Custom trained Ai Agents are the next thing.

by probo
44 replies
I believe that AI Agents are the next thing.

Currently, AI is dominated by ChatGPT, Bard, Claude, Stability, etc.

Think of them as Gaint Oracles that you go for answers with bad memory.

how do you fix that.
What you need is a AI personal assistant with memory that can learn about your day-to-day and utilise the big Gaint brains ( oracles or LLMs).

there are few vector memory services like pinecone that are now available . You don't even need to train a LLM which is super cool.

So far, According to my test, GPT4 is the most powerful INTL LLM that is fully commercial.

So how does one use AI to create an Agent that helps you 24/7

Here is an Agent idea.

1-Lawyer Agent.

Lawyer Agent, let's call him Sam. - Imagine Sam is your own AI lawyer that does not charge crazy fees like a human lawyer and has a success rate and intelligence rate better than human lawyers.

Build your Dataset so SAM understands your case files; create SAMs identity and one purpose to help your and your company.


2- Marketing Agent
One way you can use a marketing Agent to by creating an automated response or lead searching e.g on LinkedIn.


Imagine a Super Agent that could create TikTok videos relevant to your product with the goal of increasing follower count and views.

These are just some ideas to spark your thoughts; let me know if you have any other agent ideas.
#agents #custom #thing #trained
Avatar of Unregistered
  • Profile picture of the author Odahh
    I refer to these as task specific AI and we will se thousands of these available to everyone for almost any task at low costs. Basically mimic humans divisions of labor but employed as much by experts . So you have an AI agent lawyer trained on all the legal arguments from a certain lawyer.

    A small fee to consult with the AI but before any official documents are signed a expert evaluation from the actual human lawyer at the standard lawyer rates.

    Or say you learn a language from an AI modeled after a teacher. There will be the option to learn from the real person occasionally to get a real feel of advancing abilities.
    {{ DiscussionBoard.errors[11764913].message }}
  • Profile picture of the author Kay King
    What you need is a AI personal assistant
    But - I thought AI WAS your 'personal assistant'. Inserting another 'helper' could remove you further from your own site/blog/business.

    An AI 'lawyer' can't pass the bar and be licensed....I doubt a court would accept the 'legal argument' of an A.I. programs if you are sued. Just saying - keep in mind the potential limitations.

    AI for learning language? Already well covered by quite a few sites such as Babble. how could AI do it better? There are already many handheld translators so that would not be 'new'.

    I'm sure AI will have many uses we haven't even thought of yet - but i'd advise thinking carefully when you use AI in your online business. Using 'artificial' when your business relies on sales to REAL people - will require knowledge of A.I. and your customers.




    how could AI do it better

    I think the above is the question to ask. Using AI to replace something that already works well can be redundant...finding ways to use AI to accomplish tasks other apps can't do well is the challenge.
    Signature
    Saving one dog will not change the world - but the world changes forever for that one dog
    ***
    Don't count the days - Make the days count
    {{ DiscussionBoard.errors[11765019].message }}
    • Profile picture of the author Odahh
      Originally Posted by Kay King View Post

      But - I thought AI WAS your 'personal assistant'.


      I think the above is the question to ask. Using AI to replace something that already works well can be redundant...
      If it works well but still needs a human and can be done by an AI or robot. The AI or robot will be developed or used to do it as the cost to use drops below the cost of a human and the performance surpasses the majority of humans who do it.

      Look at how Amazon uses AI and robotics to replace human. But the technology is advancing so a Spanish tutor or will be able to create an AI avatar for their students to practice with. Between sessions with the human tutor.

      Robot and AI are being developed to replace fast food and restaurant workers. For the customers who are fine with that that will be their for the customers who want human interaction that will be available.

      Maybe not at Walmart if they can't replace all their tellers with self check out or the self checkout leads to to much theft they will close the stores where it doesn't work and leave the stores open where it is profitable.
      {{ DiscussionBoard.errors[11765026].message }}
    • Profile picture of the author Monetize
      Originally Posted by Kay King View Post

      An AI 'lawyer' can't pass the bar and be licensed....I doubt a court would accept the 'legal argument' of an A.I. programs if you are sued. Just saying - keep in mind the potential limitations.


      It is my understanding that lawyers are using ChatGPT
      because it can cite laws and cases, as well as adjudicate
      legal matters.

      There was a news story about a judge using it to decide
      a case, although this was not in the U.S.

      Colombian judge says he used ChatGPT in ruling

      There are others, but that was the first one that I ever
      heard of where a judge used it to decide a court case.
      {{ DiscussionBoard.errors[11765028].message }}
      • Profile picture of the author DABK
        There was a case where a lawyer used AI, and AI made up precedents (cases) and the judge figured it out and the attorney got in trouble. https://www.forbes.com/sites/mattnovak/2023/05/27/lawyer-uses-chatgpt-in-federal-court-and-it-goes-horribly-wrong/?sh=7e0d42c13494



        Got to be careful with AI.


        Though, if I were a lawyer, I could imagine my AI agent taking over dumb or already answered questions and some good but easy to answer questions:
        like: Hey, what did you say your fee was?
        Hey, do I need to sign or my wife?


        What time is my appointment?


        How much do I still owe you/


        Did you get the check I sent you last Friday?


        How do I get a copy of the agreement?

