5+ AI Workflows You Can Copy For Your Business in 2024

00:42:17
https://www.youtube.com/watch?v=R3l4m6SZaS0

Ringkasan

TLDRIn this episode of "The Next Wave" podcast, hosts Matt Wolf and Nathan L delve into advanced AI use cases beyond the basic functionalities that most people are familiar with. They emphasize exploring deeper capabilities of AI tools like Claude and perplexity and discuss innovative workflows that leverage these technologies. Matt shares how he uses Claude to create a news summarizer, which helps simplify and bullet-point complex news articles, making it easier to create YouTube content. He also highlights the "pages" feature in perplexity, likening it to creating mini Wikipedia pages on topics of interest. Nathan discusses the importance of custom instructions in AI models and how they can streamline processes like translations and meeting preparations. He emphasizes the potential of AI to reduce workloads and enhance efficiency. Matt goes into detail about his automated process for adding tools to his website, "Future Tools," using make.com. This involves scraping web content, extracting summary information with GPT-4, and posting updated details to his platform. The podcast underlines the time-saving and mental-load-reducing benefits of AI in business processes. Additionally, they discuss how advancing tools like idiogram 2.0 can now generate more sophisticated content such as YouTube thumbnails, raising interest in their potential applications. The hosts touch upon the evolving nature of AI tools and encourage listeners to experiment with these technologies.

Takeaways

  • 🤖 AI tools can do more than basic tasks; explore their deeper applications.
  • 📄 Claude is used to summarize news into bullet points, aiding in content creation.
  • 🔍 Perplexity's 'pages' feature helps in-depth learning on complex topics.
  • ⚙️ Automation in web content management saves time and effort.
  • 📈 Custom workflows can streamline business operations.
  • 🧩 AI can simplify and organize complex tasks.
  • 📝 Use AI for meeting agendas by analyzing past communications.
  • 🌐 Perplexity might have SEO advantages for content ranking.
  • 🎨 Idiogram 2.0 helps with text-based content creation like thumbnails.
  • 🗂 Make.com's automation boosts efficiency in resource management.

Garis waktu

  • 00:00:00 - 00:05:00

    The podcast begins with hosts Matt Wolf and Nathan L discussing cool real-time applications of AI, beyond just the obvious surface-level uses like conversation with ChatGPT or image creation with MidJourney. They emphasize the depth of AI's potential and Matt introduces the episode focus: exploring interesting AI use cases, including how he uses AI in his business 'Future Tools'. He mentions a news summarizer project created with Claude to simplify AI news content.

  • 00:05:00 - 00:10:00

    Matt explains one of his AI workflows using Claude for summarizing AI weekly news. He has a system prompt for Claude that summarizes uploaded content into bullet points, explains novelty, reports for news broadcast, and simplifies for general understanding. He highlights an AI development where a neural model simulates 'Doom' in real-time based on user interactions, emphasizing how AI aids in his video content creation for better accessibility and understanding.

  • 00:10:00 - 00:15:00

    Nathan and Matt discuss the time and mental energy saved by using AI tools for complex tasks, like understanding AI research papers. Matt explains how AI helps simplify intricate concepts, which he uses especially for YouTube content. Nathan suggests applying similar tools for tasks like contract review or complex document processing. Matt shares how customized GPT systems are created and used for their unique needs.

  • 00:15:00 - 00:20:00

    The conversation highlights the utility of AI's custom instructions to enhance task automation. Nathan shares his use of AI for learning Japanese, involving custom instructions for effective translations and cultural feedback, facilitated by shorthand commands. This indicates how AI can customize learning experiences. The segment explores AI's potential in translation tasks, inferred to be expanded into real-time conversational aids.

  • 00:20:00 - 00:25:00

    Matt dives into using Perplexity’s page feature to create comprehensive topic-specific AI-generated informative pages. Demonstrating on Quantum Computing, he shows how users can add sections, learn deeply, and utilize it as a knowledge-sharing tool. These pages can include various linked resources, making it a versatile tool for both personal learning and team education. Highlighted is the page's functionality for shared learning experiences.

  • 00:25:00 - 00:30:00

    Nathan discusses how idiogram 2.0 can now adeptly create YouTube thumbnails, leveraging its new text handling capabilities. Though not entirely perfect, it's effective for enhancing video presentations and slides, applicable in YouTube and professional settings. The potential for future improvements was noted, particularly around generating personalized elements like likeness in imagery for enhanced relatability.

  • 00:30:00 - 00:35:00

    Matt showcases his sophisticated workflow for adding AI tools to his website 'Future Tools'. Using Make.com, he automates tasks from scraping website data to generating tool descriptions through GPT. This streamlines site updates, showing automation’s efficiency in data handling. The discussion inspires thoughts on using AI for similar business processes, highlighting how future advances with AI agents could simplify such workflows.

  • 00:35:00 - 00:42:17

    Finally, Nathan shares how AI aids in creating meeting agendas from multiple communication sources, optimizing meeting prep time. He envisions a future where AI agents fully automate this process by aggregating relevant communication insights. This use case underlines AI's role in improving efficiency in administrative tasks, exemplifying its potential to enhance decision-making and focus within business collaborations.

Tampilkan lebih banyak

Peta Pikiran

Mind Map

Pertanyaan yang Sering Diajukan

  • What did Matt Wolf and Nathan L discuss in the podcast?

    They discussed advanced AI use cases beyond basic applications, highlighting tools like Claude, perplexity, and make.com for enhancing business processes and other tasks.

  • How is Claude used according to the podcast?

    Claude is utilized for summarizing news articles and transcripts into bullet points, helping simplify complex information for easier understanding.

  • What is perplexity's function as mentioned in the podcast?

    Perplexity is used to create detailed informational pages similar to Wikipedia entries, helping users understand complex topics by generating comprehensive summaries.

  • How does Matt automate his website updates?

    Matt uses make.com to automate adding tools to his website by scraping content, summarizing it with GPT, and uploading it to his site automatically.

  • Why are AI tools important according to Nathan and Matt?

    AI tools are important because they save time, reduce mental load, and simplify complex tasks, making them more efficient to handle.

  • What is the purpose of creating custom AI workflows mentioned in the podcast?

    Custom AI workflows help automate and simplify repetitive or complex tasks, improving efficiency and accuracy in personal or business processes.

  • How can you use AI for meetings according to Nathan?

    AI can be used to summarize past emails and discussions to create meeting agendas, helping keep meetings focused on important topics.

  • Is the content generation automated in Matt's process?

    Yes, Matt's process involves using AI and automation tools to scrape, summarize, and publish content to his website with minimal manual intervention.

  • Did the hosts provide any SEO tips related to AI?

    They briefly mentioned someone suggesting that perplexity-generated pages might rank well in SEO but haven't tested it themselves.

  • How can idiogram 2.0 be used as per the discussion?

    Idiogram 2.0 can be used to create YouTube thumbnails with text easily, though further customization might be needed for desired results.

