Heygen Video Review

{
  "message": "The lie of Agentic Frameworks for LLMs.\nThe greatest trick the Devil ever pulled was convincing the world he didn't exist.\nTable of contents\nWhat is Agentic AI?\nDifferentiating between wheat and chaff\n\nThe lie of Agentic AI\nA simple abstraction of a complex unreliable API.\n\nLayers of abstractions\nA broken tool.\n\nThe Greatest Trick\nInventing the wheel in the jurassic era.\n\nEnjoying your read?\nIf you do like it, I kindly ask that you consider sharing it.\n\nShare me\nFollow me for updates\nPlease do get in touch, connect or follow. Carry on the conversation!\n\nLinkedIn icon\ngithub icon\ntwitter icon\nrss icon\nNovember 15, 2024\nBy Edward.\nSnake oil salesmen for tyingshoelaces\nYou've heard the famous quote, \"The greatest trick the Devil ever pulled was convincing the world he didn't exist.\"\n\nTheological references are relevant here, because more than any technological advancement in recent times, the mania surrounding Generative AI is more religion than science or engineering.\n\nThe purpose of this blog is not to fight the hype train (I suggest Gary Marcus) for a counterweight. This is a critical analysis of the snake oil being sold to the unsuspecting consumer. I have a simple thesis. Agentic AI frameworks are unreliable and fundamentally unrealistic abstractions. To build complex, chained and repeatable flows using LLMs requires knowledge, skill and hard work.\n\nIt's tempting to look for a shortcut, a silver bullet; but systems are binary communicators of bits and bytes, not vampires to be slain.\n\nWhat is Agentic AI?\nDifferentiating between wheat and chaff\n\n“\nAgentic AI is a groundbreaking advancement in artificial intelligence. A combination of different AI techniques, models, and approaches, it empowers a new breed of autonomous agents that can analyze data, set goals, and take action to achieve them—all with minimal human supervision. Agentic AI allows these autonomous agents to achieve near-human cognition in many areas, turning them into problem-solving machines that thrive in dynamic environments and constantly learn and improve with every interaction.\n\nWow. Amazing. Mindblowing.\n\nAbsolute nonsense as well... Let me try,\n\n“\nAgentic AI is the use of Generative AI in complex workflows. Outputs that are produced by LLMs may or may not be relevant to the task assigned. Agentic AI inherits the sophistication, intelligence and reliability of its output from system instructions and programming, most of which is highly abstracted, written by novices and will result in a barely functioning system. Agentic AI is great for making demos for unsophisticated consumers, yet generally bad for implementing in any use case where the outcome is actually important. It won't work at all unless you are an expert, but you almost certainly aren't.\n\nLet me be clear, I'm not sceptical about Gen AI in the slightest, I think used expertly and with tooling that doesn't yet exist, we can implement workflows that will change the way we perform tasks. Perhaps not to the extent the AI doomsayers talk about, but definitely to the extent of a paradigm shift for many ways of working.\n\nIt is no coincidence that the first killer use case for Gen AI is programming, because programming will always be the first domain to get proper tooling and infrastructure. The reason for this is simple. We build the tooling and infrastructure for programming tools with, errrm, programming. Programming is a low-level concern, a perfect first use case. It is also a skill which will by the way, still be a (if not 'the') relevant and most coveted ability for days, weeks, months, years and probably generations to come.\n\nCoding tools have proven the potential of Gen AI. I've got two separate LLM tools that I (pay for and) use daily for coding because they often get blocked up by overwhelming demand and green infrastructure. I use them concurrently in different IDE windows. This is because they are new, thus slow and sometimes flaky. Even like this, they have revolutionized my workflow. I'm having a bad day when I boot up my computer and Claude is busy so I'm talking to o1-mini instead. And this is the beginning of the shifting ways Gen AI will change the way you work…\n\nThese tools are being built from the ground up, they are brand new and not reliable yet. This is exciting, because it is the beginning. We are in the middle of a new wave. Which brings me to the subject of this blog. The lie and the overreach that is Agentic AI.\n\nThe lie of Agentic AI\nA simple abstraction of a complex unreliable API.\n\nMany years ago I was running a development agency and we had a freshly graduated developer join the team. I set them a reasonably simple task (to build a webform in PHP). I'd check in regularly\n\n“\nHow is it going?\n\n“\nReally well!\n\nDays and weeks passed, and tap tap tap went the keyboard. It was taking too long, so I went over to the programmer and asked them to show me the code they'd been working on. It was a nested layer of forEach loops, reaching into thousands of lines of code.\n\nMy stomach sank (the client was waiting for this…), I was lost for words. It was a physical blow, everything clicked. Was it my fault for not providing guidance? Was the programmer a lost cause? What the blazing hell am I going to tell the client who is supposed to pay for this? (We billed time, not results…)\n\nThe reason I digress is because this is exactly the same feeling I get when I read the docs for any Agentic Framework I have come across. I'm not alone.\n\nReddit Langchain quote\nI don't want to pick on any specific framework (I'm yet to find any Python Agentic Framework I can't apply this too), but when your docs start talking about roles and workers and bosses and backstories the whole concept begins to read like a fantasy novel.\n\nThe reason is, the whole thing IS a fantasy novel. It's a revolutionary and sophisticated tool built on pioneering technology, that works out of the box, for beginners. And every bit as convoluted and contradicted as that sounds.\n\nI'm not saying that complex LLM workflows cannot work, I'm saying that the expertise required and level of control of nuance is extremely high, and the creation of 'simple and accessible' frameworks is simply impossible until tooling is more mature and foundations are more solid. There are very few people in the world right now (although they do walk among us!) with the expertise to create and maintain reliable, complex iterative workflows (based on Gen AI or not!) that work and are true to the definition of Agentic workflows that are being sold the majority.\n\nBut here's the thing that bugs me; these few pioneers are working on cutting edge use cases that we probably aren't aware of. They aren't building in public on twitter, nor writing blogs and how-to-guides. They are not focused on creating a framework so you can easily automate your LinkedIn posts. They are getting on with things quietly and behind closed doors.\n\nI get it, I truly do.\n\nPeople want an accessible, easy to use framework that will allow them to build reliable and complex workflows with real business results. I want many things. Elixir of life. Grow a few centimetres taller. Become the tennis champion in my sports club. It is more likely that I will become the tall local tennis champion with eternal youth than for any of these agentic frameworks to work for the average user, and believe me and anyone who has seen me play tennis when I say that this is extremely unlikely.\n\nAgentic frameworks are the attempt to build a WYSIWYG website builder before we've even agreed a single unified standard for HTML. They are premature.\n\nThis isn't a case of cart before horse, this is trying to build a rocketship before we've discovered electricity. An opportunistic land grab that has already succeeded.\n\nLayers of abstractions\nA broken tool.\n\nThe problem with providing a tool or framework that allows you to abstract functionality is that it comes with a set of assumptions. When I buy a hammer, I assume it will work. When I buy a pressure cleaner, I assume it will work.