        Originally Posted by Monetize View Post

        It is my understanding that lawyers are using ChatGPT
        because it can cite laws and cases, as well as adjudicate
        legal matters.

        There was a news story about a judge using it to decide
        a case, although this was not in the U.S.

        Colombian judge says he used ChatGPT in ruling

        There are others, but that was the first one that I ever
        heard of where a judge used it to decide a court case.
        {{ DiscussionBoard.errors[11790473].message }}
        • Profile picture of the author Monetize
          Originally Posted by DABK View Post

          There was a case where a lawyer used AI, and AI made up precedents (cases) and the judge figured it out and the attorney got in trouble. https://www.forbes.com/sites/mattnovak/2023/05/27/lawyer-uses-chatgpt-in-federal-court-and-it-goes-horribly-wrong/?sh=7e0d42c13494



          Got to be careful with AI.


          Though, if I were a lawyer, I could imagine my AI agent taking over dumb or already answered questions and some good but easy to answer questions:
          like: Hey, what did you say your fee was?
          Hey, do I need to sign or my wife?


          What time is my appointment?


          How much do I still owe you/


          Did you get the check I sent you last Friday?


          How do I get a copy of the agreement?

          All of the AI language models have a disclaimer that tells
          you that its information may not be accurate and you need
          to check the outputs.

          If some lawyer took raw output and presented it in a court
          case, that's his own fault for being just plain lazy.

          I proofread and edit just about everything the AIs provide,
          even things that are not technical, unless it's something
          it has proven it can handle.

          I was brainstorming with Gemini a couple of days ago on
          a business model having to do with domain names, and
          it told me that others could register my same domain.

          Sometimes the thing glitches.

          Just like everything else.
          {{ DiscussionBoard.errors[11790476].message }}
    • Profile picture of the author SARubin
      Originally Posted by Kay King View Post


      An AI 'lawyer' can't pass the bar and be licensed....I doubt a court would accept the 'legal argument' of an A.I. programs if you are sued. Just saying - keep in mind the potential limitations.
      I'm not sure AI is ready for legal prime time yet. Here's a link to a recent article about a federal judge who not only tossed a case out of court, but also fined the lawyer $5000 for using made up case law that openai gave him.

      He probably should have double checked before filing the case with the court...


      Lawyers have real bad day in court after citing fake cases made up by ChatGPT
      Signature

      Grow Your Copywriting Skills & Network with Other Copywriting Professionals - Join us at the Copywriters Forum

      {{ DiscussionBoard.errors[11765037].message }}
      • Profile picture of the author Odahh
        Originally Posted by SARubin View Post

        I'm not sure AI is ready for legal prime time yet. Here's a link to a recent article about a federal judge who not only tossed a case out of court, but also fined the lawyer $5000 for using made up case law that openai gave him.

        He probably should have double checked before filing the case with the court...


        Lawyers have real bad day in court after citing fake cases made up by ChatGPT
        This is my crazy view of AI at this point so don't take it as anything that might be real. But the current version of chat got and other chat AI are probably as good as they will get without finding another way to train and improve AI.

        Why. Because so much of the new content available is AI generated content that could be inaccurate or fraudulent. And with AI committing so much plagiarism to create content. A lot of human created work is being flagged as AI generated .


        Humans can be lazy even highly skilled. As with the lawyer you pointed out. Being lazy and overly trusting of the output of AI . Can be very costly for many parties involved

        I'm still pro AI and robotics but it's all still the creation and tools of humans
        {{ DiscussionBoard.errors[11765039].message }}
        • Profile picture of the author SARubin
          Originally Posted by Odahh View Post

          This is my crazy view of AI at this point so don't take it as anything that might be real. But the current version of chat got and other chat AI are probably as good as they will get without finding another way to train and improve AI.
          I agree with you.
          I think AI is a pretty cool technology, but definitely still in its infancy.

          Originally Posted by Odahh View Post

          Why. Because so much of the new content available is AI generated content that could be inaccurate or fraudulent. And with AI committing so much plagiarism to create content. A lot of human created work is being flagged as AI generated .
          Yeah that's going to be a problem for the foreseeable future. And in an effort to make it even more confusing for everyone, I'm currently taking a course in Prompt Engineering from Vanderbilt University.

          My biggest goal is to learn how to give the output a personality, because right now everything AI pumps out sounds robotic to me. Still very cool, but all very left brained with no soul at all. It's just one word, then the next logical word, then the next logical word...(which of course is what it's designed to do)

          But if I can figure out how to give it a personality, even a sudo personality, I think that would be a cool next step. Or a dangerous one? Time will tell...
          Signature

          Grow Your Copywriting Skills & Network with Other Copywriting Professionals - Join us at the Copywriters Forum

          {{ DiscussionBoard.errors[11765057].message }}
          • Profile picture of the author Odahh
            Originally Posted by SARubin View Post

            I agree with you.
            I think AI is a pretty cool technology, but definitely still in its infancy.

            But if I can figure out how to give it a personality, even a sudo personality, I think that would be a cool next step. Or a dangerous one? Time will tell...
            Will it be more work trying to create the perfect promts or talking the bland left brain out of the AI and editing it with your own right brain flavor .