Lihat lebih banyak ringkasan video

Dapatkan akses instan ke ringkasan video YouTube gratis yang didukung oleh AI!
Teks
en
Gulir Otomatis:
  • 00:00:00
    I want to try to see if I can like
  • 00:00:01
    inject my link future
  • 00:00:04
    [Music]
  • 00:00:06
    tools. we're figuring out cool use cases
  • 00:00:08
    in real time on this episode right
  • 00:00:11
    now fivey it linked to Future tools I'm
  • 00:00:15
    telling you saying this is like a a
  • 00:00:17
    great SEO hack right
  • 00:00:20
    now hey welcome to the next wave podcast
  • 00:00:23
    I'm Matt wolf and once again I'm here
  • 00:00:25
    with Nathan L and today we're talking
  • 00:00:28
    once again about AI use cases and I feel
  • 00:00:31
    like a lot of people sort of know the
  • 00:00:33
    surface level uses of AI right you kind
  • 00:00:36
    of know how to go have conversations
  • 00:00:38
    with chat GPT or Claud or maybe make an
  • 00:00:40
    image with mid Journey but so many
  • 00:00:42
    people are just sort of looking at the
  • 00:00:44
    tip of the iceberg and that Iceberg goes
  • 00:00:46
    so deep there are so many cool things
  • 00:00:48
    that you can do with these AI tools and
  • 00:00:51
    in this episode we're going to Deep dive
  • 00:00:53
    into some of the more interesting use
  • 00:00:55
    cases that you can use AI for some of
  • 00:00:58
    them fairly basic and easy Implement and
  • 00:01:01
    one of them is a little more complex and
  • 00:01:03
    sort of a deep dive into literally how
  • 00:01:05
    I'm using AI to run my business I sort
  • 00:01:07
    of pull back the curtain and show you
  • 00:01:09
    the whole process of what I'm doing to
  • 00:01:12
    run my future tools website using Ai and
  • 00:01:14
    we're going to get to all of that in
  • 00:01:16
    this episode I think you're going to get
  • 00:01:17
    a lot of value out of it and learn a
  • 00:01:19
    whole bunch of new ways to use AI so
  • 00:01:22
    let's just go ahead and jump right into
  • 00:01:24
    it I can just go ahead and kick it off
  • 00:01:27
    um I'll I'll share uh a fairly simple
  • 00:01:29
    simple basic one that I use this one is
  • 00:01:33
    uh using Claude uh I use Claude a lot so
  • 00:01:37
    a lot of my workflows use clad I love
  • 00:01:39
    making projects inside of clad but one
  • 00:01:42
    of these little projects that I made is
  • 00:01:44
    called the news summarizer so if you
  • 00:01:47
    watch my other YouTube channel every
  • 00:01:50
    Friday I put out an AI news video where
  • 00:01:52
    I break down like here's all of the
  • 00:01:54
    latest AI news for the week and this is
  • 00:01:58
    pretty much how I come up with like what
  • 00:02:00
    I'm going to say about the news for the
  • 00:02:02
    most part is I've created this new
  • 00:02:04
    summarizer here and I'll go ahead and
  • 00:02:07
    show you my system prompt I I'll read it
  • 00:02:09
    off as well for anybody that's listening
  • 00:02:11
    but I have this system prompt here that
  • 00:02:14
    says I will upload or paste in an
  • 00:02:16
    article PDF or video transcript your job
  • 00:02:19
    is to respond with bullet points with
  • 00:02:21
    the following
  • 00:02:22
    information summarize the content into
  • 00:02:25
    bullet points making it easy to
  • 00:02:26
    understand the concepts or ideas
  • 00:02:28
    presented tell me what we can do now as
  • 00:02:31
    a result of this information that we
  • 00:02:32
    could not do before what makes this
  • 00:02:34
    novel tell me how I should report on
  • 00:02:37
    this in the news if I was to give a news
  • 00:02:39
    broadcast and explain why this
  • 00:02:41
    information was important what should I
  • 00:02:43
    tell the viewers so that the most amount
  • 00:02:45
    of people can understand it the format
  • 00:02:47
    of the response should be and then I
  • 00:02:49
    give a little format here summary bullet
  • 00:02:50
    bullet bullet what makes it novel bullet
  • 00:02:53
    bullet bullet and then how should I
  • 00:02:55
    describe this news or Concept in a
  • 00:02:57
    simple way that anyone can understand
  • 00:02:59
    when I report on it it if it makes sense
  • 00:03:01
    use an analogy so that's the whole
  • 00:03:03
    system prompt there and basically what
  • 00:03:05
    I'm telling it to do is I'm going to
  • 00:03:07
    give you content might be a PDF might be
  • 00:03:08
    a document might be a copy and pasted
  • 00:03:10
    news article break it down for me so I
  • 00:03:13
    can explain it really really easily and
  • 00:03:16
    one of the things that's been sort of
  • 00:03:17
    circulating in the world of AI news this
  • 00:03:20
    week I'm sure you've probably seen this
  • 00:03:22
    Nathan is that Doom thing right the um
  • 00:03:24
    yeah the AI Doom where it's sort of
  • 00:03:27
    generating every single frame so as like
  • 00:03:30
    clicks the forward button it generates
  • 00:03:32
    the next frame in Doom if they click the
  • 00:03:34
    shoot button it automatically generates
  • 00:03:36
    the next frame there's no like
  • 00:03:38
    underlining game engine yeah no game
  • 00:03:40
    engine underneath it it's all being
  • 00:03:42
    generated as as they're playing so I
  • 00:03:45
    have that PDF and I can draw drag that
  • 00:03:47
    PDF this is literally the um the the
  • 00:03:51
    research report right off of archive.org
  • 00:03:53
    that explains the technical concept of
  • 00:03:57
    how this works and if you're looking at
  • 00:03:59
    my screen right now if you're not
  • 00:04:01
    looking at my screen you can look at it
  • 00:04:03
    over on YouTube but I just dropped the
  • 00:04:05
    PDF right here into Claud and I don't
  • 00:04:07
    even add a prompt all of my prompting
  • 00:04:09
    was done in that system prompt that I
  • 00:04:11
    just read and if I just hit enter on
  • 00:04:13
    this it's going to read that PDF for me
  • 00:04:16
    and then break it down with all of the
  • 00:04:18
    information that I asked it to give me
  • 00:04:20
    in that system
  • 00:04:22
    prompt so right now it's generating all
  • 00:04:25
    that text for me here and here's what it
  • 00:04:27
    says here's a summary and Analysis of
  • 00:04:29
    the research paper summary the paper
  • 00:04:31
    introduces game engine a neural model
  • 00:04:33
    that can simulate the classic game Doom
  • 00:04:35
    in real time at over 200 at over 20
  • 00:04:37
    frames per second on a single TPU right
  • 00:04:40
    and it sort of summarizes what the
  • 00:04:42
    article's about and then the next
  • 00:04:44
    section what makes this novel
  • 00:04:46
    interesting this is the first neural
  • 00:04:48
    model capable of simulating a complex 3D
  • 00:04:50
    game like Doom in real time with high
  • 00:04:52
    visual quality it demonstrates the
  • 00:04:55
    neural models can potentially replace
  • 00:04:56
    traditional game engines for interactive
  • 00:04:58
    3D environments etc etc and then it goes
  • 00:05:01
    on to say here's how to describe this
  • 00:05:03
    news simply right so now you can start
  • 00:05:06
    to see how like on my AI News videos yes
  • 00:05:10
    I put a lot of work into them but like
  • 00:05:12
    AI supports the hell out of me for
  • 00:05:14
    making those News videos so here's what
  • 00:05:16
    it says for how to describe this news
  • 00:05:18
    simply imagine if instead of programming
  • 00:05:20
    every detail of a video game we can just
  • 00:05:22
    show an AI system examples of the game
  • 00:05:26
    being played and it can learn to
  • 00:05:27
    recreate the entire game on its own
  • 00:05:30
    that's essentially what researchers have
  • 00:05:31
    achieved with a classic game called Doom
  • 00:05:34
    they trained in AI to watch gameplay
  • 00:05:36
    videos and then reproduce the game so so
  • 00:05:39
    accurately that even humans have trouble
  • 00:05:41
    telling the difference between the AI
  • 00:05:43
    version and the real thing this could
  • 00:05:45
    revolutionize how video games are made
  • 00:05:47
    in the future potentially making it much
  • 00:05:49
    faster and easier to create complex 3D
  • 00:05:51
    worlds and gameplay it's like teaching
  • 00:05:53
    an AI to be a game designer and
  • 00:05:55
    programmer allinone right so that's like
  • 00:05:58
    one of my most used workflows and this
  • 00:06:01
    is very relevant we're recording this on
  • 00:06:03
    a Thursday I record my News videos for
  • 00:06:06
    YouTube on a Thursday so I was in here
  • 00:06:09
    actually using this Claud project
  • 00:06:12
    earlier today to start to break down my
  • 00:06:15
    News videos now every piece of news that
  • 00:06:18
    I put in those News videos doesn't need
  • 00:06:20
    this process right like some of the news
  • 00:06:21
    is just sort of self-explanatory but
  • 00:06:23
    every once in a while I'll come across a
  • 00:06:25
    research paper and I'm like I want to go
  • 00:06:28
    a little bit more in depth on this I
  • 00:06:29
    really want to make sure that the the
  • 00:06:32
    viewers understand what this means and