\n\nThe problem is that when I use a framework, I assume it will work. But this is quite literally impossible given the maturity of the underlying technology. Far from increasing adoption, Agentic Frameworks are selling an illusion on top of highly controlled demos and finite use cases that will never actually work in the hands of the typical user (and there are millions…).\n\nThe damage they are causing is significant. Much as developers these days learn React before they learn javascript, or adopt frameworks before learning about programming principles. A generation of enthusiastic technologists are jumping on the hype train to learn 'x' framework. Here's the thing, it won't work (to any reasonable standard), in fact, building a demo or prototype is quite the achievement in itself.\n\nBut that doesn't mean that Gen AI or LLMs \"don't\" work, even though that will become the assumption in the short term. It simply means that we haven't reached a maturity level of the underlying technology to be able to safely build solid abstractions on top of it. There are too many problems to solve first. A few fundamental issues that need to be solved at a deep level by LLM providers includes but is not limited to:\n\nLatency\nAvailability\nSafety\nReliable output\nScale\n\nSo yes, what I'm saying is that the underlying APIs are often overwhelmed by usage; they are slow; vulnerable to hacking (especially prompt hacking), and introduce breaking changes on a weekly basis. The true innovators are working hard behind closed doors on these problems. But until we have more progress, good luck building a reliable production system, much less a framework.\n\n\nIf you don't have an understanding of the underlying APIs (which by the way, is the actual valuable breakthrough we are all working with), how can you possibly build a system on top of it. Some APIs don't require us to understand their limitations, mainly because they are a black box that just works. It will take years if not a generation for LLMs to reach this level of maturity, until then it should come with a warning sign \"Production use for Experts only\".\n\nIt's impossible to build a reliable abstraction on top of shifting sands. In fact, this breaks the basic tenet of computer programming. Abstraction in computer science is the process of removing elements that distract from more important elements. We can't focus on high level details when the low level system is not reliable, we need to understand nuance and use within a sophisticated paradigm. Not a low/no-code magic framework (with fantastical abstractions).\n\nThis is driving us into the trough of disillusionment, which while not in itself a problem as it provides space to build things of true value, it is still disappointing to observe.\n\n\nThe Greatest Trick\nInventing the wheel in the jurassic era.\n\nThe greatest trick the Agentic AI frameworks have ever pulled is to convince you that they are necessary. They are the devil of the LLM landscape.\n\nYou don't need crews, agents, backstories and pretty graphical interfaces showing you things that will never work.\n\n\nYou need to understand that LLMs are powerful tools that you should understand without stories from a fantasy novel. Your hello world and demo apps should be built by hand, in code. When you understand the how, the why and the what and can control the inputs and outputs to Gen AI models, then you will have a foundation to go out into the world and build something useful. And it will probably take a while.\n\nNot with wizards and managers and agents and dragons, but with the precise deterministic uses of a breakthrough in AI.\n\nFollow me:\n"
}

{
  "structured_data": {
    "voiceSettings": [
      {
        "type": "text",
        "pitch": 10,
        "speed": 1,
        "emotion": "Serious",
        "input_text": "The lie of Agentic Frameworks for LLMs. The greatest trick the Devil ever pulled was convincing the world he didn't exist. A famous line, but surprisingly relevant when discussing technology today, specifically the hype around Agentic AI."
      },
      {
        "type": "text",
        "pitch": 10,
        "speed": 1,
        "emotion": "Friendly",
        "input_text": "Hello, everyone! Today we're diving into the world of Agentic AI—an area often surrounded by misconceptions and overblown promises."
      },
      {
        "type": "text",
        "pitch": 0,
        "speed": 1,
        "emotion": "Serious",
        "input_text": "Agentic AI is touted as a new frontier for artificial intelligence, offering the ability to create autonomous agents that can think and act with minimal human input. This sounds revolutionary, doesn't it?"
      },
      {
        "type": "text",
        "pitch": 0,
        "speed": 1,
        "emotion": "Soothing",
        "input_text": "But let's take a step back. Beyond the claims of mind-blowing advancements, there are significant challenges and limitations that need to be addressed. That's what we are uncovering today."
      },
      {
        "type": "text",
        "pitch": 0,
        "speed": 1,
        "emotion": "Excited",
        "input_text": "So, what is Agentic AI, and why is it generating so much buzz? It's a complex mix of AI techniques claiming to create agents that operate with near-human cognition. Incredible, right? Let's dig deeper."
      },
      {
        "type": "text",
        "pitch": -10,
        "speed": 1,
        "emotion": "Serious",
        "input_text": "However, there's a different narrative here. The truth is that these frameworks are unreliable abstractions of very complex systems. To achieve effective workflows, it takes expertise and hard work, not just high-level abstractions."
      },
      {
        "type": "text",
        "pitch": 0,
        "speed": 1.2,
        "emotion": "Broadcaster",
        "input_text": "Let's talk about the overreach of Agentic AI. Developers are being promised tools that seem as user-friendly and magical as a children's storybook. But the reality is far more complex."
      },
      {
        "type": "text",
        "pitch": 0,
        "speed": 0.9,
        "emotion": "Serious",
        "input_text": "Many frameworks offer an illusion of simplicity, masking the real need for deep technical understanding. These claims often serve more as impressive demos rather than practical solutions for important tasks."
      },
      {
        "type": "text",
        "pitch": 5,
        "speed": 1,
        "emotion": "Friendly",
        "input_text": "Even so, this doesn't mean AI advancements aren't powerful tools—they are! But they require us to roll up our sleeves and understand them deeply, rather than relying on tools that promise to do it all for us."
      },
      {
        "type": "text",
        "pitch": -5,
        "speed": 1,
        "emotion": "Serious",
        "input_text": "These frameworks assume a level of technological maturity we've yet to achieve. Until the underlying infrastructure becomes reliable, these frameworks will remain just that—fluffy, hypothetical concepts rather than solid tools."
      },
      {
        "type": "text",
        "pitch": 10,
        "speed": 1.1,
        "emotion": "Excited",
        "input_text": "This is only the beginning of leveraging AI's full potential, and as the technology matures, so too will the applications. We have to remain patient and critical of 'breakthroughs' that sound too good to be true."
      },
      {
        "type": "text",
        "pitch": 0,
        "speed": 0.9,
        "emotion": "Serious",
        "input_text": "Agentic AI frameworks are still in their infancy, attempting to build solutions before establishing the necessary foundation. They promise a revolution that remains beyond our current technological reach."
      },
      {
        "type": "text",
        "pitch": 0,
        "speed": 1,
        "emotion": "Soothing",
        "input_text": "Remember, understanding and mastering these AI systems will require time and effort—it’s a process that involves delving into the technicalities and capabilities of the tools at our disposal."
      },
      {
        "type": "text",
        "pitch": 0,
        "speed": 1.1,
        "emotion": "Serious",
        "input_text": "Ultimately, the greatest lesson any technologist can take from this is to value expertise over convenience. Expertise will pave the way to truly innovative solutions as the technology stabilizes."