            My view is as the technology advances in the next few years you as an individual can feed a number of examples of how you write and how you speak. Then the AI will turn out work that sounds really close to something you would write. As an individual it's then up to you to check accuracy and add any personal flavor you deem needed.

            Personalized AI avatars that sound more human because they copy individual humans. Who willing provide as much data as possible so an AI agent can duplicate them.
            {{ DiscussionBoard.errors[11765061].message }}
            • Profile picture of the author SARubin
              Originally Posted by Odahh View Post

              My view is as the technology advances in the next few years you as an individual can feed a number of examples of how you write and how you speak. Then the AI will turn out work that sounds really close to something you would write. As an individual it's then up to you to check accuracy and add any personal flavor you deem needed.
              Yes, I agree. So far, when using ai for writing I still find myself editing about 50% for it to feel authentic. But it's getting better all the time, a year ago I had to edit around 90% just to make it passable.


              Originally Posted by Odahh View Post

              Personalized AI avatars that sound more human because they copy individual humans. Who willing provide as much data as possible so an AI agent can duplicate them.
              Yeah, I think the vision I have in mind, if it's even possible, is helping create a sort of empathy / back and forth communication between the machine and a person.

              That's why I'm taking the class, to learn about core prompts. The prompts that work in the background and tell the LLMs how to advance the conversation.
              It's really pretty fascinating stuff so far.

              My main motivation, right now, is because I've seen a lot of stories lately about lonely older people having robot companions to keep them company. And the studies I've seen show a positive effect on both mental and physical health.

              Now imagine an AI companion showing genuine empathy, or at least giving the genuine appearance of it. I wonder what that might look like in the real world?

              Don't know if it'll ever be possible, certainly not with the current technology because let's face it, right now Chat GPT is basically a high tech, high speed, article spinner. But what if it becomes possible in the near future? Even if I'm not the guy who finally cracks the code I'd still love to be on the team when it happens.
              Signature

              Grow Your Copywriting Skills & Network with Other Copywriting Professionals - Join us at the Copywriters Forum

              {{ DiscussionBoard.errors[11765233].message }}
  • Profile picture of the author Odahh
    I believe the people being able to create personal avatars after themselves. Will get close to the empathy results you are looking for.

    Now complete bias here it's probably better to try to give AI the empathy level of a great dog than attempt a human level of empathy. The more human you
    Make something synthetic the more it goes from cute to creepy.

    Take the AI agent/avatars concept trained by individuals to mimic them at closely as possible or the aspect of their personality they want replicated to interact with others.

    So the individual human gets choose the avatars they find more pleasing to interact with for any task. I'm pessimistic on the level of empathy most humans actually show each other. That's why my bias is really lovable dog level empathy.
    {{ DiscussionBoard.errors[11765266].message }}
    • Profile picture of the author Frank Donovan
      Originally Posted by Odahh View Post

      Now complete bias here it's probably better to try to give AI the empathy level of a great dog than attempt a human level of empathy.
      We should remember that any form of writing, like human language itself, is representational. We attach empathy to words only by combining them with a complex cocktail of learned and acquired signals, expectations and context.

      Dogs, or people come to that, don't have to speak to engender empathy.
      Signature


      {{ DiscussionBoard.errors[11765269].message }}
  • Profile picture of the author Kay King
    the level of empathy most humans actually show each other

    When it seems everyone claims victimhood of some sort and demands 'empathy' from others....the empathy pool becomes shallow.
    Signature
    Saving one dog will not change the world - but the world changes forever for that one dog
    ***
    Don't count the days - Make the days count
    {{ DiscussionBoard.errors[11765268].message }}
    • Profile picture of the author SARubin
      Originally Posted by Odahh View Post

      I'm pessimistic on the level of empathy most humans actually show each other. That's why my bias is really lovable dog level empathy.
      Yes, I see your point. Sadly most people only show empathy to people in their own tribe. Sometimes not even then.

      I guess that's where the AI robot of the future comes in...

      Originally Posted by Kay King View Post

      When it seems everyone claims victimhood of some sort and demands 'empathy' from others....the empathy pool becomes shallow.
      Didn't you hear the news, Kay? Victimhood is the new courageous. At least that's what I keep hearing on the news...
      And that's why I stop watching the news every once in a while. So I can detox from all the BS.

      Sad times indeed.
      Signature

      Grow Your Copywriting Skills & Network with Other Copywriting Professionals - Join us at the Copywriters Forum

      {{ DiscussionBoard.errors[11765271].message }}
      • Profile picture of the author Odahh
        Originally Posted by SARubin View Post

        Yes, I see your point. Sadly most people only show empathy to people in their own tribe. Sometimes not even then.

        I guess that's where the AI robot of the future comes in...


        Didn't you hear the news, Kay? Victimhood is the new courageous. At least that's what I keep hearing on the news...
        And that's why I stop watching the news every once in a while. So I can detox from all the BS.

        Sad times indeed.
        Yeah empathy and understanding are generally reserved for family and friends and cute animals. It requires way too much personal energy to invest it in total strangers on a regular basis.

        I deal with hundreds of people on a daily basis generally all have had pretty bad stuff happen to them to horrible. I don't have much empathy or sympathy for people who have been overly coddled and crack at the smallest taste of real life. But I will stay away from sensitive issues with people who are just trying to live their lives.