  • 00:06:35
    this is how I do that I'll just plug in
  • 00:06:36
    the research paper and it's going to
  • 00:06:38
    simplify it for me and then tell me how
  • 00:06:40
    I should break it down as a piece of
  • 00:06:43
    news that's awesome like I'm curious
  • 00:06:45
    like how much time do you think this
  • 00:06:47
    saves you there's a Time Savings but
  • 00:06:49
    there's also like a mental savings right
  • 00:06:51
    so it probably saves me a good 15 or 20
  • 00:06:54
    minutes of actually like going through
  • 00:06:56
    and reading all of the the concepts
  • 00:06:59
    within the re search paper but at the
  • 00:07:01
    same time these research papers are
  • 00:07:02
    often over my head like let's be honest
  • 00:07:04
    like once they start breaking down like
  • 00:07:06
    math and there's like letters within the
  • 00:07:08
    math equation and it's like IX over 15%
  • 00:07:13
    + 32 a d + 6 I'm like you lost me with I
  • 00:07:17
    right like the very first letter in
  • 00:07:19
    there right so like um like the mental
  • 00:07:22
    low that it takes off of my plate of
  • 00:07:24
    like me reading this research paper like
  • 00:07:27
    three times to try to understand what
  • 00:07:29
    the hell it's actually saying is
  • 00:07:31
    probably the bigger savings than the
  • 00:07:32
    actual time savings for me yeah I feel
  • 00:07:35
    like I should be using this for my
  • 00:07:36
    newsletter probably like I don't use you
  • 00:07:38
    I don't really do anything like this
  • 00:07:40
    that that much for my newsletter but it
  • 00:07:41
    feels like any anytime there's like a
  • 00:07:43
    complicated paper or something like that
  • 00:07:44
    if I get like a nice summary of it uh
  • 00:07:47
    that'd be useful for people I do Wonder
  • 00:07:49
    like it feels like you know maybe a lot
  • 00:07:51
    of our listeners they're probably not
  • 00:07:52
    YouTubers so they're probably like okay
  • 00:07:53
    how do I use this you know it feels like
  • 00:07:55
    you could use this in a lot of different
  • 00:07:57
    parts of business right like even if it
  • 00:07:58
    was like contracts or or whatever any
  • 00:08:01
    kind of complicated paperwork papers
  • 00:08:03
    almost anything any kind of PDF you you
  • 00:08:05
    you could you could make a project in
  • 00:08:06
    Claude or even using you know chbt maybe
  • 00:08:09
    Cloud's kind of better for this right
  • 00:08:10
    now though yeah you can create a custom
  • 00:08:12
    project and kind of tell it like okay
  • 00:08:13
    here's the kind of data to expect and
  • 00:08:14
    here's kind of like what I want out of
  • 00:08:16
    the data here's what here's the kind of
  • 00:08:17
    stuff that's actually important to me
  • 00:08:19
    and then just feed the data in and like
  • 00:08:20
    it does an amazing job of just like
  • 00:08:22
    spitting out stuff that's actually
  • 00:08:24
    important for you yeah yeah I mean this
  • 00:08:26
    I originally came up with this concept
  • 00:08:28
    for my newsletter now I have help I have
  • 00:08:31
    extra writers and editors helping write
  • 00:08:33
    the newsletter uh for me um but I
  • 00:08:35
    originally generated this as a custom
  • 00:08:38
    GPT over in chat GPT but I actually like
  • 00:08:41
    the way Claude does it better I like
  • 00:08:42
    claud's projects better I like the way
  • 00:08:45
    it's organize better I think Sonet 3.5
  • 00:08:47
    gives me better outputs than what the
  • 00:08:49
    current model of gp4 gives me so I sort
  • 00:08:53
    of reimplemented it over in Claud to get
  • 00:08:56
    the same thing but now I've started
  • 00:08:58
    using it for video instead of my
  • 00:09:00
    newsletter but honestly like this can
  • 00:09:02
    just be used for anything you want
  • 00:09:04
    explained more simply right like if you
  • 00:09:07
    come across some sort of article that
  • 00:09:09
    you're like that seems like it's pretty
  • 00:09:11
    interesting but it's over my head like
  • 00:09:13
    how do I understand this like how do I
  • 00:09:15
    just like grasp it stuff like this
  • 00:09:17
    really helps to just like grasp Concepts
  • 00:09:19
    like use some analogies give me bullet
  • 00:09:22
    points explain it as simply as possible
  • 00:09:24
    yeah it feels like opening eye really
  • 00:09:25
    dropped the ball with like not having
  • 00:09:27
    something like the uh Claud projects
  • 00:09:29
    yeah like the projects are so useful I
  • 00:09:31
    think like you first learned about the
  • 00:09:33
    projects or how to use them like on this
  • 00:09:34
    podcast right like I I told you like oh
  • 00:09:36
    that just came out and you're like oh
  • 00:09:37
    cool I heard about that like what is it
  • 00:09:39
    and like yeah here here's how you use it
  • 00:09:42
    I think so the first time I ever used
  • 00:09:43
    them I think you you sort of like I knew
  • 00:09:45
    artifacts and I was using the artifacts
  • 00:09:47
    thing and I was like I saw projects but
  • 00:09:49
    in my mind projects were just like a
  • 00:09:50
    folder to store stuff in I didn't
  • 00:09:52
    realize it had the custom instructions
  • 00:09:55
    and the ability to upload additional
  • 00:09:57
    like information that it would you know
  • 00:09:59
    sort of retrieve when you're prompting
  • 00:10:01
    and I didn't know it had that extra
  • 00:10:03
    stuff yet when we talked about it on the
  • 00:10:05
    podcast and now I'm like this is the
  • 00:10:07
    best thing ever I'm in here like all day
  • 00:10:10
    yeah look in the next 30 seconds I'm
  • 00:10:12
    about to let you in on something that
  • 00:10:13
    will completely change your marketing
  • 00:10:15
    strategy with AI back when we met with
  • 00:10:17
    Greg Eisenberg he summed up the
  • 00:10:19
    widespread fascination with AI perfectly
  • 00:10:21
    everyone from your Barber to Jeff Bezos
  • 00:10:24
    is interested in AI so if you want to
  • 00:10:26
    drop some AI knowledge during your next
  • 00:10:28
    haircut HubSpot has a new trend report
  • 00:10:30
    that is packed with original research
  • 00:10:32
    and data on all the AI Trends to watch
  • 00:10:34
    for this year this PDF is the ultimate
  • 00:10:36
    resource for you you'll learn the top
  • 00:10:38
    five AI marketing tools everyone's
  • 00:10:40
    talking about the tools that are driving
  • 00:10:43
    significant return on investment and it
  • 00:10:45
    answers the question on everyone's mind
  • 00:10:47
    is AI helping or hurting your career
  • 00:10:49
    trajectory you can steal the research
  • 00:10:51
    findings from this report to
  • 00:10:53
    revolutionize your business strategy and
  • 00:10:55
    transform the way you work so if you
  • 00:10:57
    want it grab your free copy from the
  • 00:10:58
    link in the description below now back
  • 00:11:00
    to the show I mean I think most
  • 00:11:02
    listeners probably don't even understand
  • 00:11:04
    even like custom instructions or that
  • 00:11:05
    chat gbt has that they've had it they
  • 00:11:07
    were the kind of the first to have
  • 00:11:08
    custom instructions where you know and I
  • 00:11:11
    I feel like sometimes when people are
  • 00:11:12
    talking about the quality of these
  • 00:11:13
    models I do Wonder like have they ever
  • 00:11:15
    used custom instructions like like like
  • 00:11:18
    maybe they haven't like I I I you know I
  • 00:11:20
    probably should show my custom
  • 00:11:21
    instructions at some point but uh you
  • 00:11:23
    know I feel like it it gives me a lot
  • 00:11:25
    better results from from the model like
  • 00:11:27
    I tell it to like okay be
  • 00:11:29
    politically correct be more
  • 00:11:31
    straightforward tell me the truth uh you
  • 00:11:34
    know and all these kind of things and I
  • 00:11:35
    kind of tell like what level of detail
  • 00:11:36
    do I want and right and and you you can
  • 00:11:39
    give it you know give the models that
  • 00:11:40
    those kind of custom instructions now
  • 00:11:42
    the thing that cla's doing better than
  • 00:11:43
    open ey is that you can have different
  • 00:11:45
    projects with different custom
  • 00:11:46
    instructions based on what you're trying
  • 00:11:48
    to accomplish well you can technically
  • 00:11:49
    do it in chat GPT you just create a GPT
  • 00:11:52
    for each thing because you can give
  • 00:11:54
    every GPT its own custom instructions
  • 00:11:56
    the problem I was having with chat GPT
  • 00:11:58
    was actually built a GPT to do this kind
  • 00:12:01
    of thing for me and it worked really
  • 00:12:04
    really well for several weeks and then
  • 00:12:06
    one day I went in there and tried to use
  • 00:12:08
    it and it wasn't formatting the
  • 00:12:10
    information correctly and it wasn't
  • 00:12:12
    giving me analogies like I was asking it
  • 00:12:14
    to just like one day it started ignoring
  • 00:12:16
    my custom instructions and I'm like what
  • 00:12:19
    happened and I went and open the custom
  • 00:12:20
    instructions in that GPT nothing changed