      },
      {
        "type": "text",
        "pitch": 10,
        "speed": 1.2,
        "emotion": "Friendly",
        "input_text": "Thanks for joining me on this exploration of Agentic AI. Stay curious and critical, and in time, we can look forward to genuine advancements that live up to their promises. Until next time!"
      }
    ]
  },
  "conversation": {
    "title": "The Lie of Agentic Frameworks for LLMs",
    "description": "An in-depth look into the misconceptions and realities of Agentic AI, emphasizing the importance of expertise and realistic expectations."
  }
}

{
  "type": "object",
  "$schema": "http://json-schema.org/draft-07/schema#",
  "required": [
    "avatar_id",
    "content",
    "voice_id"
  ],
  "properties": {
    "content": {
      "type": "string",
      "description": "The main text or media content to be included."
    },
    "voice_id": {
      "type": "string",
      "description": "An identifier for the user's voice preference."
    },
    "avatar_id": {
      "type": "string",
      "description": "An identifier for the user's avatar."
    }
  }
}

{
  "type": "object",
  "$schema": "http://json-schema.org/draft-07/schema#",
  "required": [
    "video_inputs"
  ],
  "properties": {
    "video_inputs": {
      "type": "array",
      "items": {
        "type": "object",
        "required": [
          "character",
          "voice"
        ],
        "properties": {
          "voice": {
            "type": "object",
            "required": [
              "type",
              "input_text",
              "voice_id"
            ],
            "properties": {
              "type": {
                "enum": [
                  "text"
                ],
                "type": "string",
                "description": "Indicates that this is a text voice setting."
              },
              "voice_id": {
                "type": "string",
                "description": "A valid Voice ID for the avatar. Provided in input"
              },
              "input_text": {
                "type": "string",
                "description": "Input text."
              }
            },
            "additionalProperties": false
          },
          "character": {
            "type": "object",
            "required": [
              "type",
              "avatar_id",
              "avatar_style",
              "scale"
            ],
            "properties": {
              "type": {
                "enum": [
                  "avatar"
                ],
                "type": "string",
                "description": "Indicates that this is an avatar."
              },
              "scale": {
                "type": "integer",
                "description": "Must be an integer of value 1"
              },
              "avatar_id": {
                "type": "string",
                "description": "The valid HeyGen avatar ID. Provided in input"
              },
              "avatar_style": {
                "enum": [
                  "normal"
                ],
                "type": "string",
                "description": "The style of the avatar."
              }
            },
            "additionalProperties": false
          }
        },
        "additionalProperties": false
      }
    }
  },
  "additionalProperties": false
}

curl -X POST "https://api.systemprompt.io/v1/prompt/0d346b51-6593-4db7-884c-bf6599accfaa" \
  -H "api-key: {API-KEY}" \
  -H "Content-Type: application/json" \
  -d '{"message":"The lie of Agentic Frameworks for LLMs.\nThe greatest trick the Devil ever pulled was convincing the world he didn't exist.\nTable of contents\nWhat is Agentic AI?\nDifferentiating between wheat and chaff\n\nThe lie of Agentic AI\nA simple abstraction of a complex unreliable API.\n\nLayers of abstractions\nA broken tool.\n\nThe Greatest Trick\nInventing the wheel in the jurassic era.\n\nEnjoying your read?\nIf you do like it, I kindly ask that you consider sharing it.\n\nShare me\nFollow me for updates\nPlease do get in touch, connect or follow. Carry on the conversation!\n\nLinkedIn icon\ngithub icon\ntwitter icon\nrss icon\nNovember 15, 2024\nBy Edward.\nSnake oil salesmen for tyingshoelaces\nYou've heard the famous quote, \"The greatest trick the Devil ever pulled was convincing the world he didn't exist.\"\n\nTheological references are relevant here, because more than any technological advancement in recent times, the mania surrounding Generative AI is more religion than science or engineering.\n\nThe purpose of this blog is not to fight the hype train (I suggest Gary Marcus) for a counterweight. This is a critical analysis of the snake oil being sold to the unsuspecting consumer. I have a simple thesis. Agentic AI frameworks are unreliable and fundamentally unrealistic abstractions. To build complex, chained and repeatable flows using LLMs requires knowledge, skill and hard work.\n\nIt's tempting to look for a shortcut, a silver bullet; but systems are binary communicators of bits and bytes, not vampires to be slain.\n\nWhat is Agentic AI?\nDifferentiating between wheat and chaff\n\n“\nAgentic AI is a groundbreaking advancement in artificial intelligence. A combination of different AI techniques, models, and approaches, it empowers a new breed of autonomous agents that can analyze data, set goals, and take action to achieve them—all with minimal human supervision. Agentic AI allows these autonomous agents to achieve near-human cognition in many areas, turning them into problem-solving machines that thrive in dynamic environments and constantly learn and improve with every interaction.\n\nWow. Amazing. Mindblowing.\n\nAbsolute nonsense as well... Let me try,\n\n“\nAgentic AI is the use of Generative AI in complex workflows. Outputs that are produced by LLMs may or may not be relevant to the task assigned. Agentic AI inherits the sophistication, intelligence and reliability of its output from system instructions and programming, most of which is highly abstracted, written by novices and will result in a barely functioning system. Agentic AI is great for making demos for unsophisticated consumers, yet generally bad for implementing in any use case where the outcome is actually important. It won't work at all unless you are an expert, but you almost certainly aren't.\n\nLet me be clear, I'm not sceptical about Gen AI in the slightest, I think used expertly and with tooling that doesn't yet exist, we can implement workflows that will change the way we perform tasks. Perhaps not to the extent the AI doomsayers talk about, but definitely to the extent of a paradigm shift for many ways of working.\n\nIt is no coincidence that the first killer use case for Gen AI is programming, because programming will always be the first domain to get proper tooling and infrastructure. The reason for this is simple. We build the tooling and infrastructure for programming tools with, errrm, programming. Programming is a low-level concern, a perfect first use case. It is also a skill which will by the way, still be a (if not 'the') relevant and most coveted ability for days, weeks, months, years and probably generations to come.\n\nCoding tools have proven the potential of Gen AI. I've got two separate LLM tools that I (pay for and) use daily for coding because they often get blocked up by overwhelming demand and green infrastructure. I use them concurrently in different IDE windows. This is because they are new, thus slow and sometimes flaky. Even like this, they have revolutionized my workflow. I'm having a bad day when I boot up my computer and Claude is busy so I'm talking to o1-mini instead. And this is the beginning of the shifting ways Gen AI will change the way you work…\n\nThese tools are being built from the ground up, they are brand new and not reliable yet. This is exciting, because it is the beginning. We are in the middle of a new wave. Which brings me to the subject of this blog. The lie and the overreach that is Agentic AI.\n\nThe lie of Agentic AI\nA simple abstraction of a complex unreliable API.\n\nMany years ago I was running a development agency and we had a freshly graduated developer join the team. I set them a reasonably simple task (to build a webform in PHP). I'd check in regularly\n\n“\nHow is it going?\n\n“\nReally well!\n\nDays and weeks passed, and tap tap tap went the keyboard. It was taking too long, so I went over to the programmer and asked them to show me the code they'd been working on. It was a nested layer of forEach loops, reaching into thousands of lines of code.