        AI will be far more manipulative than empathetic if you take into account who is developing them and why eventually the developers want to make money off those who use AI. So if you forgive me for defaulting to this. I will assume those who win and provide the AI and robotics to elderly and lonely person. Will do so in a way that makes the most money.
        {{ DiscussionBoard.errors[11765276].message }}
        • Profile picture of the author SARubin
          Originally Posted by Odahh View Post

          AI will be far more manipulative than empathetic if you take into account who is developing them and why
          That my friend is the existential dilemma. The technology itself is neither good nor bad, but whoever wields the power makes it so.

          Unfortunately that's way above my paygrade, so for now all I can do is try to make the world a little better than I found it, and hope that anything I create isn't used to abuse.
          Signature

          Grow Your Copywriting Skills & Network with Other Copywriting Professionals - Join us at the Copywriters Forum

          {{ DiscussionBoard.errors[11765277].message }}
          • Profile picture of the author Odahh
            Originally Posted by SARubin View Post

            That my friend is the existential dilemma. The technology itself is neither good nor bad, but whoever wields the power makes it so.

            Unfortunately that's way above my paygrade, so for now all I can do is try to make the world a little better than I found it, and hope that anything I create isn't used to abuse.
            I personally see the world mainly getting better for individuals willing to use all these tools and new knowledge being made available to them. To make their own life better.

            AI and robotics as they become better cheaper and more widely available. Will be a net positive. And these AI agent/avatars will be part of that and there will be immoral and self destructive ways to use them. As well as a wide number of ways to waste a large amount of time while a person convinces themselves they are being more productive.

            Overall a net positive but there will be those who fall in love with and marry their AI / robots or other wise completely replace all human interaction with robots and AI agents.

            I'm basically observing stuff already happening and watching the progression. There are a range of other possible things that can happen. But I would need far better writing skills than I currently have to explain them.

            So the AI agents/ avatars is a good thing to talk about
            {{ DiscussionBoard.errors[11765282].message }}
  • What are the advantages of lackin' artifice?

    Beyond the easy-fix bots, you gaht real people on the enda the line.

    K, so they hail from Peru or wherevah (which is fine if'n you live in Peru or wherevah) ... but what 'bout we get smoochiah?

    Like mebbe sum gal from where you at, readin' a script?

    Hmm.

    You know instinctively your great granma would presoom the world gone crazy.

    So what 'bout the gal really knows her stuff?

    An' can interact meaningfully without appearin' no klutz or accidentally pissin' her pants?

    Likley you would prefer this dialog to be 1 to 1.

    In her own store, where she knows zackly what is gowin' on, an' can converse with you direct 'bout anythin' you ask, even if'n you wearin' a PARROT COSTOOM.

    Note to self: parrot costoom for Hallowe'en. Done propah, it is intrinsically more scary than sum dumb green guy with bolts through his neck, or even no bloodsuckah.

    * googles BEAK *
    Signature

    Lightin' fuses is for blowin' stuff togethah.

    {{ DiscussionBoard.errors[11765278].message }}
  • Profile picture of the author max5ty
    In May, Bill Gates did a speech and called it "AI personal assistants" ...which I think is what you may be referring to?

    He also went on to say:

    "...a future AI personal assistant will be so profound that the first company to develop it will have a leg up on competitors."

    I don't believe he was saying AI as your personal Attorney, Doctor, etc. Just an assistant that you could personalize so it knew your wants and needs and could help you with your daily stuff...pointing you in the right direction.
    {{ DiscussionBoard.errors[11765291].message }}
    • Profile picture of the author GordonJ
      MJ DeMarco has been testing an AI assistant at his BusinessFastlane forum. Nothing Held Back facebook group is about to morph into a site, where Alan Sultanic is developing his AI assistant from his massive amounts of work he has done. Should be very interesting to see how they work.

      My AI PA, so far, just can't capture my "get off the lawn", curmudgeon, smart ass personality but I'm sure with some more training...I can be sipping cool ones on the beach whilst GJ AI continues to give the forums his sarcasm, meaness and the true love behind each and every post IT makes.

      GordonJ


      Originally Posted by max5ty View Post

      In May, Bill Gates did a speech and called it "AI personal assistants" ...which I think is what you may be referring to?

      He also went on to say:

      "...a future AI personal assistant will be so profound that the first company to develop it will have a leg up on competitors."

      I don't believe he was saying AI as your personal Attorney, Doctor, etc. Just an assistant that you could personalize so it knew your wants and needs and could help you with your daily stuff...pointing you in the right direction.
      {{ DiscussionBoard.errors[11765366].message }}
      • Profile picture of the author Odahh
        Originally Posted by GordonJ View Post

        MJ DeMarco has been testing an AI assistant at his BusinessFastlane forum. Nothing Held Back facebook group is about to morph into a site, where Alan Sultanic is developing his AI assistant from his massive amounts of work he has done. Should be very interesting to see how they work.

        My AI PA, so far, just can't capture my "get off the lawn", curmudgeon, smart ass personality but I'm sure with some more training...I can be sipping cool ones on the beach whilst GJ AI continues to give the forums his sarcasm, meaness and the true love behind each and every post IT makes.

        GordonJ
        Gordon you are not mean . And this is me saying that. Calling out people's BS . Which is going to be a natural function of AI with all the data it can collect about a person interacting with it regularly.