  • 00:12:22
    they were the same it just felt like
  • 00:12:24
    chat GPT got Dumber one day and so I
  • 00:12:27
    went and tested over on Claude And I'm
  • 00:12:28
    like okay clot is is out custom gping
  • 00:12:32
    GPT you know yeah yeah I mean opening
  • 00:12:36
    eyes done this a lot where they they've
  • 00:12:37
    been making changes so quickly and
  • 00:12:39
    sometimes they're great and sometimes
  • 00:12:40
    like oh yeah they broke something I they
  • 00:12:42
    they came out with that memory feature
  • 00:12:44
    which they barely talked about and I I
  • 00:12:45
    thought it was an amazing feature where
  • 00:12:47
    it's like it'll just like start adding
  • 00:12:48
    stuff to memory I'm like oh this is
  • 00:12:50
    amazing the model's actually remembering
  • 00:12:52
    stuff and then oh I can actually add I
  • 00:12:53
    can manually add stuff to the memory and
  • 00:12:55
    then and now it's going to remember it
  • 00:12:56
    forever that's awesome and you find out
  • 00:12:58
    like there's like a your limit to the
  • 00:12:59
    memory and and not only that but it'll
  • 00:13:01
    actually go in and like remove things
  • 00:13:03
    from the
  • 00:13:04
    memory it'll be like oh and even if you
  • 00:13:07
    don't tell to remember something you
  • 00:13:08
    might type something to it and it's like
  • 00:13:09
    oh it wants to remember that now and it
  • 00:13:11
    says memory updated and like literally
  • 00:13:13
    you go check memory and then sometimes
  • 00:13:14
    that means it actually remove something
  • 00:13:16
    else it will remember stuff that you're
  • 00:13:18
    like why are you remembering that
  • 00:13:19
    specifically that doesn't need to be
  • 00:13:21
    remembered right right yeah I'll be
  • 00:13:23
    asking it for advice on something I'm
  • 00:13:24
    doing in business or whatever in life
  • 00:13:26
    whatever and like it'll be like memory
  • 00:13:27
    update I'm like what the
  • 00:13:30
    what what's your first one you want to
  • 00:13:32
    share yeah mine is you know kind of a
  • 00:13:34
    personal one it's also I I Ed Cloud
  • 00:13:37
    project for it uh I think I've talked
  • 00:13:39
    before my entire marriage and
  • 00:13:42
    relationship you know in Japan like when
  • 00:13:44
    I first moved to Japan my Japanese is
  • 00:13:46
    very basic you know I can't have
  • 00:13:47
    full-blown conversations with anyone I
  • 00:13:49
    can have like I can say small little
  • 00:13:51
    words enough to kind of get some basic
  • 00:13:53
    points across but then I started using
  • 00:13:55
    uh chat BT for translation and I I I did
  • 00:13:59
    through like basically the custom
  • 00:14:00
    instructions where I would like say like
  • 00:14:02
    cuz CU obviously you can tell chbt to
  • 00:14:05
    translate something for you but it's
  • 00:14:06
    kind of tedious when you're doing it all
  • 00:14:07
    the time like yeah you know if you're
  • 00:14:09
    like yeah if you're translating a page
  • 00:14:11
    or you copy and paste everything and and
  • 00:14:13
    you explain what you want but that gets
  • 00:14:15
    annoying if you're actually in a
  • 00:14:16
    conversation with someone right so
  • 00:14:19
    basically I would use custom
  • 00:14:20
    instructions to make like short hands
  • 00:14:21
    like saying hey if you see if I put T
  • 00:14:24
    you know at the beginning of a sentence
  • 00:14:26
    that means I want this translated I
  • 00:14:28
    don't want to tell you hey please
  • 00:14:29
    translate all of this and so that's
  • 00:14:31
    something I haven't seen people do much
  • 00:14:32
    is you can actually make a lot of cool
  • 00:14:34
    short hands yourself with chat BT or
  • 00:14:36
    with claw uh and I I don't really see
  • 00:14:38
    anyone talking about that but you it's
  • 00:14:40
    it's really it's nice you can make Shor
  • 00:14:41
    hands like here's this letter or
  • 00:14:43
    whatever and this is what this what it
  • 00:14:44
    means and this is what I want to
  • 00:14:45
    accomplish now and you just put that at
  • 00:14:47
    the beginning of a of a sentence and uh
  • 00:14:49
    and now it'll it'll do what you want so
  • 00:14:51
    I use that for translations I want
  • 00:14:53
    something translated put a t then I put
  • 00:14:55
    it it knows to translate it um and then
  • 00:14:58
    similar to like how you give you know
  • 00:15:00
    with u with your YouTube channel you
  • 00:15:02
    tell it all this other stuff you wanted
  • 00:15:03
    to do I do the same thing with the
  • 00:15:04
    translations like hey when I'm
  • 00:15:06
    translating something uh to her like uh
  • 00:15:09
    maybe give me cultural feedback too
  • 00:15:10
    because Japan's quite a bit different so
  • 00:15:12
    don't just change what I'm saying but do
  • 00:15:14
    give me feedback like if like maybe for
  • 00:15:15
    Japanese people this is very odd to say
  • 00:15:17
    something like this like like like give
  • 00:15:20
    me cultural feedback but don't change it
  • 00:15:22
    without asking me right like I tell all
  • 00:15:24
    this kind of stuff uh so don't change it
  • 00:15:25
    without asking me also show the English
  • 00:15:27
    underneath so I know
  • 00:15:29
    that you didn't change anything CU
  • 00:15:31
    occasionally when you tell it to
  • 00:15:32
    translate it does change things without
  • 00:15:34
    telling you uh so you should always get
  • 00:15:37
    a version in your your own language to
  • 00:15:39
    kind of double check but also I I've
  • 00:15:41
    used it for more like people were saying
  • 00:15:42
    when we did the episode before when I
  • 00:15:44
    talked about this they were like oh so
  • 00:15:45
    you're literally not going to learn
  • 00:15:46
    Japanese I'm like no I'm trying to learn
  • 00:15:48
    Japanese too like I'm I'm actually
  • 00:15:49
    trying to do both at the same time so
  • 00:15:50
    when I have it translate I also have it
  • 00:15:52
    break things down for me which has been
  • 00:15:54
    one of the cool the coolest things like
  • 00:15:56
    um Japanese has uh like three different
  • 00:15:59
    uh types of characters character systems
  • 00:16:01
    for their language was like makes it
  • 00:16:03
    very complicated yeah kirana Katakana
  • 00:16:05
    and and kanji uh hiana being the most
  • 00:16:08
    basic one so uh I know H herana very
  • 00:16:11
    well I know Katakana kind of okay and
  • 00:16:13
    kanji oh my God that takes like a long
  • 00:16:15
    long time to learn so what I what I have
  • 00:16:17
    it do is basically uh anytime I
  • 00:16:19
    translate something I'm like okay what
  • 00:16:20
    are the key words in that sentence that
  • 00:16:22
    I should learn and then like break them
  • 00:16:24
    down for me uh in hiana you know so
  • 00:16:27
    it'll be like okay you made this s s and
  • 00:16:29
    here's like three bullet points I'll
  • 00:16:31
    tell like I don't want to learn like
  • 00:16:32
    five words at a time make it like three
  • 00:16:34
    I'll learn like three words what are the
  • 00:16:36
    three most important words in that
  • 00:16:37
    sentence and then break them down for me
  • 00:16:39
    and then show me the hirana because I
  • 00:16:40
    can actually read that in Japanese and
  • 00:16:42
    uh and and that's actually helped my you
  • 00:16:44
    know it's it's made it a lot easier for
  • 00:16:46
    me to learn the language it's like okay
  • 00:16:47
    here's that word I just was wanting to
  • 00:16:49
    translate and now break it down for me
  • 00:16:51
    and and teach it to me uh the next level
  • 00:16:53
    I mean ideally in the future like you
  • 00:16:54
    can like the memory be really good right
  • 00:16:57
    and you could like have the model help
  • 00:16:58
    you learn like oh Nathan needed to learn
  • 00:17:00
    these words or he learned these before
  • 00:17:01
    test him on it to see if he actually
  • 00:17:02
    learned it but unfortunately they're not
  • 00:17:04
    really like good enough to like remember
  • 00:17:06
    all that yet but I feel like that's
  • 00:17:08
    probably coming soon yeah yeah no I I I
  • 00:17:10
    noticed um Riley Brown do you know you
  • 00:17:12
    know Riley brown right he he does like
  • 00:17:14
    short form content around AI uh I saw
  • 00:17:17
    him use a similar trick because he's
  • 00:17:19
    doing a lot of stuff where he's using
  • 00:17:20
    Claude to help him code up his his like
  • 00:17:24
    he's trying to make like an app a day
  • 00:17:25
    for 30 days or something like that right
  • 00:17:27
    yeah yeah and one thing Claude tends to
  • 00:17:29
    do when you're asking it to write code
  • 00:17:31
    is it will write just the code that
  • 00:17:33
    needs to be swapped out as opposed to
  • 00:17:36
    rewriting the entire thing so let's say
  • 00:17:38
    just for like a real basic example it
  • 00:17:40
    writes out a whole like HTML web page
  • 00:17:43
    for you and then you say oh the font at
  • 00:17:45
    the top needs to be changed well what
  • 00:17:47
    it'll do is it'll say okay here's the
