17
    matter um if you're one of the early
  • 00:36:19
    adopters and you learned it and you
  • 00:36:20
    applied it it was a massive Arbitrage
  • 00:36:22
    opportunity because as uh as common
  • 00:36:25
    place as we think these things are they
  • 00:36:26
    take years to kind of get absorbed into
  • 00:36:28
    the where the average person whatever
  • 00:36:30
    discipline or whatever role and so right
  • 00:36:32
    now we are at the precipice of that so
  • 00:36:34
    by the way this is not to Pander to your
  • 00:36:36
    audience uh but if you're watching this
  • 00:36:38
    podcast right now or listening to it
  • 00:36:40
    right now that means you have one of the
  • 00:36:42
    most underrated skills that exists uh in
  • 00:36:45
    business right now which is curiosity
  • 00:36:47
    right you could have been doing
  • 00:36:48
    something else with your time you could
  • 00:36:49
    have been watching something else on
  • 00:36:50
    YouTube or whatever you're watching this
  • 00:36:52
    because you have this Nate curiosity
  • 00:36:53
    about what's happening uh and my message
  • 00:36:55
    to you would be dig into that Curiosity
  • 00:36:59
    this happens to be um a profitable
  • 00:37:01
    curiosity to have uh this will advance
  • 00:37:03
    your career learn about AI use AI apply
  • 00:37:07
    AI help your team use AI help your
  • 00:37:09
    company use AI that will raise your
  • 00:37:11
    currency I guarantee you there are very
  • 00:37:13
    few guarantees of life that if you
  • 00:37:14
    actually kind of commit those calories
  • 00:37:16
    uh you will emerge a uh better person
  • 00:37:19
    for your company for yourself or even
  • 00:37:20
    your own individual currency um I think
  • 00:37:22
    it will be
  • 00:37:23
    helpful very few times has that happened
  • 00:37:25
    where you had a broadly applicable Tech
  • 00:37:27
    technology uh you know on the internet
  • 00:37:29
    really was the last time mobile was big
  • 00:37:31
    didn't apply as much to B2B you
  • 00:37:33
    everybody built an app but you know we
  • 00:37:35
    only use like eight apps or something
  • 00:37:36
    like that on most of our phones um but
  • 00:37:38
    this is this is big this is like the
  • 00:37:39
    internet was right it's uh and and all
  • 00:37:42
    of you if you're watching right now you
  • 00:37:43
    have the inside Advantage it's still
  • 00:37:45
    very very early days I'm still figuring
  • 00:37:47
    it out that's kind of what's possible
  • 00:37:48
    and trying to piece it together but yeah
  • 00:37:50
    yeah I think I think you're right darvan
  • 00:37:52
    that's one of the core takeways here
  • 00:37:53
    it's never been a more profitable time
  • 00:37:55
    to be curious especially about
  • 00:37:57
    artificial intelligence because like the
  • 00:37:59
    Social Web Web 2.0 all those things that
  • 00:38:01
    was kind of like a derivation of the
  • 00:38:04
    internet right like you had it's like
  • 00:38:05
    having the iPhone and then getting the
  • 00:38:07
    iPhone like 3G it it was it was a really
  • 00:38:10
    great Improvement but it was still very
  • 00:38:12
    similar this revolution of the internet
  • 00:38:15
    is going to be vastly different I mean
  • 00:38:17
    you have you have Kieran who I mean you
  • 00:38:20
    self self- profess dropped out of like
  • 00:38:22
    being a web programmer and now you're
  • 00:38:24
    like wait I love prompting I love
  • 00:38:26
    programming I'm building all these
  • 00:38:27
    things that I never thought I'd be able
  • 00:38:29
    to do and this is and this technology is
  • 00:38:31
    in like the very earliest of its stages
  • 00:38:34
    right I I think I think it's the best
  • 00:38:38
    time in human history to monetize good
  • 00:38:41
    ideas yes yes I think that is the time
  • 00:38:44
    we live in because if you have had good
  • 00:38:46
    ideas and you like I have had a lot of
  • 00:38:48
    good ideas I told Kip This and like you