        Calling BS when someone talks about what they are going to do. When you have the data on them that says it won't happen. You have that down. And it's something AI is perfect for.
        {{ DiscussionBoard.errors[11765402].message }}
    • Originally Posted by max5ty View Post

      [SNIP]and could help you with your daily stuff...[/SNIP]
      Yep. That's how I use multi-modal LLMs. Here are some day-to-day stuff that I ask it to do for me:

      1) Accessibility Assistant (I'm completely blind): "Analyze this [ input_image_showing_visual_graph ]. Convert the data into a spreadsheet that's accessible to blind users of PC screenreaders." Then, code to save the output in CSV format.

      2) From Logic Gates to Contextual Gates in Writing Programmatic Workflows: "Analyze this data: [ input_data ]. If it makes sense based on these parameters, then write 0 and nothing else: [ input_parameters ]. Otherwise, write 1 and nothing else." Then, "If 0, [ do_this ]. Otherwise, ignore and proceed to next item."

      3) Navigation Assistant: "Analyze this top view map of my immediate (indoor or outdoor) environment: [ top_view_photo_from_drone_or_overhead_cam ]. I am currently at [ input_3D_location_from_IMU_sensor_data ]. Tell me how to go from my current position to the nearest male restroom."

      4) Orientation and Placement Assistant: "I am looking at this scene: [ current_frame_from_wearable_video_cam ]. What are the largest objects around me? Describe each briefly. Also tell me each object's XY position in clockface orientation, and each object's Z distance in feet. Use my current position as reference." Then, I can send a follow-up question like: "Are there vacant chairs? If yes, then guide me to it."

      5. OCR for Text in the Wild and Physical Documents: "Analyze this frame: [ current_input_frame_from_wearable_video_cam ]. If there's an area with text, write 0 and nothing else. Otherwise, ignore and move on to the next frame." Then, "If 0, write 0 if text is clearly readable. Otherwise, write 1." Next, "If 0, read text out loud using [ prefered_TTS_voice ]. Otherwise, generate and playback spatial 3D sound of scribbling pen from XY location and Z distance of text area." This way, I can navigate towards and centrally focus the text area, so as to input a much clearer text area for subsequent OCR.

      Signature
      • Deep Learning & Machine Vision Engineer: ARIA Research (Sydney, AU)
      • Founder: Grayscale (Manila, PH) & SEO Campaign Manager: Kiteworks, Inc. (SF, US)
      {{ DiscussionBoard.errors[11768143].message }}
    • Profile picture of the author savidge4
      Originally Posted by max5ty View Post

      In May, Bill Gates did a speech and called it "AI personal assistants" ...which I think is what you may be referring to?

      He also went on to say:

      "...a future AI personal assistant will be so profound that the first company to develop it will have a leg up on competitors."

      I don't believe he was saying AI as your personal Attorney, Doctor, etc. Just an assistant that you could personalize so it knew your wants and needs and could help you with your daily stuff...pointing you in the right direction.
      Google PaLM is going to be exactly this... a true personal assistant with a closed portion of "Personal Data"

      Basically a set of personal preferences that is obtained through Google / android devices that will then reach out to the greater internet set of "knowledge" and return customized personal responses.

      AI as we are using it now is anything and everything from personal... its basically a jacked up search engine. Ask a question get an answer - Try asking ChatGPT what your favorite song is and it would not have a clue - in my mind this is where sh stuff gets interesting - and dangerous
      Signature
      Success is an ACT not an idea
      {{ DiscussionBoard.errors[11769395].message }}
      • Savidge4,

        Originally Posted by savidge4 View Post

        Basically a set of personal preferences that is obtained through Google / android devices that will then reach out to the greater internet set of "knowledge" and return customized personal responses.
        And it's starting.
        • Just went to bard.google.com > Extensions > Grant permission; and
        • Tested with "What are the 100 newest MP3s in my Google Drive? What is each Mp3 all about? Organize your answers in a spreadsheet with the following column headers: [ enter_column_headers ]"

        https://blog.google/products/bard/go...and%20services.
        Signature
        • Deep Learning & Machine Vision Engineer: ARIA Research (Sydney, AU)
        • Founder: Grayscale (Manila, PH) & SEO Campaign Manager: Kiteworks, Inc. (SF, US)
        {{ DiscussionBoard.errors[11769613].message }}
        • Profile picture of the author savidge4
          Originally Posted by Marx Vergel Melencio View Post

          Savidge4,



          And it's starting.
          It is indeed starting. I suspect from Google specifically some new Google home devices that will actually have ROM type memory installed. Home devices already have the ability to recognize different users within a household. At my house Google says "good morning" to my son and I, and not the wife hahaha. and thats about as far is it goes.

          Once it has the ability to collect specific personal end user data... "Good Morning" becomes "Good Morning there is a specific happening today" ( say a programming update related to me, or my sone there is a DLC being released today... or a favorite sports team or a favorite artist released a song etc etc )

          The more noticable change I for see is Google having the ability to say You have been searching for an X... and Google will prompt you with "Hey X is on sale today would you like to look?" AKA AD revenue... but more than that, ridiculously down to a thing and that things interest to the specific person.
          Signature
          Success is an ACT not an idea
          {{ DiscussionBoard.errors[11769639].message }}
          • Profile picture of the author GordonJ
            Originally Posted by savidge4 View Post

            It is indeed starting. I suspect from Google specifically some new Google home devices that will actually have ROM type memory installed. Home devices already have the ability to recognize different users within a household. At my house Google says "good morning" to my son and I, and not the wife hahaha. and thats about as far is it goes.