  • 00:17:50
    code that needs to be changed and it'll
  • 00:17:51
    just give you like three lines of code
  • 00:17:53
    and you're like okay if I don't know
  • 00:17:55
    anything about code I don't know where
  • 00:17:56
    that code is supposed to go like what am
  • 00:17:58
    I supposed to replace well well well
  • 00:17:59
    even if you know how to code I mean I
  • 00:18:00
    know how to code decently like it's
  • 00:18:02
    still annoying so I feel kind of dumb
  • 00:18:03
    now for not having done this so I get
  • 00:18:04
    where you're going so like he has short
  • 00:18:05
    hands to tell it and he Mak sure tell
  • 00:18:08
    like give me the entire code he has a
  • 00:18:10
    short hand that's like give I don't
  • 00:18:12
    remember what his actual short code is
  • 00:18:14
    but it might be like uh CCC or something
  • 00:18:16
    like that he'll put like CCC at the end
  • 00:18:18
    of uh one of his prompts but it's a
  • 00:18:20
    short code to like remind it give me the
  • 00:18:23
    entire code don't just give me like what
  • 00:18:26
    needs to be replaced right and I thought
  • 00:18:28
    that was really I started implementing
  • 00:18:29
    that when I start playing around with
  • 00:18:31
    code as well I always tell it give me
  • 00:18:33
    the entire code don't give me just what
  • 00:18:35
    I need to replace and like the little
  • 00:18:37
    short code thing saves quite a bit of
  • 00:18:38
    time on that aspect yeah that's I need
  • 00:18:42
    to be doing that because i' I've
  • 00:18:43
    definitely used uh Claude for coding and
  • 00:18:45
    like been I've done that many times
  • 00:18:46
    where like it'll it'll give me the small
  • 00:18:47
    little snippet I'm like damn it now I
  • 00:18:49
    gotta go like read through the code and
  • 00:18:50
    it'll take like 30 seconds or a minute
  • 00:18:51
    like give me the whole thing and and
  • 00:18:54
    always wonder why it doesn't do that
  • 00:18:56
    maybe because it's like you know it's
  • 00:18:57
    saving energy or I have no idea why
  • 00:18:59
    they're doing it like saving compute or
  • 00:19:00
    something I don't know but it is
  • 00:19:02
    annoying where it'll just like give you
  • 00:19:03
    a small snippet and it's like yeah yeah
  • 00:19:05
    it it and you can even the funny thing
  • 00:19:07
    is I even have it in my system prompt of
  • 00:19:10
    like when I ask you to generate code
  • 00:19:12
    generate the entire code don't just tell
  • 00:19:14
    me what needs to be replaced and it
  • 00:19:16
    literally always ignores that like you
  • 00:19:17
    need to for whatever reason put it in
  • 00:19:19
    the prompt it doesn't work in the system
  • 00:19:21
    prompt it's so annoying I'm sure they'll
  • 00:19:23
    like that'll get patched up soon but for
  • 00:19:26
    whatever reason it ignores that part of
  • 00:19:28
    the system prompt um so the next thing
  • 00:19:31
    I'll share is I'm going to share a
  • 00:19:33
    little perplexity tip this one will be
  • 00:19:35
    pretty quick here um one thing that I
  • 00:19:38
    think is pretty overlooked inside of
  • 00:19:39
    perplexity but I feel is pretty powerful
  • 00:19:42
    is they have a feature here where if I
  • 00:19:46
    go to my library and go to this page
  • 00:19:50
    feature here this page feature inside of
  • 00:19:53
    perplexity will essentially create like
  • 00:19:55
    a little mini like Wikipedia on any
  • 00:19:58
    topic you want it to create right so if
  • 00:20:00
    I hear about like a concept that I don't
  • 00:20:03
    understand very well but I want to learn
  • 00:20:05
    more about it I'll come to this page
  • 00:20:07
    feature instead of perplexity and have
  • 00:20:10
    it actually generate like a little mini
  • 00:20:12
    Wikipedia for me so for example uh let's
  • 00:20:14
    say I want to learn Quantum Computing
  • 00:20:17
    right like I don't understand Quantum
  • 00:20:18
    Computing I can literally just type
  • 00:20:19
    Quantum Computing hit the little button
  • 00:20:21
    to tell it to go and then it will
  • 00:20:24
    literally just start generating
  • 00:20:26
    essentially a Wikipedia page that
  • 00:20:28
    explains everything I can possibly want
  • 00:20:30
    to know well probably not everything
  • 00:20:31
    there's probably quite a bit of depth at
  • 00:20:33
    quantum computer but it can explain
  • 00:20:34
    pretty much everything a surface level
  • 00:20:37
    person might want to know about Quantum
  • 00:20:39
    Computing um so you can see Quantum
  • 00:20:41
    speedups and database searches Quantum
  • 00:20:43
    algorithms for optimization problems
  • 00:20:46
    Quantum error M mitigation techniques
  • 00:20:50
    and it's just got like this whole page
  • 00:20:51
    now with sources that I can click into
  • 00:20:54
    and click over to the websites to learn
  • 00:20:55
    more and maybe there's something that
  • 00:20:58
    it's missing like um let's say overlap
  • 00:21:02
    of
  • 00:21:03
    quantum Computing and AI I'll hit plus
  • 00:21:09
    and now it's just going to add a little
  • 00:21:10
    section that tells me all about the
  • 00:21:12
    overlap of quantum Computing and Ai and
  • 00:21:15
    how they're related and things like that
  • 00:21:18
    so this to me has been really really
  • 00:21:19
    powerful whenever I sort of get a New
  • 00:21:22
    Concept in my mind that I'm like I need
  • 00:21:24
    to dig into that a little bit more like
  • 00:21:26
    this tool is really powerful for just
  • 00:21:28
    like diving deeper and deeper and deeper
  • 00:21:31
    because I can go here just tell it to
  • 00:21:33
    insert another section and add more
  • 00:21:34
    about it um Quantum speedups and
  • 00:21:37
    database searches okay there's not
  • 00:21:39
    enough info there for me I'll click more
  • 00:21:41
    um oh that doesn't do what I thought it
  • 00:21:43
    did I thought it was going to actually
  • 00:21:44
    add more to it but it doesn't um but I
  • 00:21:47
    can just come down here it saves us to
  • 00:21:49
    an actual it saves this as an actual
  • 00:21:51
    page too right yeah saves it at an
  • 00:21:53
    actual page so if I come back to my
  • 00:21:54
    library here um I I have a tab here for
  • 00:21:58
    Threads which are like the questions
  • 00:21:59
    I've asked perplexity then I have Pages
  • 00:22:01
    here and you can see like it's right
  • 00:22:04
    there I can get right back to this and
  • 00:22:06
    it's also sharable um I can click
  • 00:22:08
    publish up here in the top right of this
  • 00:22:10
    perplexity page and when I publish it
  • 00:22:13
    it's going to give me a link that I can
  • 00:22:14
    share and now I can send this link to
  • 00:22:17
    anybody I want and they can see my
  • 00:22:19
    Quantum Computing Revolution page that I
  • 00:22:21
    had generated for me here in in a matter
  • 00:22:25
    of a minute so I found that to be really
  • 00:22:26
    powerful yeah it's awesome people on SEO
  • 00:22:30
    Twitter talking about this saying that
  • 00:22:31
    like actually these pages are ranking
  • 00:22:33
    quite well right now so for listeners
  • 00:22:35
    that might be something I I haven't
  • 00:22:37
    tested it myself I don't know if it
  • 00:22:38
    works well but it definitely probably if
  • 00:22:40
    you're ring on SEO maybe it's a good
  • 00:22:42
    strategy to to test how would we get our
  • 00:22:44
    website in here though let's say um so
  • 00:22:47
    if I get to this version where I can
  • 00:22:48
    edit I want to try to see if I can like
  • 00:22:50
    inject my link um
  • 00:22:55
    Quantum
  • 00:22:56
    Computing and Future
  • 00:23:02
    tools. Let's see if it'll actually
  • 00:23:04
    generate
  • 00:23:05
    content uh around that relationship five
  • 00:23:09
    s it link to Future tools yeah I'm I'm
  • 00:23:13
    telling you tell you I've been hearing
  • 00:23:14
    people are saying like this is like a
  • 00:23:16
    like a kind of a SEO hack right now I'm
  • 00:23:17
    not sure I haven't tested it myself but
  • 00:23:20
    uh people are saying this is like a
  • 00:23:21
    great way to like get links to your site
  • 00:23:23
    and maybe some traffic I'm not sure if
  • 00:23:25
    the back links are follow links or not
  • 00:23:27
    who I'm not sure but um we're figuring
  • 00:23:29
    out cool use cases in real time on this
  • 00:23:31
    episode right now yeah yeah that they've
  • 00:23:34
    really nailed that I've actually I mean
  • 00:23:35
    I read about this but I haven't actually
  • 00:23:37
    tried it I should have tried it uh it
  • 00:23:39
    looks so cool it's like cooler than I
  • 00:23:41
    expected my my buddy Jude who was like
  • 00:23:43
    an early guy at YC he did something
  • 00:23:45
    similar or he tried to called Golden uh
  • 00:23:48
    back a few years ago and and I think
  • 00:23:50
    they were mostly relying on manual and
  • 00:23:52
    they tried to shift to Ai and it was
  • 00:23:54
    like right when AI was just starting and
  • 00:23:55
    just ended up not go you know I think he
  • 00:23:58
    sold the company for a decent amount but