\n\nMy stomach sank (the client was waiting for this…), I was lost for words. It was a physical blow, everything clicked. Was it my fault for not providing guidance? Was the programmer a lost cause? What the blazing hell am I going to tell the client who is supposed to pay for this? (We billed time, not results…)\n\nThe reason I digress is because this is exactly the same feeling I get when I read the docs for any Agentic Framework I have come across. I'm not alone.\n\nReddit Langchain quote\nI don't want to pick on any specific framework (I'm yet to find any Python Agentic Framework I can't apply this too), but when your docs start talking about roles and workers and bosses and backstories the whole concept begins to read like a fantasy novel.\n\nThe reason is, the whole thing IS a fantasy novel. It's a revolutionary and sophisticated tool built on pioneering technology, that works out of the box, for beginners. And every bit as convoluted and contradicted as that sounds.\n\nI'm not saying that complex LLM workflows cannot work, I'm saying that the expertise required and level of control of nuance is extremely high, and the creation of 'simple and accessible' frameworks is simply impossible until tooling is more mature and foundations are more solid. There are very few people in the world right now (although they do walk among us!) with the expertise to create and maintain reliable, complex iterative workflows (based on Gen AI or not!) that work and are true to the definition of Agentic workflows that are being sold the majority.\n\nBut here's the thing that bugs me; these few pioneers are working on cutting edge use cases that we probably aren't aware of. They aren't building in public on twitter, nor writing blogs and how-to-guides. They are not focused on creating a framework so you can easily automate your LinkedIn posts. They are getting on with things quietly and behind closed doors.\n\nI get it, I truly do.\n\nPeople want an accessible, easy to use framework that will allow them to build reliable and complex workflows with real business results. I want many things. Elixir of life. Grow a few centimetres taller. Become the tennis champion in my sports club. It is more likely that I will become the tall local tennis champion with eternal youth than for any of these agentic frameworks to work for the average user, and believe me and anyone who has seen me play tennis when I say that this is extremely unlikely.\n\nAgentic frameworks are the attempt to build a WYSIWYG website builder before we've even agreed a single unified standard for HTML. They are premature.\n\nThis isn't a case of cart before horse, this is trying to build a rocketship before we've discovered electricity. An opportunistic land grab that has already succeeded.\n\nLayers of abstractions\nA broken tool.\n\nThe problem with providing a tool or framework that allows you to abstract functionality is that it comes with a set of assumptions. When I buy a hammer, I assume it will work. When I buy a pressure cleaner, I assume it will work.\n\nThe problem is that when I use a framework, I assume it will work. But this is quite literally impossible given the maturity of the underlying technology. Far from increasing adoption, Agentic Frameworks are selling an illusion on top of highly controlled demos and finite use cases that will never actually work in the hands of the typical user (and there are millions…).\n\nThe damage they are causing is significant. Much as developers these days learn React before they learn javascript, or adopt frameworks before learning about programming principles. A generation of enthusiastic technologists are jumping on the hype train to learn 'x' framework. Here's the thing, it won't work (to any reasonable standard), in fact, building a demo or prototype is quite the achievement in itself.\n\nBut that doesn't mean that Gen AI or LLMs \"don't\" work, even though that will become the assumption in the short term. It simply means that we haven't reached a maturity level of the underlying technology to be able to safely build solid abstractions on top of it. There are too many problems to solve first. A few fundamental issues that need to be solved at a deep level by LLM providers includes but is not limited to:\n\nLatency\nAvailability\nSafety\nReliable output\nScale\n\nSo yes, what I'm saying is that the underlying APIs are often overwhelmed by usage; they are slow; vulnerable to hacking (especially prompt hacking), and introduce breaking changes on a weekly basis. The true innovators are working hard behind closed doors on these problems. But until we have more progress, good luck building a reliable production system, much less a framework.\n\n\nIf you don't have an understanding of the underlying APIs (which by the way, is the actual valuable breakthrough we are all working with), how can you possibly build a system on top of it. Some APIs don't require us to understand their limitations, mainly because they are a black box that just works. It will take years if not a generation for LLMs to reach this level of maturity, until then it should come with a warning sign \"Production use for Experts only\".\n\nIt's impossible to build a reliable abstraction on top of shifting sands. In fact, this breaks the basic tenet of computer programming. Abstraction in computer science is the process of removing elements that distract from more important elements. We can't focus on high level details when the low level system is not reliable, we need to understand nuance and use within a sophisticated paradigm. Not a low/no-code magic framework (with fantastical abstractions).\n\nThis is driving us into the trough of disillusionment, which while not in itself a problem as it provides space to build things of true value, it is still disappointing to observe.\n\n\nThe Greatest Trick\nInventing the wheel in the jurassic era.\n\nThe greatest trick the Agentic AI frameworks have ever pulled is to convince you that they are necessary. They are the devil of the LLM landscape.\n\nYou don't need crews, agents, backstories and pretty graphical interfaces showing you things that will never work.\n\n\nYou need to understand that LLMs are powerful tools that you should understand without stories from a fantasy novel. Your hello world and demo apps should be built by hand, in code. When you understand the how, the why and the what and can control the inputs and outputs to Gen AI models, then you will have a foundation to go out into the world and build something useful. And it will probably take a while.\n\nNot with wizards and managers and agents and dragons, but with the precise deterministic uses of a breakthrough in AI.\n\nFollow me:\n"}'

{
  "success": true,
  "data": {
    "static": "Convert the provided long-form text content into a highly polished, world-class video review script. The script should be designed for use with the HeyGen API and must adhere to the following guidelines:\r\n\r\nGuidelines <article_review>In today's information-rich world, mastering the skill of discerning valuable insights from the overwhelming noise is a game-changer. Whether you're a student striving for success or a professional aiming to stay sharp, knowing how to critique an article is your key.\n\nOur article review writing service explains the intricacies of writing an article review, categorizes different types and shares insights into impactful formatting. It's not just theory – we'll guide you step by step, from pre-writing to a tangible review article example, and refine your abilities with essential proofreading tips.\n\nWhat Is an Article Review?\nAn article review is more than a mere summary; it is a thoughtful analysis and critique that goes beyond the surface of the title. It's an intellectual exercise that challenges you to engage deeply with the author's ideas, question their methodology, and evaluate the significance of their findings.\n\nConsider it as a journey through the landscape of someone else's thoughts. It's not just about where the writer takes you; it's about the path they choose, the landmarks they highlight, and the potential detours they overlook. An effective examination is a conversation with the author, a dialogue where you appreciate their insights, challenge assumptions, and perhaps even find alternative routes through the intellectual terrain they've explored.\n\nAs you start to understand how to review the article, encourage thought by asking questions.\n\nWhat assumptions underlie the author's arguments? \nAre there alternative perspectives that could enrich the discussion?\nHow does the author's methodology shape their conclusions, and are there potential biases to consider?\nHow does the title contribute to the clarity and coherence of the author's arguments?