            Once it has the ability to collect specific personal end user data... "Good Morning" becomes "Good Morning there is a specific happening today" ( say a programming update related to me, or my sone there is a DLC being released today... or a favorite sports team or a favorite artist released a song etc etc )

            The more noticable change I for see is Google having the ability to say You have been searching for an X... and Google will prompt you with "Hey X is on sale today would you like to look?" AKA AD revenue... but more than that, ridiculously down to a thing and that things interest to the specific person.
            About a decade and a half ago, I went shopping as a 16 year old pregnant teen, as a part of a research project...to see what the AI of the time (search results) would serve up. And some followed me for a couple of years with diaper offers, then clothing, all things for babies, kids and such.

            Today, I would run text into an AI speech generator, with Southern accent, and let it tell Suri, et al, what to shop for, and never never have a camera in the house for face recognition. We're not that far away from total AI dossier's which follow you all day and night...and is of course, fed directly into Elon Musk's brain for his enjoyment/manipulation.

            He sees you when you're sleeping, and when you are awake,
            He knows what you are doing, when your lit or baked...
            He knows when to refill your prescriptions, sends the weed right to your door,
            Knows when to take your money, and knows when to take some more.

            The future Santa Claus has his elves make custom one of a kind gifts, just for you.

            GordonJ
            {{ DiscussionBoard.errors[11769884].message }}
          • Profile picture of the author NetMan
            Originally Posted by savidge4 View Post

            It is indeed starting. I suspect from Google specifically some new Google home devices that will actually have ROM type memory installed.

            ...

            The more noticable change I for see is Google having the ability to say You have been searching for an X... and Google will prompt you with "Hey X is on sale today would you like to look?" AKA AD revenue... but more than that, ridiculously down to a thing and that things interest to the specific person.
            Personally, I would at all cost AVOID this, like I would for the last plague...
            {{ DiscussionBoard.errors[11770926].message }}
        • Profile picture of the author NetMan
          Originally Posted by Marx Vergel Melencio View Post

          Savidge4,

          And it's starting.
          • Just went to bard.google.com > Extensions > Grant permission; and
          • Tested with "What are the 100 newest MP3s in my Google Drive? What is each Mp3 all about? Organize your answers in a spreadsheet with the following column headers: [ enter_column_headers ]"

          https://blog.google/products/bard/go...and%20services.
          On the other hand, does anyone really wants Google's AI to connect one's stuff?
          {{ DiscussionBoard.errors[11770925].message }}
  • Profile picture of the author Kay King
    It requires way too much personal energy to invest it in total strangers on a regular basis.
    I don't think it's 'personal energy' - it's time management.

    In generations past, we were expected (as humans) to have empathy and sympathy for family members, for neighbors, for people in our own small town or the neighborhood of our larger city.

    Enter WWW and social media - now we are expected to 'empathize' with people we have never and will never meet - with people we really DO NOT care about because we don't know them. We are told we should be 'involved' and we 'should care'...there are petitions we should sign ...

    .....and there is not enough time in life to care about everyone or every cause.


    But - maybe an AI assistant would give the world the impression you/we do empathize everywhere all the time with everyone.... That might decrease depression and lower the suicide rate?
    Signature
    Saving one dog will not change the world - but the world changes forever for that one dog
    ***
    Don't count the days - Make the days count
    {{ DiscussionBoard.errors[11765406].message }}
  • Profile picture of the author Odahh
    Although personalize AI can easily be designed to work with and support people with physical and mental disabilities. The AI can already detect the onset of depression well before an individual notices they are depressed. While people may also share their feelings with an AI. Well before they would seek human counseling.

    On the other hand an obese person with diabetes who wants others to cover the cost of medication and medical treatment. Will probably be required to have a continuous glucose monitor implanted. With an AI tracking every thing they eat and all physical activity.

    I don't know what you mean by empathy. But I'm inferring rather than empathy. It's more people demanding others except their self destructive behavior and cover the costs of the issue that result.

    Back to more serious implications of these as it will probably be corporations who begin using these agents in mass to replace humans. Which will speed up the decline in commercial real estate currently being observed. Even if every office worker went back to an office 5 days a week.
    {{ DiscussionBoard.errors[11765415].message }}
  • Profile picture of the author mattfrancis40
    Absolutely, AI agents are shaping up to be the next big thing in technology. The current leaders in the field, like ChatGPT and Bard, are like powerful oracles with vast knowledge but a lack of memory.

    To address this, the focus is on creating AI personal assistants that learn from our daily interactions and leverage large AI models like GPT-4. Tools like vector memory services, such as Pinecone, are already helping in this direction.

    GPT-4 stands out as one of the most powerful commercial AI models available, making it a strong candidate for enhancing AI agents.

    Two interesting agent ideas are:

    Lawyer Agent (SAM): Imagine an AI lawyer named SAM, equipped with better success rates and intelligence than human lawyers. SAM learns from case files and provides valuable legal guidance.