  • 00:38:51
    know the typical person who had good
  • 00:38:52
    ideas and then see the business being
  • 00:38:55
    done that was my idea but the thing the
  • 00:38:57
    bar for me was always like hey can I
  • 00:38:59
    have a technical finder or can I just
  • 00:39:00
    code the thing myself right and I think
  • 00:39:03
    whatever your idea is that's the thing
  • 00:39:05
    AI unlocks is the ability to like bring
  • 00:39:08
    that to life in a really incredibly fast
  • 00:39:10
    way and you can iterate much fast or
  • 00:39:12
    much faster with with AI so when I every
  • 00:39:15
    day I'm doing something with Claude and
  • 00:39:16
    maybe it Sut my lifestyle that like one
  • 00:39:18
    of my better friends now is like an AI
  • 00:39:20
    assistant and that's not healthy
  • 00:39:22
    everything you need you need to know
  • 00:39:23
    about you but but I'm like back and
  • 00:39:25
    forth like hey take this like tell me
  • 00:39:27
    think about this like what's the first
  • 00:39:28
    draft of this oh what if we did this
  • 00:39:30
    like but one thing I'll give everyone a
  • 00:39:32
    quick quick piece of advice for cl if
  • 00:39:34
    you want to make right and better short
  • 00:39:36
    from content just ask it to create
  • 00:39:37
    something more ask it to do the same
  • 00:39:39
    thing but make it punchier the punchier
  • 00:39:41
    Thing Really Works you just make a
  • 00:39:42
    punchier make a punch year but it's it's
  • 00:39:45
    the it's how it unlocks your creativity
  • 00:39:48
    and the speed you can iterate with it
  • 00:39:50
    that I just think is the most exciting
  • 00:39:53
    thing for me to be you know to be living
  • 00:39:56
    in this time where you can actually
  • 00:39:57
    bring those ideas to life uh through
  • 00:39:59
    through AI yeah yeah I totally agree and
  • 00:40:04
    not um I'll go on a little side quest
  • 00:40:07
    here uh because you we were talking
  • 00:40:08
    about writing styles earlier and you
  • 00:40:09
    have that kind of embedded in the agent
  • 00:40:11
    that you built Kieran um like one of the
  • 00:40:13
    things I think that that's kind of super
  • 00:40:15
    fascinating and this one again wasn't
  • 00:40:16
    possible for AI so one of the um agents
  • 00:40:19
    I built a week and a half ago is to take
  • 00:40:22
    a uh piece of writing or a YouTube video
  • 00:40:24
    and extract the genetic code of that
  • 00:40:26
    writing like what is it that defines
  • 00:40:28
    that piece of writing it's like oh the
  • 00:40:29
    average length is this how us how long
  • 00:40:31
    paragraphs are U does it use figures of
  • 00:40:33
    speech if so which ones and so it can do
  • 00:40:35
    that right that's a very doable thing um
  • 00:40:37
    kind of extracting the genetic code and
  • 00:40:39
    so what I want to um I what I'm thinking
  • 00:40:41
    about doing is say let's say we extract
  • 00:40:43
    those uh gentic code elements whatever
  • 00:40:45
    and I may have an idea of what my
  • 00:40:47
    audience uh that maybe they they like
  • 00:40:48
    punching things maybe they don't and
  • 00:40:50
    they're something they're likely
  • 00:40:51
    Universal but sometimes uh in certain
  • 00:40:53
    industries certain kinds of writing will
  • 00:40:55
    work and certain kinds won't right and
  • 00:40:56
    we just sort of don't know and it's
  • 00:40:58
    never really been possible to sort of
  • 00:41:00
    test that but now imagine that says okay
  • 00:41:02
    I'm going to pick uh styles of writing
  • 00:41:04
    that I like or that I know have been
  • 00:41:05
    successful extract that genetic code but
  • 00:41:08
    then I want to say um in the same way we
  • 00:41:10
    run AB tests for like a subject line or
  • 00:41:12
    whatever I want to take these four
  • 00:41:14
    writing styles exactly and put them in a
  • 00:41:15
    loop in my agent and just cycle through
  • 00:41:18
    and then track the activity post uh post