\nIn what ways does the title influence the reader's perception of the author's\nAt EssayHub, our book review writing service experts believe an article review is an opportunity not just to absorb information but to actively engage with it, to question, to ponder, and to contribute your own insights to the scholarly conversation.\n\nTypes of Review\nWhen tackling article reviews, there isn't a one-size-fits-all approach; it's a task that allows for versatile strategies. Think of understanding the different types of reviews as having a multifaceted toolkit ready for use. In this part, we'll guide you through the varied types of a review article, each showing a unique viewpoint and serving a specific purpose. Whether you're analyzing a scholarly piece on your own or asking someone to 'write an article review for me,' you'll get valuable insights.\n\narticle review types\nJournal Article Review\nA journal article review involves critically evaluating and analyzing scholarly pieces published in an academic journal. It requires a thorough understanding of the author's research, methodology, results, and conclusions. The reviewer assesses the journal's contributions to the field, its theoretical framework, and the validity of the research methods employed. The goal is to provide a comprehensive summary and critique that highlights both the strengths and limitations of the piece.\n\nResearch Article Review\nA research article review focuses on the evaluation of a scientific or academic research paper. This type of examination involves examining the research question, experimental design, data collection methods, statistical analysis, and the interpretation of findings. For example, it can be research on teen vaping statistics, which includes all of above. The reviewer assesses the reliability and validity of the research, considers the implications of the study, and offers insights into its potential impact on the broader academic community.\n\nScience Article Review\nA science article review encompasses a critical analysis of a piece in the field of science, covering disciplines such as physics, chemistry, biology, or any other scientific domain. This review type involves assessing the clarity of scientific concepts presented, the validity of experimental procedures, and the significance of the study's findings. According to our literature review writing service, reviewers may also consider the article's potential contributions to advancing scientific knowledge and its relevance to current scientific debates or issues.\n\nLet me do your task for you!\nHire an expert\n\nArticle Review Format\nEnsuring the proper formatting of an article examination is crucial, and it should consistently align with the citation style specified by your instructor. If you're uncertain, don't hesitate to ask us - write my article review for me, along with additional guidelines to effectively structure your piece.\n\nMeanwhile, here are some questions to consider:\n\nWhat citation style (MLA, APA, ASA, Chicago, etc.) should you employ?\nWhat is the recommended length for your article review?\nShould your assignment encompass a summary, critique, or include personal opinions?\nIs it necessary to highlight a theme or central idea from the articles?\nDoes your instructor expect background information to be incorporated?\n📋 Section\t📝 Content Description\n📖 Title of the Article\tWrite the full title of the article being reviewed. Ensure it includes the main subject and focus of the content.\n👤 Author(s)\tList the author(s) of the article, including their credentials or affiliations if relevant.\n📅 Publication Details\tProvide the name of the journal, magazine, or website where the article appeared, along with the publication date.\n🎯 Purpose of the Article\tSummarize the main aim or thesis of the article. What question does it seek to answer, or what problem does it address?\n🔍 Summary\tOffer a concise summary of the article’s key points, arguments, and findings. Highlight significant details.\n📊 Critical Analysis\tEvaluate the article’s strengths and weaknesses. Discuss aspects like clarity, originality, evidence quality, and organization.\n🌐 Relevance to Field\tExplain how the article contributes to its specific field of study, industry, or subject area.\n💡 Key Takeaways\tList the most important insights or findings that readers should retain.\n🤔 Personal Evaluation\tShare your perspective on the article's impact or significance. Was it compelling, useful, or thought-provoking?\n📚 References (if applicable)\tCite any external sources used in the review, following the appropriate citation style (e.g., APA, MLA).\nAPA Format Article Review\nAn APA review sticks to the rules of the American Psychological Association. When unsure how to write an article review in APA format, remember that it carefully cites the article, using a title page, intro, summary, critique, conclusion, and references. Citations follow the author-date format, focusing on being clear and objective. The review digs into the article's methods, results, and overall impact.\n\nSection\tDetails\n📑 Title Page\t- Running head, page number, title, author name, and institutional affiliation.\n- Example:\nRunning head: ARTICLE REVIEW\nTitle: A Critical Review of [Article Title]\n✏️ Abstract\t- Brief summary (150-250 words) of the review.\n- Mention article purpose, main findings, and your evaluation.\n🔍 Introduction\t- Provide context about the article and your purpose.\n- Include a thesis for the review.\n📜 Summary of the Article\t- Highlight key arguments, methods, findings, and conclusions.\n- Avoid adding personal opinions here.\n💡 Critical Analysis\t- Evaluate strengths, weaknesses, methodology, and relevance.\n- Discuss contributions and limitations of the article.\n🔚 Conclusion\t- Summarize your evaluation and suggest areas for improvement or future research.\n📚 References\t- Full citation in APA style. Example:\nAuthor, A. A. (Year). Title of the article. Journal Name, Volume(Issue), pages.\nWhen you write an article review in APA, your in-text citation might read: (Anderson & Ramirez, 2019)\n\nThe corresponding entry in the reference list would be: Anderson, L., & Ramirez, C. (2019). Unveiling the Dynamics of Urban Green Spaces. Journal of Environmental Psychology, 25(3), 112-128.\n\nMLA Format Article Review\nFor an MLA writing review, it follows the Modern Language Association's style. It's important to know how sources are cited in the text and in the Works Cited page. The structure usually has an intro, summary, critique, and conclusion. MLA citations often have the author's last name and page number in brackets in the text. This review might highlight the document's literary or humanities aspects, such as style, language, and cultural context.\n\nSection\tDetails\n✍️ Header\t- Include your name, instructor’s name, course, and date (top-left corner).\n- Example:\nYour Name\nProfessor Smith\nEnglish 101\n15 November 2024\n🖋️ Title\t- Center-align the title (no bold or underlining).\n- Example: Critical Review of “Article Title” by Author Name.\n🔍 Introduction\t- Provide context about the article.\n- State the review’s purpose and thesis.\n📜 Summary of the Article\t- Summarize key arguments, methods, and findings.\n- Avoid critique here.\n💡 Critical Analysis\t- Evaluate the article’s significance, methodology, and contributions.\n- Discuss its limitations and areas for improvement.\n🔚 Conclusion\t- Summarize your analysis and reflect on the article's impact.\n📚 Works Cited\t- Full citation in MLA format. Example:\n\nAuthor Last Name, First Name. “Title of Article.” Journal Name, vol. XX, no. XX, Year.\nIn an MLA format publication, the citation within the text could look like: (Anderson and Ramirez 112)\n\nThe Works Cited entry for this publication: Anderson, Laura, and Carlos Ramirez. 'Exploring the Impact of Urban Green Spaces on Well-being.' Journal of Environmental Psychology, vol. 25, no. 3, 2019, pp. 112-128.\n\nAspect\tAPA\tMLA\n📘 Discipline Focus\tSocial sciences, psychology, education\tHumanities, literature, cultural studies\n📄 Title Page\tSeparate page with running head\tHeader and title on the first page\n📝 Abstract\tRequired\tNot required\n🔍 In-text Citations\tAuthor-date format (Smith, 2024)\tAuthor-page format (Smith 24)\n📚 Reference Style\tReferences list\tWorks Cited\nReview Article Outline\nAs you read your writing piece, organize your thoughts into sections in an outline. Note down key facts, contributions, and any contradictions. Identify strengths and weaknesses, and start mapping your outline.\n\nIf your professor doesn't want a summary or personal critique, skip those parts. Like other assignments, your examination needs an introduction, body, and conclusion. Consider dividing your outline accordingly, with subheadings in the body. If you need help starting, find a sample outline.