    Marketing Agent: AI can be employed to automate responses and lead searches, streamlining marketing efforts. Think about an AI agent creating TikTok videos to boost product visibility.

    In essence, AI agents have the potential to revolutionize various industries. Exploring new agent concepts and refining existing ones will drive technological advancement and redefine how we engage with information and services.
    {{ DiscussionBoard.errors[11767198].message }}
  • Profile picture of the author Kay King
    Lawyer Agent (SAM): Imagine an AI lawyer named SAM, equipped with better success rates and intelligence than human lawyers. SAM learns from case files and provides valuable legal guidance.

    As mentioned above - AI has a tendency to 'create' what it needs - it has 'made up' references and books titles....and has 'made up legal precedent cases, too.


    Automating responses in marketing, etc, will work until the public realizes they are talking to AI - and I expect at that point there will be some pushback.
    Signature
    Saving one dog will not change the world - but the world changes forever for that one dog
    ***
    Don't count the days - Make the days count
    {{ DiscussionBoard.errors[11767200].message }}
    • Profile picture of the author max5ty
      Originally Posted by Kay King View Post

      As mentioned above - AI has a tendency to 'create' what it needs - it has 'made up' references and books titles....and has 'made up legal precedent cases, too.


      Automating responses in marketing, etc, will work until the public realizes they are talking to AI - and I expect at that point there will be some pushback.
      I think some people get 'giddy' with the whole idea of AI replacing this and that.

      It will replace some things. It can't replace some things.

      AI replacing an attorney is ok for things like Privacy Policies, and other paperwork.

      Having a daughter who is an attorney, and having had conversations with her about this subject...

      an attorney represents you in court in front of a jury. A big part of the representation is the way they conduct your defense. What they say. How they say it. It's giving a speech with persuasion. How they connect with the jury, etc.

      Remember OJ and 'if the glove doesn't fit, you must acquit"

      That one phrase by an attorney was believed to help seal the Defense argument.

      Imagine if you went to court to contest a traffic citation and you just played back an AI-written article to mount your defense...

      won't work. The judge looks at your demeanor, your attitude, and a whole list of everything else.

      In order to have an AI attorney, you would need an AI judge, an AI jury...even an AI Defendant and plaintiff.

      So, I think some people are getting a little carried away at times with their love of everything AI and how it is going to replace humans.

      I like AI, but let's keep it real.
      {{ DiscussionBoard.errors[11767203].message }}
      • Profile picture of the author Monetize
        Originally Posted by max5ty View Post

        an attorney represents you in court in front of a jury.


        It's not my intention to pick your post apart but attorneys
        perform other tasks besides representing people before
        a jury. Most things that lawyers are hired for don't involve
        juries at all.

        Lawyers research, write letters, contracts, wills, dispense
        legal advice, and formulate legal arguments. A.I. is able
        to assist in all those tasks, freeing up the lawyer's time to
        do other things.

        If I asked ChatGPT about all the ways that it could assist
        a lawyer, it would give me a list as long as your arm, but
        I won't do that, or post the results because of rules.

        Another thing about lawyers, and this is from my own
        personal experience, is that you have to hire different
        ones for different matters. You need one for wills and
        probates, another one about a consumer matter, and
        a different one for real estate.

        A.I. is going to change all of that and put a lot of people
        out of business.
        {{ DiscussionBoard.errors[11767213].message }}
        • Profile picture of the author max5ty
          Originally Posted by Monetize View Post

          It's not my intention to pick your post apart but attorneys
          perform other tasks besides representing people before
          a jury. Most things that lawyers are hired for don't involve
          juries at all.

          Lawyers research, write letters, contracts, wills, dispense
          legal advice, and formulate legal arguments. A.I. is able
          to assist in all those tasks, freeing up the lawyer's time to
          do other things.

          If I asked ChatGPT about all the ways that it could assist
          a lawyer, it would give me a list as long as your arm, but
          I won't do that, or post the results because of rules.

          Another thing about lawyers, and this is from my own
          personal experience, is that you have to hire different
          ones for different matters. You need one for wills and
          probates, another one about a consumer matter, and
          a different one for real estate.

          A.I. is going to change all of that and put a lot of people
          out of business.
          I said AI was good for some things when it came to attorneys.

          Now, let's get real with your argument...

          the first people it will put out of business are people who write articles etc. Which from past comments, I think you fall into that category?

          Very soon there will be no need for anyone to read an article since AI will change from search engines to answer engines.

          Why would anyone read an article you had AI write when all they will need to do is ask AI themselves to tell them what they want?

          It seems some are getting a little wacky when they're talking about how all those 'other' people are going to lose their jobs...when their job (article writing, etc.) is the main purpose of the free AI platforms we have now.

          Thanks for your comment though.
          {{ DiscussionBoard.errors[11767215].message }}
          • Profile picture of the author Monetize
            Originally Posted by max5ty View Post

            I said AI was good for some things when it came to attorneys.

            Now, let's get real with your argument...

            the first people it will put out of business are people who write articles etc. Which from past comments, I think you fall into that category?

            Very soon there will be no need for anyone to read an article since AI will change from search engines to answer engines.

            Why would anyone read an article you had AI write when all they will need to do is ask AI themselves to tell them what they want?