  • 00:41:20
    production and see what happens it's
  • 00:41:22
    like okay now for my audience for my
  • 00:41:24
    people here's the thing that seems to
  • 00:41:26
    work and then use that for future post
  • 00:41:28
    that I do right now right like that was
  • 00:41:31
    inconceivable even if you were a coder
  • 00:41:32
    like you could not do it right that not
  • 00:41:34
    a d you could you could start an entire
  • 00:41:35
    company and an entire Army of Engineers
  • 00:41:37
    and you still wouldn't be able to pull
  • 00:41:38
    that off now you can do that in like two
  • 00:41:41
    weekends and some coffee right it's uh
  • 00:41:43
    that that's exactly what that's exactly
  • 00:41:46
    what I was thinking through when I built
  • 00:41:47
    that agent because you can see I have
  • 00:41:48
    the five different riding Styles and
  • 00:41:49
    what I would love to do is like
  • 00:41:50
    categorize all of the posts under a
  • 00:41:52
    rideing style and see for each rideing
  • 00:41:54
    style which one is performing best and
  • 00:41:56
    you can imagine world in the future
  • 00:41:58
    where you can look at the riding Style
  • 00:41:59
    by like certain segments because you're
  • 00:42:01
    because you're right you now have the
  • 00:42:02
    ability to that's the thing again coming
  • 00:42:04
    back to where we start a podcast I have
  • 00:42:06
    just been fascinated when AI can
  • 00:42:08
    extrapolate like to you you say like the
  • 00:42:10
    DNA of this thing like what what makes
  • 00:42:13
    up this thing this morning I was kind of
  • 00:42:14
    doing um we have in man coming up I for
  • 00:42:17
    for some reason sign up to seven talks I
  • 00:42:19
    was like yes yes no problem yes and then
  • 00:42:22
    now the all the slides are du and so I
  • 00:42:25
    was doing slides this morning and I
  • 00:42:27
    started working with Claud I was like I
  • 00:42:28
    wonder if you could just like build a a
  • 00:42:31
    prompt to begin with to basically do
  • 00:42:33
    what you said for like a deck like
  • 00:42:35
    extrapolate from a PDF that I upload and
  • 00:42:37
    from that PD like if this is a deck I
  • 00:42:39
    like just build a template that an AI
  • 00:42:42
    assistant can then just replicate into
  • 00:42:45
    that style and that to me works with
  • 00:42:48
    Claude I have to maybe it works in chat
  • 00:42:49
    PD I've got so obsessed with Claude and
  • 00:42:51
    the cool thing about agent the is
  • 00:42:53
    actually you can pick different llms for
  • 00:42:55
    each step and I've been doing that as
  • 00:42:56
    well
  • 00:42:58
    but I still think that that is the most
  • 00:43:01
    I don't know if people understand how
  • 00:43:02
    incredible that is like you can actually
  • 00:43:04
    have ai take something and build the
  • 00:43:08
    extract as like the DNA and then like
  • 00:43:11
    templae it yes
  • 00:43:14
    yes okay yeah um I get really worked up
  • 00:43:17
    all this stuff two things kind of pull
  • 00:43:18
    on one is around composing things across
  • 00:43:21
    LM so as you've noted so I'm I'm a huge
  • 00:43:23
    fan of of claw 3.5 by the way if we were
  • 00:43:26
    having this conversation 6 months ago
  • 00:43:28
    and I'm a huge Fanboy of open AI right
  • 00:43:31
    and fully I'm an investor in open AI
  • 00:43:33
    they were so far ahead that it's like
  • 00:43:35
    okay well yes there will be always the
  • 00:43:36
    second Rams and there and maybe the but
  • 00:43:38
    I could not imagine a world in which
  • 00:43:40
    open a was not at the kind of Frontier
  • 00:43:42
    of it and then all a sudden CLA 3.5
  • 00:43:44
    comes along it's like it's measurably
  • 00:43:46
    better right and I find myself using it
  • 00:43:48
    more and so the nice thing about tools
  • 00:43:50
    like agent. is you can so you have
  • 00:43:52
    access to GPT 40 GPT 4 mini Cloud 3.5
  • 00:43:55