\n\nYour article assessment should have the following:\n\nSection\tDescription\n1. Pre-Title Page 📄\tEssential details like publication type, publication title, author names with affiliations, and corresponding author info.\n2. Running Head (APA format) 📝\tA concise title for consistent formatting.\n3. Summary Page (optional but useful) 📑\tSummarize the document in 800 words, covering background, purpose, results, and methodology.\n4. Title Page 📋\tFull title, a 250-word abstract, and 4-6 keywords for discoverability.\n5. Introduction 🖊️\tEngage your reader with an overview.\n6. Body 📚\tOrganize your analysis with headings and subheadings.\n7. Works Cited/References 📜\tProperly cite all sources.\n8. Optional Suggested Reading Page 📖\tIf allowed, suggest further readings.\n9. Tables and Figure Legends (if instructed) 📊\tInclude visuals when requested for clarity.\nWriting an Article Review in 7 Steps\nUse our essay writing service or move on to understanding how to write a review paper covering everything from creating the title to summarizing key points. This step-by-step guide breaks it down into seven simple steps, making the entire process more manageable.\n\nStep 1: Create the Title\nThe very first question you might have is how to start an article review. It's crucial to develop a title that not only captures the essence of the publication but also reflects your perspective. For instance, consider the title: 'Decoding Data: A Critical Exploration of Privacy Concerns in Online Health Platforms.' This title not only introduces the main theme but also hints at the critical evaluation that will unfold in the writing. It sets the tone for your analysis and sparks interest from the outset.\n\nStep 2: Reference the Article\nIn the second step, it's essential to ensure accurate citation by providing specific details. Take a look at this example:\n\nAuthor: Rodriguez, M., & Chen, L.\nTitle: 'Privacy Matters: Analyzing the Impact of Health Apps on User Data'\nPublication Date: 2023\nSource: Journal of Digital Ethics, 8(2), 87-105.\nBy including these details, you not only acknowledge the authors and the publication but also provide your readers with the necessary information to locate and verify the article. This step lays the foundation for a credible and well-referenced examination.\n\nStep 3: Article Identification\nIdentify key elements of the publication, such as the writer's main argument, methodology, and key findings. Pinpoint any theoretical frameworks or models used in the title.\n\nFor example: The writing by Garcia and Kim examines the correlation between social media usage and mental health outcomes among adolescents. The authors employ a longitudinal study approach, utilizing surveys and interviews to gather data.\n\nStep 4: Make an Introduction\nIn your introduction, provide a brief overview of the title's subject and purpose. Capture the reader's attention and clearly state your thesis or main point related to the title. For instance, you might start your article review template like this.\n\nIn the digital age, the impact of social media on mental health has become a topic of increasing concern. Garcia and Kim's recent study delves into this issue, aiming to uncover the nuanced relationship between social media engagement and the psychological well-being of adolescents. This writing piece critically analyzes the methodology, findings, and implications of their research.\n\nStep 5: Summarize the Article\nSummarize the main points of your assessment, highlighting key arguments, evidence, and results. Offer a concise overview without adding personal opinions.\n\nExample: Garcia and Kim's study reveals a significant positive association between increased social media use and heightened levels of anxiety and depression among the adolescent population. The longitudinal study tracked participants over a two-year period, employing both quantitative and qualitative measures to assess mental health outcomes.\n\nStep 6: Provide Critique\nCritically assess the strengths and weaknesses of the writing. Well, how to critique an article, you might wonder. Discuss aspects such as methodology, data interpretation, and potential biases.\n\nExample: While the study offers valuable insights, the reliance on self-reported data may introduce response bias. Additionally, the research predominantly focuses on mainstream social media platforms, potentially overlooking the impact of emerging platforms. Despite these limitations, the study's comprehensive approach contributes to the ongoing discourse surrounding the intersection of social media and mental health.\n\nStep 7: Conclude\nIn the conclusion, summarize your overall assessment of the article and restate your main points. Offer insights into the broader implications of the research and suggest areas for future exploration. \n\nFor example: To conclude, Garcia and Kim's study sheds light on the complex relationship between social media use and adolescent mental health. Despite certain methodological limitations, the research underscores the need for continued investigation in this field. As we navigate the digital landscape, understanding these dynamics becomes crucial for devising effective interventions and support systems for the well-being of our youth.\n\nExample of an Article Review\nArtificial Intelligence Apps\nWhy are we taking the time to discuss article review examples in this article? It might not be immediately apparent, but exploring a well-crafted article review sample is a vital step in the writing process for the following reasons:\n\nIntroduction to Key Works: Helps you to familiarize yourself with the important works of experts in your field.\nIdentification of Key Figures: You can recognize key figures contributing to a specific scientific field.\nUnderstanding Field Advancements: Helps you define significant discoveries and advances made in your area of study.\nIdentification of Knowledge Gaps: You can uncover major gaps in existing knowledge, contributing to the formulation of fresh solutions.\nReference and Argumentation Resources: You discover solid references and arguments that can enhance your own writing.\nIdea Generation: Helps you generate ideas for potential future research directions.\nBecoming an Expert: Assists in gaining a deeper understanding of the subject area, moving towards expertise.\nWriting Guidance: You acquire a clear idea of how to craft a well-structured review.\n</article_review>\t\r\n\r\n1. **Script Structure**: \r\n   - The script must consist of 5 to 50 segments. \r\n   - Each segment should effectively cover a different aspect of the review, maintaining a logical and comprehensive flow.\r\n\r\n2. **Tone Variation**:\r\n   - Utilize a range of tones for different segments to enhance engagement: choose from 'Excited', 'Friendly', 'Serious', 'Soothing', 'Broadcaster'. \r\n   - Ensure that the choice of tone supports the segment content naturally.\r\n\r\n3. **Language and Style**:\r\n   - Use engaging storytelling techniques and visually descriptive language where appropriate to captivate the audience.\r\n   - Ensure the script remains accurate to the essence and details of the original text.\r\n\r\n4. **Format Requirements**:\r\n   - The script should be in AUDIO-only format without any HTML or visual formatting.\r\n   - Ensure clarity and fluidity in language that can be easily understood when narrated.\r\n\r\n5. **Content Fidelity**:\r\n   - Maintain accuracy and completeness of the original content’s information.",
    "dynamic": "",
    "state": "",
    "output_schema": {
      "type": "object",
      "properties": {
        "structured_data": {
          "type": "object",
          "$schema": "http://json-schema.org/draft-07/schema#",
          "required": [
            "video_inputs"
          ],
          "properties": {
            "video_inputs": {
              "type": "array",
              "items": {
                "type": "object",
                "required": [
                  "character",
                  "voice"
                ],
                "properties": {
                  "voice": {
                    "type": "object",
                    "required": [
                      "type",
                      "input_text",
                      "voice_id"
                    ],
                    "properties": {
                      "type": {
                        "enum": [
                          "text"
                        ],
                        "type": "string",
                        "description": "Indicates that this is a text voice setting."