            It seems some are getting a little wacky when they're talking about how all those 'other' people are going to lose their jobs...when their job (article writing, etc.) is the main purpose of the free AI platforms we have now.

            Thanks for your comment though.


            I wasn't making an argument, I was just stating my
            thoughts and my real life experience with lawyers,
            which I am dealing with right now.

            Last month I telephoned the Texas Bar Association
            and was referred to three different ones for different
            issues, and an additional one because of something
            that happened in another part of Texas that a local
            lawyer can't handle.

            I disagree that content or article writing is the main
            function of A.I. but different people use it for what
            they want to use it for. Most of us don't even know
            its full potential.

            Even if A.I. were just a content development tool,
            and even if it were to displace me personally, which
            it won't because I am diversified, it doesn't change
            the fact that there are several professions that won't
            require humans anymore.

            10 Jobs Most at Risk of Being Replaced by AI
            {{ DiscussionBoard.errors[11767227].message }}
            • Profile picture of the author max5ty
              Originally Posted by Monetize View Post

              I wasn't making an argument, I was just stating my
              thoughts and my real life experience with lawyers,
              which I am dealing with right now.

              Last month I telephoned the Texas Bar Association
              and was referred to three different ones for different
              issues, and an additional one because of something
              that happened in another part of Texas that a local
              lawyer can't handle.

              I disagree that content or article writing is the main
              function of A.I. but different people use it for what
              they want to use it for. Most of us don't even know
              its full potential.

              Even if A.I. were just a content development tool,
              and even if it were to displace me personally, which
              it won't because I am diversified, it doesn't change
              the fact that there are several professions that won't
              require humans anymore.

              10 Jobs Most at Risk of Being Replaced by AI
              I probably didn't make myself very clear...

              I meant the AI we talk about around here.

              There is AI that is used in the medical profession that is phenomenal because it can detect anomalies quicker than humans can. There is AI in the transportation industry that can route and reroute airplanes quicker than humans...there is AI that can write code better than some can...

              there are also some professions (like attorneys) that I think have more of a limited use.
              {{ DiscussionBoard.errors[11767238].message }}
              • Profile picture of the author Odahh
                Originally Posted by max5ty View Post

                I probably didn't make myself very clear...

                I meant the AI we talk about around here.

                There is AI that is used in the medical profession that is phenomenal because it can detect anomalies quicker than humans can. There is AI in the transportation industry that can route and reroute airplanes quicker than humans...there is AI that can write code better than some can...

                there are also some professions (like attorneys) that I think have more of a limited use.
                The potential for AI based education/ skills training is amazing. The AI and algorithms we currently have can probably identify human potential and educate on an individual basis exponentially better than what we have now
                {{ DiscussionBoard.errors[11767241].message }}
  • Profile picture of the author Odahh
    I'm not really worried about who gets replaced. It's generally going to be people retiring Or roles that there are not enough people to do.

    Skilled/ productive people will use it as a tool to be far more productive and replace interns and assistants who don't have much actual potential.

    Novices who want to rapidly improve skills and get productive faster have a tool that will help them do that.

    I'd say managers and the large numbers of staff that don't accomplish much except giving managers people to manage will those who get replaced.

    But that is only based off who technology companies are currently claiming to to be getting rid of.

    The manager over seeing 50-10 people or more or the team of managers managing the thousands of employees. Get replaced by a few dozen skilled people with advanced AI.

    The fun part will be watching the multiple failures as people believe they can replace skilled humans with AI and get massive scale. Vertical farming is going down that path. House flipping has not worked out well. And robot truck drivers and taxi cabs are probably still decades away. There is probably more.

    I'm sure there are real farmers using the latest technology in their practices to be far more productive and profitable.
    {{ DiscussionBoard.errors[11767226].message }}
  • Profile picture of the author WF- Enzo
    Administrator
    This isn't really a "next thing"

    Majority of websites already have frontdesk chatbots.



    Originally Posted by probo View Post

    I believe that AI Agents are the next thing.
    Signature
    {{ DiscussionBoard.errors[11768117].message }}
    • Profile picture of the author NetMan
      Originally Posted by WF- Enzo View Post

      This isn't really a "next thing"

      Majority of websites already have frontdesk chatbots.
      Actually there's a BIG difference between AI Agents and ChatBots, but that doesn't prevent from programming a ChatBot as an AI Agent.
      {{ DiscussionBoard.errors[11772960].message }}
  • Profile picture of the author Karshit Bhargava
    AI Agents is a great concept wherein a well trained agent give responses to your query and with self training it goes on to improve itself. For example a banking AI Agent may suggest you good investment plans based on your profile.
    {{ DiscussionBoard.errors[11790466].message }}
  • Creating an AI personal assistant with memory capabilities opens up exciting possibilities for efficiency and personalization. Here's a comment you can use:

    ---

    AI has truly become like giant oracles with vast knowledge but often lacks memory for personalized assistance. Introducing AI personal assistants with memory, leveraging tools like Pinecone for vector memory, can revolutionize user interaction. Imagine an AI like GPT-4 integrated with memory, tailored to assist 24/7, whether as a legal advisor (like SAM) with deep case file understanding or a dynamic marketer creating TikTok content. It's a promising frontier in AI's evolution.
    {{ DiscussionBoard.errors[11796775].message }}
Avatar of Unregistered

Trending Topics