    all the versions you have all the imag
  • 00:43:57
    generation models so you can choose and
  • 00:43:59
    then swap out as things evolves like you
  • 00:44:01
    may decide in your agent that says you
  • 00:44:03
    know for this step in it I used to use
  • 00:44:04
    cloud 3.5 but now GPT 5 does a better
  • 00:44:07
    job and your customer users don't need
  • 00:44:09
    to know right that's they don't need to
  • 00:44:11
    be trying to decide which one's better
  • 00:44:13
    and take a prompt that you you put on
  • 00:44:15
    your blog post and copy paste that into
  • 00:44:17
    this versus that now your agent takes
  • 00:44:19
    care of that's what software is supposed
  • 00:44:20
    to do is raise the level of abstraction
  • 00:44:22
    for the M user so they don't have to
  • 00:44:23
    think about things like here's the
  • 00:44:25
    problem they're trying to solve I'm
  • 00:44:26
    trying to create a LinkedIn post for
  • 00:44:27
    YouTube video Kieran has that figured
  • 00:44:29
    out and will keep his agent updated so
  • 00:44:31
    it's using the latest and greatest uh
  • 00:44:33
    you know technology whatever happens to
  • 00:44:34
    be um so the other kind of idea here is
  • 00:44:37
    like this happens often in image
  • 00:44:39
    Generation Um the same way that happens
  • 00:44:41
    with regular llms is that different
  • 00:44:43
    image generation models are good at
  • 00:44:45
    different use cases I'll give you a very
  • 00:44:47
    very simple example there are some uh
  • 00:44:48
    image generation models that absolutely
  • 00:44:50
    suck at anything that involves
  • 00:44:52
    text but are better at 90% of the other
  • 00:44:56
    things let's just say uh but this is
  • 00:44:57
    actually true so when someone puts a
  • 00:45:00
    prompt in you can actually make a decent
  • 00:45:02
    guess it's like based on the prompt
  • 00:45:04
    which image generation model should I
  • 00:45:05
    use the user is trying to generate an
  • 00:45:07
    image or four Images uh that they can
  • 00:45:09
    pick from maybe you mix and match images
  • 00:45:11
    maybe you learn that says oh you tend to
  • 00:45:12
    like idiogram uh images better than
  • 00:45:15
    flux. one I don't know why but the the
  • 00:45:17
    agent can get smarter on a user's behalf
  • 00:45:19
    as well automatically B based on which
  • 00:45:21
    ones were chosen and kind of cross so it
  • 00:45:23
    doesn't have to be one LM doesn't have
  • 00:45:25
    to be one image generation model the
  • 00:45:27
    Asian can get smarter just like a human
  • 00:45:28
    would it's like you know if you had an
  • 00:45:30
    assistant or had an intern it's like
  • 00:45:31
    Kieran seems to approve of these kinds
  • 00:45:33
    of things more than those kinds of
  • 00:45:35
    things I'm going to give him more of
  • 00:45:36
    these kinds of things in the future
  • 00:45:38
    makes sense right that's what we all do
  • 00:45:39
    I will say the next one that I'm going
  • 00:45:40
    to build is um for YouTube shorts like
  • 00:45:43
    replicate this because we had a
  • 00:45:44
    incredible Creator on called Jenny hoers
  • 00:45:47
    she has a billion views on shorts and
  • 00:45:49
    she gave this master class of how to do
  • 00:45:51
    shorts and there's four a 4minute
  • 00:45:52
    segment where she went through like
  • 00:45:54
    every second of a 60 second short and
  • 00:45:58
    exactly why that second belongs there
  • 00:46:00
    and I'm like oh wait a minute I can take
  • 00:46:02
    that 4 minute I can transcribe it I can
  • 00:46:04
    turn that into like a prompt for the llm
  • 00:46:07
    to teach it and then I can put it into
  • 00:46:09
    agents that AI in a couple of steps and
  • 00:46:10
    actually replicate the kind of same
  • 00:46:12
    process to actually help people build
  • 00:46:15
    YouTube short transcripts and why it's
  • 00:46:18
    structured that way and again I had that
  • 00:46:20
    thought last night right I had that I