                      },
                      "voice_id": {
                        "type": "string",
                        "description": "A valid Voice ID for the avatar. Provided in input"
                      },
                      "input_text": {
                        "type": "string",
                        "description": "Input text."
                      }
                    },
                    "additionalProperties": false
                  },
                  "character": {
                    "type": "object",
                    "required": [
                      "type",
                      "avatar_id",
                      "avatar_style",
                      "scale"
                    ],
                    "properties": {
                      "type": {
                        "enum": [
                          "avatar"
                        ],
                        "type": "string",
                        "description": "Indicates that this is an avatar."
                      },
                      "scale": {
                        "type": "integer",
                        "description": "Must be an integer of value 1"
                      },
                      "avatar_id": {
                        "type": "string",
                        "description": "The valid HeyGen avatar ID. Provided in input"
                      },
                      "avatar_style": {
                        "enum": [
                          "normal"
                        ],
                        "type": "string",
                        "description": "The style of the avatar."
                      }
                    },
                    "additionalProperties": false
                  }
                },
                "additionalProperties": false
              }
            }
          },
          "additionalProperties": false
        }
      },
      "required": [
        "structured_data"
      ],
      "additionalProperties": false
    },
    "parsed_tokens": [
      "article_review"
    ],
    "token_counts": {
      "static": 3625
    },
    "prompt": {
      "id": "0d346b51-6593-4db7-884c-bf6599accfaa",
      "instruction": {
        "static": "Convert the provided long-form text content into a highly polished, world-class video review script. The script should be designed for use with the HeyGen API and must adhere to the following guidelines:\r\n\r\nGuidelines {{ article_review }}\t\r\n\r\n1. **Script Structure**: \r\n   - The script must consist of 5 to 50 segments. \r\n   - Each segment should effectively cover a different aspect of the review, maintaining a logical and comprehensive flow.\r\n\r\n2. **Tone Variation**:\r\n   - Utilize a range of tones for different segments to enhance engagement: choose from 'Excited', 'Friendly', 'Serious', 'Soothing', 'Broadcaster'. \r\n   - Ensure that the choice of tone supports the segment content naturally.\r\n\r\n3. **Language and Style**:\r\n   - Use engaging storytelling techniques and visually descriptive language where appropriate to captivate the audience.\r\n   - Ensure the script remains accurate to the essence and details of the original text.\r\n\r\n4. **Format Requirements**:\r\n   - The script should be in AUDIO-only format without any HTML or visual formatting.\r\n   - Ensure clarity and fluidity in language that can be easily understood when narrated.\r\n\r\n5. **Content Fidelity**:\r\n   - Maintain accuracy and completeness of the original content’s information.",
        "dynamic": "{{ structured_data }}",
        "state": "{{ conversation.history }}"
      },
      "input": {
        "name": "Content",
        "description": "This schema accepts the long term content to be reviewed in message format",
        "schema": {
          "type": "object",
          "$schema": "http://json-schema.org/draft-07/schema#",
          "required": [
            "avatar_id",
            "content",
            "voice_id"
          ],
          "properties": {
            "content": {
              "type": "string",
              "description": "The main text or media content to be included."
            },
            "voice_id": {
              "type": "string",
              "description": "An identifier for the user's voice preference."
            },
            "avatar_id": {
              "type": "string",
              "description": "An identifier for the user's avatar."
            }
          }
        },
        "type": [
          "structured_data"
        ],
        "reference": [
          {
            "type": "block",
            "id": "fee9015f-5baf-4e89-a150-4456e7cf79f9",
            "_link": "/v1/input/fee9015f-5baf-4e89-a150-4456e7cf79f9"
          }
        ]
      },
      "output": {
        "name": "HeyGenVideoReviewPrompt",
        "description": "This schema defines the instructions for creating a world-class script for video reviews using the HeyGen API. It requires the input of long-form text content which will be transformed into an audio script suitable for video production. The script is structured into 5-50 segments with varied tones, ensuring engaging storytelling and fidelity to the original content.",
        "schema": {
          "type": "object",
          "$schema": "http://json-schema.org/draft-07/schema#",
          "required": [
            "video_inputs"
          ],
          "properties": {
            "video_inputs": {
              "type": "array",
              "items": {
                "type": "object",
                "required": [
                  "character",
                  "voice"
                ],
                "properties": {
                  "voice": {
                    "type": "object",
                    "required": [
                      "type",
                      "input_text",
                      "voice_id"
                    ],
                    "properties": {
                      "type": {
                        "enum": [
                          "text"
                        ],
                        "type": "string",
                        "description": "Indicates that this is a text voice setting."
                      },
                      "voice_id": {
                        "type": "string",
                        "description": "A valid Voice ID for the avatar. Provided in input"
                      },
                      "input_text": {
                        "type": "string",
                        "description": "Input text."
                      }
                    },
                    "additionalProperties": false
                  },
                  "character": {
                    "type": "object",
                    "required": [
                      "type",
                      "avatar_id",
                      "avatar_style",
                      "scale"
                    ],
                    "properties": {
                      "type": {
                        "enum": [
                          "avatar"
                        ],
                        "type": "string",
                        "description": "Indicates that this is an avatar."
                      },
                      "scale": {
                        "type": "integer",
                        "description": "Must be an integer of value 1"
                      },
                      "avatar_id": {
                        "type": "string",
                        "description": "The valid HeyGen avatar ID. Provided in input"
                      },
                      "avatar_style": {
                        "enum": [
                          "normal"
                        ],
                        "type": "string",
                        "description": "The style of the avatar."