  • 00:23:59
    it didn't it didn't get as big as he
  • 00:24:00
    hoped but he was trying to do the same
  • 00:24:02
    kind of thing cuz actually I was sitting
  • 00:24:03
    thinking like hey maybe lore.com is a
  • 00:24:04
    great name for that too or something
  • 00:24:06
    like like you know the same kind of
  • 00:24:08
    thing like the lore behind some kind of
  • 00:24:09
    maybe like more like fantasy and
  • 00:24:11
    consumer kind of stuff versus like
  • 00:24:12
    businesses but uh no it's it's really
  • 00:24:14
    cool it's it's a great tool for just
  • 00:24:15
    sort of deep diving on a concept and you
  • 00:24:17
    can just keep adding new sections and
  • 00:24:19
    deep diving into into various topics on
  • 00:24:23
    this you know on this page I can even
  • 00:24:26
    like select and highlight certain area
  • 00:24:28
    and it will sort of extrapolate on
  • 00:24:30
    certain areas that I highlight as well
  • 00:24:32
    so just just a cool Deep dive tool you
  • 00:24:34
    could yeah you could use it as a
  • 00:24:36
    learning tool but also you probably
  • 00:24:37
    could use it as a way to communicate
  • 00:24:38
    things to your team or something
  • 00:24:40
    something they should know about or to
  • 00:24:41
    learn you could kind of curate like
  • 00:24:43
    here's all the information on a topic
  • 00:24:44
    that's actually relevant that you should
  • 00:24:46
    learn and yeah I was building out a page
  • 00:24:48
    um not too long ago that was all about
  • 00:24:50
    YouTube growth strategies right and I
  • 00:24:53
    just kept on having it add new sections
  • 00:24:55
    about ways to grow on YouTube and it
  • 00:24:57
    just kept on finding more information
  • 00:24:59
    about YouTube growth strategies and
  • 00:25:01
    adding new sections related to how to
  • 00:25:03
    grow on YouTube and I'm like this is so
  • 00:25:05
    cool like I could just get lost in
  • 00:25:07
    that's so cool I I yeah I've read about
  • 00:25:09
    it I thought it was cool and I'm like
  • 00:25:11
    now like thinking like yeah I gotta go
  • 00:25:12
    play with this so I'm probably gonna do
  • 00:25:14
    that after the show yeah
  • 00:25:16
    yeah cool so this one is from our our
  • 00:25:19
    friend uh ball sidu who does the uh Ted
  • 00:25:22
    AI podcast yep um I thought this was
  • 00:25:24
    kind of cool he's showing how you can
  • 00:25:26
    use idiogram 2.0 like the new one of the
  • 00:25:28
    new AI art tools that we recently talked
  • 00:25:30
    about uh to make uh YouTube thumbnails
  • 00:25:34
    um it's not perfect I mean I would say
  • 00:25:36
    it's probably not as good as like the
  • 00:25:37
    ones we're currently making or on your
  • 00:25:38
    YouTube channel um but it's really good
  • 00:25:40
    at text like used to that was a thing
  • 00:25:42
    that air really struggled with was like
  • 00:25:44
    getting text on an image and now
  • 00:25:47
    apparently like they've just like
  • 00:25:49
    completely nailed text like it works the
  • 00:25:52
    best at text for sure nothing come close
  • 00:25:54
    to what idiogram can do yeah and uh and
  • 00:25:57
    he says apparently you know actually he
  • 00:25:59
    tags you here um apparently you know
  • 00:26:02
    they're they're also they're working on
  • 00:26:03
    likeness so you probably can do you know
  • 00:26:06
    kind of like how you do where you have
  • 00:26:07
    like AI art that's similar to your face
  • 00:26:09
    uh and use that as a thumbnail like
  • 00:26:11
    apparently there that's in the works so
  • 00:26:13
    Pro probably in the next few months
  • 00:26:15
    you'll be able to make like you know
  • 00:26:16
    your own thumbnails just using idiogram
  • 00:26:18
    versus like having to pay someone or
  • 00:26:20
    spend a bunch of time in Photoshop or
  • 00:26:21
    whatever so I think for people who are
  • 00:26:23
    doing YouTube or have YouTube channels
  • 00:26:25
    or you know uh this is a great use case
  • 00:26:27
    and and also may even for like
  • 00:26:28
    presentations right like if you're
  • 00:26:29
    making a a corporate presentation like
  • 00:26:32
    people love to have like fancy looking
  • 00:26:34
    art with text over it and presentations
  • 00:26:36
    it makes your takes your presentation to
  • 00:26:38
    the next level uh I think people right
  • 00:26:40
    now should be using idiogram for that
  • 00:26:42
    yeah does he break down a workflow there
  • 00:26:43
    is it like is there like a a specific
  • 00:26:46
    process he's using to get those results
  • 00:26:48
    or is he basically just saying like hey
  • 00:26:50
    idiogram is is great at this as far as I
  • 00:26:53
    can as far as I can tell he's just
  • 00:26:54
    saying idiogram is great at this I don't
  • 00:26:55
    I don't think there's any like yeah the
  • 00:26:58
    thumbnails that it made and it looks
  • 00:26:59
    good they almost look like a type of
  • 00:27:01
    thumbnail you might see on like a like
  • 00:27:04
    an NPR sort of YouTube channel or
  • 00:27:06
    something like that yeah but I'm just
  • 00:27:09
    curious what the prompts would be to get
  • 00:27:11
    it looking like that right because the
  • 00:27:12
    way it's sort of uh the the text is
  • 00:27:15
    Justified to the left in the bottom left
  • 00:27:17
    corner of the images like on the reality
  • 00:27:20
    bending 3D captures the word captures is
  • 00:27:22
    highlighted in blue like I wonder if
  • 00:27:25
    there was like some prompt tricks that
  • 00:27:27
    he used to get that or if he just like
  • 00:27:29
    said make an image that says this you
  • 00:27:31
    know yeah I mean I right right now like
  • 00:27:33
    my process for YouTube thumbnails is
  • 00:27:35
    pretty much create an image with AI um
  • 00:27:38
    and then uh inject my face into it and
  • 00:27:41
    then pull it into canva and add the text
  • 00:27:43
    on top because I haven't really gotten
  • 00:27:45
    the AI generators to sort of uh stylize
  • 00:27:48
    the text I like or like put it in the
  • 00:27:50
    exact location that I like yet so yeah
  • 00:27:54
    you know if idiogram if if you can go
  • 00:27:55
    and get that granular I want the text
  • 00:27:57
    Left Alive in the bottom left corner of
  • 00:27:59
    the image and I want this uh you know
  • 00:28:01
    this image aspect this the aspect of the
  • 00:28:04
    thumbnail to be up in the top right cor
  • 00:28:06
    like if you could sort of dial that
  • 00:28:07
    stuff in that'll be that that'll be
  • 00:28:10
    really really awesome that that would
  • 00:28:11
    save a lot of time and it'll save a few
  • 00:28:13
    extra steps for people yeah for sure
  • 00:28:16
    very cool do you have any spicy ones
  • 00:28:18
    like uh on our episode the other day oh
  • 00:28:20
    how how I make all of my AI girlfriends
  • 00:28:25
    um my my Herm of AI girlfriends
  • 00:28:29
    here's here's how you can get around the
  • 00:28:30
    restrictions or
  • 00:28:33
    whatever so here's what I'll share so
  • 00:28:35
    this one is going to be a little bit
  • 00:28:37
    more in the weeds this is actually how I
  • 00:28:40
    add tools to Future tools now this is
  • 00:28:43
    probably not going to be super relevant
  • 00:28:46
    to everybody watching but I think this
  • 00:28:48
    should give people ideas of the types of
  • 00:28:51
    workflows they can build to sort of run
  • 00:28:54
    their businesses and make them a little
  • 00:28:55
    bit more automated behind the scenes
  • 00:28:57
    right obviously this exact step-by-step
  • 00:29:00
    process is not going to be something
  • 00:29:02
    you'll need in your business but
  • 00:29:04
    hopefully it gets the wheels turning of
  • 00:29:06
    what these tools are capable of so when
  • 00:29:10
    I've come across a new tool that I want
  • 00:29:11
    to add to Future tools I actually came
  • 00:29:13
    up there's a tool that I'm going to add
  • 00:29:15
    to Future tools right now going through
  • 00:29:18
    this process there's this tool called
  • 00:29:19
    spotter which is a a YouTube Tool which
  • 00:29:24
    helps you come up with titles and
  • 00:29:25
    thumbnails and you know break down hooks
  • 00:29:27
    for your YouTube videos and stuff like
  • 00:29:29
    that it's a tool that that I use right
  • 00:29:31
    but it's not on Future tools yet so when
  • 00:29:33
    I want to add it to Future tools here's
  • 00:29:35
    what I do I copy the URL in my browser I
  • 00:29:39
    come over to this uh Google sheet
  • 00:29:42
    spreadsheet um I have a little tab that
  • 00:29:44
    I call the machine and I plug in the URL
  • 00:29:47
    here to this input URL then I have a
  • 00:29:51
    make.com automation setup that does a
  • 00:29:54
    whole bunch of steps so all I do is I
  • 00:29:57
    plug in this input URL I go to
  • 00:30:00
    make.com I have they call like they you
  • 00:30:03
    know in zaper they call them zaps
  • 00:30:04
    make.com they call them scenarios uh so
  • 00:30:07
    I have this scenario called integration
  • 00:30:09
    scraping be here and what it does is it