  • 00:46:22
    was like wow you know she I reminded me
  • 00:46:24
    because a clip come up on my feet of
  • 00:46:25
    Jenny and I was like hey wait a minute I
  • 00:46:27
    can take that and I can templa it and
  • 00:46:29
    again not you know I'll edit it and
  • 00:46:31
    things like that but and then I can
  • 00:46:33
    probably build out an agents at AI
  • 00:46:34
    tonight right if I am allowed to I get
  • 00:46:37
    the time free from being a new parent
  • 00:46:40
    and uh that's how quick again you can go
  • 00:46:42
    from idea to execution so I really first
  • 00:46:44
    of all appreciate you coming on the show
  • 00:46:46
    we really appreciate it
  • 00:46:48
    appreciate okay for anyone out there
  • 00:46:50
    listening it's e. it works on a credit
  • 00:46:53
    system right now so you get 100 credits
  • 00:46:55
    when you sign up for free um yeah you
  • 00:46:57
    don't have to pay any money listeners of
  • 00:46:59
    the Pod um if you go to agent. sign up
  • 00:47:03
    you will get your 100 credits if you put
  • 00:47:05
    the promo code in there it's blindingly
  • 00:47:08
    obvious you click on the little credits
  • 00:47:09
    button up there and here's the code it's
  • 00:47:11
    matg rocks as in marketing Against the
  • 00:47:14
    Grain rocks matg rocks you put that in
  • 00:47:17
    you get 50 bonus credits on top of the
  • 00:47:18
    100 that you would get uh for free sign
  • 00:47:20
    up so and I would like to point out that
  • 00:47:22
    we talked about kieran's sick agent that
  • 00:47:26
    first of all Karan are you going to
  • 00:47:27
    share it or not have you decided I I I
  • 00:47:29
    I'll share it I just won't share the
  • 00:47:31
    code but I will share the agent for all
  • 00:47:32
    of you listening because the Des agent
  • 00:47:35
    which is buildon agent. there's also a
  • 00:47:37
    bunch of really other cool agents there
  • 00:47:40
    that I would just be remiss if we didn't
  • 00:47:42
    talk about one that I loved daresh was
  • 00:47:45
    the company research agent where you
  • 00:47:47
    just put in a URL you and I love the
  • 00:47:50
    folks over at similar web I literally
  • 00:47:51
    put in similarweb.com and I got all of
  • 00:47:54
    the information I would ever want like
  • 00:47:56
    all the office locations specific
  • 00:47:59
    breakdowns of who the CEO is and how how
  • 00:48:03
    are they positioning the company so many
  • 00:48:05
    things that you could use for sales
  • 00:48:07
    Outreach marketing automation really
  • 00:48:09
    really cool engagement work that could
  • 00:48:12
    come off of that and I thought that was
  • 00:48:14
    another awesome awesome agent and also
  • 00:48:17
    by the way guys there's a meme generator
  • 00:48:21
    that's awesome I'm really excited about
  • 00:48:23
    o Oasis getting back together um Oasis
  • 00:48:27
    tour 2025 and so agent. built me a sweet
  • 00:48:31
    meme all about Oasis getting back
  • 00:48:33
    together but the fact that like there
  • 00:48:36
    are so many different use cases in in in
  • 00:48:39
    the early version of this platform I
  • 00:48:40
    think is super cool so dares we thank
  • 00:48:42
    you very much for not only coming on the
  • 00:48:44
    show but building agents. and helping
  • 00:48:46
    bring agents to folks uh in a world
  • 00:48:49
    where we're still very early on in this
  • 00:48:52
    technology this data is wrong every
  • 00:48:54
    freaking time have you heard of
  • 00:48:57
    HubSpot HubSpot is a CRM platform where
  • 00:49:00
    everything is fully integrated W I can
  • 00:49:02
    see the client's whole history call
  • 00:49:04
    support tickets emails and here's a task
  • 00:49:07
    from 3 days ago I totally
  • 00:49:09
    missed HubSpot grow better
Etiquetas
  • agent.ai
  • AI tools
  • AI agents
  • marketing automation
  • content creation
  • AI integration
  • productivity
  • automation
  • non-programmers
  • AI democratization