  • 00:30:12
    watches that Google sheet that I just
  • 00:30:14
    showed you for a a new entry into this
  • 00:30:17
    input URL here and then it uses a site
  • 00:30:20
    called scraping B which goes and takes
  • 00:30:25
    all of the content that's on that web
  • 00:30:28
    site so it will go and look at spotter.
  • 00:30:30
    La or whatever the URL was it will look
  • 00:30:33
    at that site and it will look at all of
  • 00:30:35
    the text on that site and it will scrape
  • 00:30:37
    it all once it scrapes all of that
  • 00:30:40
    content it then goes into a chat GPT
  • 00:30:44
    this is actually using GPT 4 um right
  • 00:30:47
    here it will actually go into GPT 4 and
  • 00:30:51
    you can see my my prompt that I give it
  • 00:30:53
    here summarize what this tool does in a
  • 00:30:55
    single paragraph and include what it can
  • 00:30:58
    be used for and why people might want to
  • 00:31:00
    use it um your response should begin
  • 00:31:03
    with the name of the tool right so
  • 00:31:05
    basically like spotter does X Y and Z
  • 00:31:08
    and so it takes everything that was
  • 00:31:10
    scraped you can see the message content
  • 00:31:12
    here is um what was scraped and then
  • 00:31:15
    additional message content is to
  • 00:31:17
    summarize this so it summarizes it into
  • 00:31:19
    a paragraph and then it takes that
  • 00:31:22
    paragraph and it tells it to summarize
  • 00:31:24
    it into a short little snippet right so
  • 00:31:27
    if you go to to the Future tools website
  • 00:31:29
    the homepage shows just like a one
  • 00:31:31
    sentence snippet of what a tool does and
  • 00:31:33
    when you click in you see a longer
  • 00:31:35
    paragraph about what that tool does so
  • 00:31:38
    this second run through of chat GPT here
  • 00:31:41
    it says in as few words as possible
  • 00:31:43
    describe what this tool does in one sort
  • 00:31:45
    sentence your response should begin with
  • 00:31:47
    the name of the tool so same idea it
  • 00:31:50
    just takes the longer paragraph moves it
  • 00:31:52
    to the next step in this workflow here
  • 00:31:55
    and makes a one- sentence version of it
  • 00:31:58
    and then the last step here oh so this
  • 00:32:02
    one actually if you look back at my
  • 00:32:04
    spreadsheet here you see you got the
  • 00:32:05
    input URL we've got our short
  • 00:32:07
    description our long description and the
  • 00:32:10
    tool name so the last one basically just
  • 00:32:12
    pulls in the tool name right it just
  • 00:32:14
    looks at the website says what is this
  • 00:32:16
    tool called and it pulls in the tool
  • 00:32:18
    Name by looking at the website here so
  • 00:32:21
    what is the name of this tool your reply
  • 00:32:22
    should have no extra sentence or details
  • 00:32:24
    please simply output the name of the
  • 00:32:26
    tool and nothing else
  • 00:32:28
    and then once it does that it takes
  • 00:32:30
    everything it just pulled it scrapes it
  • 00:32:33
    makes a long paragraph makes a short
  • 00:32:34
    sentence figures out the title and then
  • 00:32:36
    injects it back into the Google sheet so
  • 00:32:38
    let's let's pretend I'm pulling in
  • 00:32:40
    otter.ai and it's not on the website yet
  • 00:32:42
    I could put the input URL is
  • 00:32:44
    otter.ai come over
  • 00:32:49
    here I'm going to make sure it's running
  • 00:32:51
    in the right thing I'll press run once
  • 00:32:53
    and now it'll run through the whole
  • 00:32:54
    process it's scraping the sales page
  • 00:32:57
    using chat GPT to summarize it so we can
  • 00:33:00
    see it pulled in the tool name of otter
  • 00:33:02
    AI here over on the short description
  • 00:33:05
    otter AI provides automated note-taking
  • 00:33:07
    and real-time transcription for meetings
  • 00:33:09
    over in the long description you can see
  • 00:33:11
    here that it's a much more in-depth
  • 00:33:13
    explanation of what otter AI does and so
  • 00:33:17
    now it's basically created my short
  • 00:33:18
    description the title um the long
  • 00:33:21
    description and it also automatically
  • 00:33:23
    pulled in the um the featured image from
  • 00:33:27
    the website that's not an AI thing
  • 00:33:29
    that's just something that I have it
  • 00:33:30
    import through Google Sheets with and
  • 00:33:33
    then once that's done I actually have a
  • 00:33:34
    separate little automation that runs
  • 00:33:36
    after
  • 00:33:37
    that that basically creates the short
  • 00:33:39
    link and uploads it to webflow for me so
  • 00:33:43
    I would go that's awesome choose where
  • 00:33:45
    it is I run this one once and this one's
  • 00:33:49
    giving me an error but the only reason
  • 00:33:50
    it's giving me an error is because I've
  • 00:33:51
    already got this tool on my site but
  • 00:33:53
    what you'll notice here is it would have
  • 00:33:56
    filled in this Main tool URL with the
  • 00:33:58
    new URL that it created and then it
  • 00:34:00
    would just automatically add it to the
  • 00:34:02
    web flow website but if I look over at
  • 00:34:04
    the Future tools website here it's
  • 00:34:07
    giving me an error right now but the
  • 00:34:08
    reason it's giving me an error is
  • 00:34:09
    because if I search out
  • 00:34:11
    otter there it is so you can see it's
  • 00:34:13
    already loaded on the website so it's
  • 00:34:14
    giving me an error it won't actually let
  • 00:34:16
    me add a new tool that's already on the
  • 00:34:18
    website but that's the whole workflow I
  • 00:34:20
    told you it's in depth and in the weeds
  • 00:34:23
    but I wanted to show it because I really
  • 00:34:25
    love this make.com site right if if
  • 00:34:28
    you've ever used that beers Zer can do a
  • 00:34:30
    lot of the same kind of stuff but you
  • 00:34:31
    can create these amazing workflows where
  • 00:34:34
    it does things like scrape a website and
  • 00:34:36
    then create a long paragraph create a
  • 00:34:38
    short paragraph grab the title for me
  • 00:34:40
    and then put all of that information
  • 00:34:42
    into a spreadsheet for me so whether
  • 00:34:44
    you're doing any sort of like data
  • 00:34:45
    analysis and you're going in trying to
  • 00:34:48
    pull a bunch of data off a whole bunch
  • 00:34:49
    of websites you can use a workflow like
  • 00:34:51
    this right um it lets you do a whole
  • 00:34:54
    bunch of stuff in bulk and this will
  • 00:34:56
    also work if let's say I have like 20
  • 00:34:58
    tools listed here it'll just one at a
  • 00:35:00
    time go through the list and follow that
  • 00:35:02
    same process for every tool in the list
  • 00:35:05
    yeah I mean that workflow is awesome I
  • 00:35:07
    just the whole time you're talking I'm
  • 00:35:08
    sitting there thinking like that's
  • 00:35:10
    really cool but it's also very
  • 00:35:11
    complicated and like God this this is
  • 00:35:13
    the kind of stuff that that agents are
  • 00:35:15
    going to like
  • 00:35:16
    simplify right right and so I was
  • 00:35:18
    wondering like okay make's cool zpp is
  • 00:35:20
    cool do those sites still exist in the
  • 00:35:23
    future like when you have agents like
  • 00:35:24
    like maybe that's why dares is making
  • 00:35:26
    that agent. thing
  • 00:35:28
    maybe maybe's maybe he's realized like
  • 00:35:30
    those kind of sites are going to be
  • 00:35:31
    simplified down into like okay here are
  • 00:35:33
    the agents and here's like the kind of
  • 00:35:35
    the instructions for the agents yeah and
  • 00:35:37
    you just and you just hand the agents
  • 00:35:38
    your API keys and it does all that like
  • 00:35:41
    automatically I mean it's a little bit
  • 00:35:43
    more hands off than I I think I made it
  • 00:35:45
    look because I had to like turn off the
  • 00:35:48
    automations from running automatically
  • 00:35:50
    so that I could demo them on the video
  • 00:35:53
    um so literally what I'm doing when I
  • 00:35:55
    add a new tool to the website is I'm
  • 00:35:57
    plugging in the URL of the website and
  • 00:36:01
    then it just goes and does the rest
  • 00:36:03
    right it's watching the Google sheet for
  • 00:36:05
    new information being added in like I
  • 00:36:08
    don't know every half hour or every hour
  • 00:36:09
    or something like that it comes back
  • 00:36:11
    double checks the Google sheet to see if
  • 00:36:13
    there's any new information on the
  • 00:36:14
    Google sheet if there is it goes through
  • 00:36:17
    the list and runs the automation so my
  • 00:36:19
    role in all of this is literally just
  • 00:36:21
    adding a URL to the input URL on the
  • 00:36:23
    Google sheet and then all of the rest of
  • 00:36:25
    it happens behind the scenes automat
  • 00:36:27
    ially so it's sort of like I created my
  • 00:36:29
    own little AI agent but you're right I
  • 00:36:32
    think in the future I won't have to
  • 00:36:34
    build all that I'll just have to tell
  • 00:36:36
    like an AI or like a chatbot here's what
  • 00:36:39
    I want you to do and it will build all
  • 00:36:41
    those automations that I just sort of
  • 00:36:42
    showed you the workflow for yeah but
  • 00:36:45
    then it'll like I said it'll probably
  • 00:36:46
    just ask you your API Keys it'll save
  • 00:36:47
    that somewhere secure so it remembers it
  • 00:36:49
    for the future and it'll just do all of
  • 00:36:51
    that that's going to be so awesome yeah
  • 00:36:53
    yeah so share something less complex cuz
  • 00:36:58
    that one was really in the
  • 00:36:59
    weeds mine's Mine mine's way less
  • 00:37:01
    complex so um yeah so this is something
  • 00:37:05
    I I I started doing recently um so so
  • 00:37:07
    right now I'm kind of working you know
  • 00:37:09
    I've had lore.com for a long time and
  • 00:37:10
    I've been trying to figure out what to
  • 00:37:11
    do with it I'm I'm still kind of trying
  • 00:37:12
    to figure out what to do with it Beyond
  • 00:37:14
    like I like writing my newsletter but I
  • 00:37:16
    feel like it's such a great domain I
  • 00:37:17
    should be doing something more with
  • 00:37:19
    it uh you know originally I bought the
  • 00:37:21
    domain cuz I was trying to do the movie
  • 00:37:23
    studio with with Barry Osborne uh and so
  • 00:37:25
    I've been interested in using it for
  • 00:37:27
    something more more entertainment
  • 00:37:28
    related for a long time so so right now
  • 00:37:30
    I'm I'm I'm doing this thing called Lore
  • 00:37:31
    Labs I'm like it's kind of like a AI
  • 00:37:33
    video agency and a kind of community to
  • 00:37:35
    help people figure out how to create AI
  • 00:37:37
    videos as well still figuring it out so
  • 00:37:39
    I've been working with these two great
  • 00:37:40
    guys one of them who who makes some of
  • 00:37:41
    the best AI videos I've ever seen I'm
  • 00:37:43
    not announcing who he is yet but we're
  • 00:37:44
    collaborating on it and uh so we've got
  • 00:37:46
    a slack full of you know uh different
  • 00:37:50
    conversations about like okay what
  • 00:37:51
    should be our strategy
  • 00:37:53
    pricing what's the initial strategy what
  • 00:37:55
    are we going to try if that doesn't work
  • 00:37:56
    like you know lots and lots of
  • 00:37:58
    conversations figure figure everything
  • 00:38:00
    out and then we're having weekly
  • 00:38:01
    meetings and then one thing I realize AI
  • 00:38:03
    is really good at and I think other
  • 00:38:05
    people could use it this way too is like
  • 00:38:06
    using AI is almost like a meeting agenda
  • 00:38:09
    like okay you're having a meeting you've
  • 00:38:11
    been having all these conversations
  • 00:38:12
    through email or slack or whatever just
  • 00:38:14
    throw them all in there into into Claud
  • 00:38:17
    and you you make a custom project for it
  • 00:38:19
    or whatever and uh and say hey what are
  • 00:38:22
    the actual like things that need to be
  • 00:38:23
    decided upon what are the like action
  • 00:38:25
    items you need to do you know key
  • 00:38:28
    takeaways from the
  • 00:38:29
    conversations uh and then just have it
  • 00:38:31
    create the meeting agenda automatically
  • 00:38:33
    that way versus like well you gonna sit
  • 00:38:34
    there like 30 minutes and like type up
  • 00:38:36
    the meeting agenda uh and have to re-
  • 00:38:38
    remember what you actually talked about
  • 00:38:40
    yeah yeah yeah that's so that's what I'm
  • 00:38:41
    doing now is I literally just copy and
  • 00:38:42
    paste all of it in there and like I told
  • 00:38:44
    it here's how I want my meetings
  • 00:38:46
    structured here's the conversations
  • 00:38:48
    figure out the agenda yeah and then and
  • 00:38:51
    then send it out so that's super smart
  • 00:38:53
    so you're basically let's say you've got
  • 00:38:54
    some emails you've got some DMS um
  • 00:38:57
    you've got some tweets I don't know
  • 00:38:59
    you've got all of these various uh
  • 00:39:01
    places that you've had communication you
  • 00:39:03
    just pull all of these pieces of content
  • 00:39:06
    in these communications you've had put
  • 00:39:09
    them into clad and say I'm about to meet
  • 00:39:11
    with this person based on everything I
  • 00:39:13
    just put into the your input how should
  • 00:39:16
    we like H how should we keep this
  • 00:39:18
    meeting on track write an agenda for me
  • 00:39:20
    yeah totally and and and ideally there's
  • 00:39:23
    you have like an agent for that in the
  • 00:39:24
    future right like it actually goes out
  • 00:39:25
    there and like yeah we're talking onil
  • 00:39:27
    we're talking on slack and you know
  • 00:39:29
    whatever yeah go and get that for us and
  • 00:39:30
    then actually create the invite that's
  • 00:39:32
    like the next step right to actually do
  • 00:39:33
    all of it um but right now yeah I'm
  • 00:39:36
    manually copying and pasting and then
  • 00:39:37
    just it but it still saves me probably
  • 00:39:39
    20 minutes or so um but uh you know
  • 00:39:42
    because some meetings it's like okay
  • 00:39:43
    what did we talk about in the past or
  • 00:39:44
    what's actually important to discuss you
  • 00:39:46
    know just tell the AI like what you want
  • 00:39:48
    like what's you know how do you like
  • 00:39:49
    meetings to be structured yeah Fe feed
  • 00:39:52
    it feed the data in it's pretty it's
  • 00:39:54
    pretty good what's even cooler too now
  • 00:39:55
    is this is just a brand new feature came
  • 00:39:57
    out this week actually well the week
  • 00:39:59
    that we're recording this uh uh the week
  • 00:40:01
    that we that we're recording this there
  • 00:40:03
    was a new feature in Google um Google
  • 00:40:05
    meet where now we'll summarize the
  • 00:40:07
    meetings right this has been in Zoom for
  • 00:40:08
    a little while now but Google meet just
  • 00:40:10
    rolled out the feature where it will if
  • 00:40:12
    you turn it on it's not always listening
  • 00:40:13
    to your calls but if you turn it on it
  • 00:40:15
    will listen to your calls and basically
  • 00:40:17
    summarize the meeting for you and make
  • 00:40:20
    like a a to-do list off of the meeting
  • 00:40:22
    like an action item list kind of thing
  • 00:40:24
    and so let's say you need to have a
  • 00:40:26
    second call with that person the a
  • 00:40:27
    followup call well on that follow-up
  • 00:40:29
    call you just grab the meeting notes
  • 00:40:31
    from your last call plug that in and say
  • 00:40:34
    here's what we talked about last time
  • 00:40:36
    you know and yeah yeah yeah yeah what
  • 00:40:37
    progress has been made what's unresolved
  • 00:40:40
    you know what do we need to discuss yeah
  • 00:40:42
    it's awesome so really really cool no I
  • 00:40:45
    think I think uh these are some awesome
  • 00:40:47
    ideas and uh I really like this format
  • 00:40:50
    like I love this format of just here's
  • 00:40:52
    some of the cool ways we're using them I
  • 00:40:53
    think we're going to get to a point
  • 00:40:55
    where we're like all right we're running
  • 00:40:56
    out of use cases that use ourself but
  • 00:40:59
    we're constantly out there like fishing
  • 00:41:00
    for other use cases we're watching
  • 00:41:02
    Twitter and YouTube and Reddit and all
  • 00:41:05
    these places where people are sharing
  • 00:41:06
    their cool AI use cases so I think it'll
  • 00:41:09
    be cool I think maybe next time we do
  • 00:41:10
    one of these we'll bring like a bunch of
  • 00:41:12
    like here's some cool workflows that we
  • 00:41:15
    came across that other people are using
  • 00:41:16
    that we think you should know about no
  • 00:41:18
    but this has been a fun discussion uh
  • 00:41:20
    once again let us know your thoughts in
  • 00:41:22
    the comments wherever you're listening
  • 00:41:23
    to this if you're on YouTube let us know
  • 00:41:25
    your thoughts if you have some cool use
  • 00:41:26
    cases that you want share and you want
  • 00:41:27
    us to maybe shout them out on a future
  • 00:41:29
    episode share them in the comments that
  • 00:41:31
    might be a really great way to discover
  • 00:41:33
    new workflows if we talk about your
  • 00:41:35
    workflow we'll shout out your name on
  • 00:41:37
    the episode and let everybody know that
  • 00:41:39
    that came from you um but I think we're
  • 00:41:41
    probably going to do more of these
  • 00:41:42
    episodes so let us know what you think
  • 00:41:44
    of them in the comments either on
  • 00:41:45
    YouTube or Spotify or wherever you're
  • 00:41:47
    listening to podcasts and uh if you want
  • 00:41:49
    to make sure more of these podcasts show
  • 00:41:51
    up at your feed make sure you like And
  • 00:41:52
    subscribe we really really appreciate it
  • 00:41:54
    makes us feel all warm and fuzzy and
  • 00:41:56
    thank you so much for tuning in yeah
  • 00:41:58
    thank you so much
  • 00:42:06
    [Music]
  • 00:42:12
    [Music]
Tags
  • AI tools
  • Claude
  • perplexity
  • make.com
  • automation
  • workflows
  • AI use cases
  • news summarization
  • business efficiency
  • idiogram 2.0