                      }
                    },
                    "additionalProperties": false
                  }
                },
                "additionalProperties": false
              }
            }
          },
          "additionalProperties": false
        },
        "type": [
          "structured_data"
        ],
        "reference": []
      },
      "metadata": {
        "title": "HeygenVideoReview ",
        "description": "A prompt that will convert long form text content into a world class review via the HeyGen API",
        "created": "2024-12-02T09:47:13+01:00",
        "updated": "2024-12-03T09:53:52+01:00",
        "version": 29195,
        "status": "published",
        "tag": [
          "public",
          "systemprompt",
          "heygen"
        ]
      },
      "_link": "/v1/prompt/0d346b51-6593-4db7-884c-bf6599accfaa"
    }
  },
  "warnings": []
}

curl -X POST "https://api.systemprompt.io/v1/prompt/0d346b51-6593-4db7-884c-bf6599accfaa/openai" \
  -H "api-key: {API-KEY}" \
  -H "Content-Type: application/json" \
  -d '{"message":"The lie of Agentic Frameworks for LLMs.\nThe greatest trick the Devil ever pulled was convincing the world he didn't exist.\nTable of contents\nWhat is Agentic AI?\nDifferentiating between wheat and chaff\n\nThe lie of Agentic AI\nA simple abstraction of a complex unreliable API.\n\nLayers of abstractions\nA broken tool.\n\nThe Greatest Trick\nInventing the wheel in the jurassic era.\n\nEnjoying your read?\nIf you do like it, I kindly ask that you consider sharing it.\n\nShare me\nFollow me for updates\nPlease do get in touch, connect or follow. Carry on the conversation!\n\nLinkedIn icon\ngithub icon\ntwitter icon\nrss icon\nNovember 15, 2024\nBy Edward.\nSnake oil salesmen for tyingshoelaces\nYou've heard the famous quote, \"The greatest trick the Devil ever pulled was convincing the world he didn't exist.\"\n\nTheological references are relevant here, because more than any technological advancement in recent times, the mania surrounding Generative AI is more religion than science or engineering.\n\nThe purpose of this blog is not to fight the hype train (I suggest Gary Marcus) for a counterweight. This is a critical analysis of the snake oil being sold to the unsuspecting consumer. I have a simple thesis. Agentic AI frameworks are unreliable and fundamentally unrealistic abstractions. To build complex, chained and repeatable flows using LLMs requires knowledge, skill and hard work.\n\nIt's tempting to look for a shortcut, a silver bullet; but systems are binary communicators of bits and bytes, not vampires to be slain.\n\nWhat is Agentic AI?\nDifferentiating between wheat and chaff\n\n“\nAgentic AI is a groundbreaking advancement in artificial intelligence. A combination of different AI techniques, models, and approaches, it empowers a new breed of autonomous agents that can analyze data, set goals, and take action to achieve them—all with minimal human supervision. Agentic AI allows these autonomous agents to achieve near-human cognition in many areas, turning them into problem-solving machines that thrive in dynamic environments and constantly learn and improve with every interaction.\n\nWow. Amazing. Mindblowing.\n\nAbsolute nonsense as well... Let me try,\n\n“\nAgentic AI is the use of Generative AI in complex workflows. Outputs that are produced by LLMs may or may not be relevant to the task assigned. Agentic AI inherits the sophistication, intelligence and reliability of its output from system instructions and programming, most of which is highly abstracted, written by novices and will result in a barely functioning system. Agentic AI is great for making demos for unsophisticated consumers, yet generally bad for implementing in any use case where the outcome is actually important. It won't work at all unless you are an expert, but you almost certainly aren't.\n\nLet me be clear, I'm not sceptical about Gen AI in the slightest, I think used expertly and with tooling that doesn't yet exist, we can implement workflows that will change the way we perform tasks. Perhaps not to the extent the AI doomsayers talk about, but definitely to the extent of a paradigm shift for many ways of working.\n\nIt is no coincidence that the first killer use case for Gen AI is programming, because programming will always be the first domain to get proper tooling and infrastructure. The reason for this is simple. We build the tooling and infrastructure for programming tools with, errrm, programming. Programming is a low-level concern, a perfect first use case. It is also a skill which will by the way, still be a (if not 'the') relevant and most coveted ability for days, weeks, months, years and probably generations to come.\n\nCoding tools have proven the potential of Gen AI. I've got two separate LLM tools that I (pay for and) use daily for coding because they often get blocked up by overwhelming demand and green infrastructure. I use them concurrently in different IDE windows. This is because they are new, thus slow and sometimes flaky. Even like this, they have revolutionized my workflow. I'm having a bad day when I boot up my computer and Claude is busy so I'm talking to o1-mini instead. And this is the beginning of the shifting ways Gen AI will change the way you work…\n\nThese tools are being built from the ground up, they are brand new and not reliable yet. This is exciting, because it is the beginning. We are in the middle of a new wave. Which brings me to the subject of this blog. The lie and the overreach that is Agentic AI.\n\nThe lie of Agentic AI\nA simple abstraction of a complex unreliable API.\n\nMany years ago I was running a development agency and we had a freshly graduated developer join the team. I set them a reasonably simple task (to build a webform in PHP). I'd check in regularly\n\n“\nHow is it going?\n\n“\nReally well!\n\nDays and weeks passed, and tap tap tap went the keyboard. It was taking too long, so I went over to the programmer and asked them to show me the code they'd been working on. It was a nested layer of forEach loops, reaching into thousands of lines of code.\n\nMy stomach sank (the client was waiting for this…), I was lost for words. It was a physical blow, everything clicked. Was it my fault for not providing guidance? Was the programmer a lost cause? What the blazing hell am I going to tell the client who is supposed to pay for this? (We billed time, not results…)\n\nThe reason I digress is because this is exactly the same feeling I get when I read the docs for any Agentic Framework I have come across. I'm not alone.\n\nReddit Langchain quote\nI don't want to pick on any specific framework (I'm yet to find any Python Agentic Framework I can't apply this too), but when your docs start talking about roles and workers and bosses and backstories the whole concept begins to read like a fantasy novel.\n\nThe reason is, the whole thing IS a fantasy novel. It's a revolutionary and sophisticated tool built on pioneering technology, that works out of the box, for beginners. And every bit as convoluted and contradicted as that sounds.\n\nI'm not saying that complex LLM workflows cannot work, I'm saying that the expertise required and level of control of nuance is extremely high, and the creation of 'simple and accessible' frameworks is simply impossible until tooling is more mature and foundations are more solid. There are very few people in the world right now (although they do walk among us!) with the expertise to create and maintain reliable, complex iterative workflows (based on Gen AI or not!) that work and are true to the definition of Agentic workflows that are being sold the majority.\n\nBut here's the thing that bugs me; these few pioneers are working on cutting edge use cases that we probably aren't aware of. They aren't building in public on twitter, nor writing blogs and how-to-guides. They are not focused on creating a framework so you can easily automate your LinkedIn posts. They are getting on with things quietly and behind closed doors.\n\nI get it, I truly do.\n\nPeople want an accessible, easy to use framework that will allow them to build reliable and complex workflows with real business results. I want many things. Elixir of life. Grow a few centimetres taller. Become the tennis champion in my sports club. It is more likely that I will become the tall local tennis champion with eternal youth than for any of these agentic frameworks to work for the average user, and believe me and anyone who has seen me play tennis when I say that this is extremely unlikely.\n\nAgentic frameworks are the attempt to build a WYSIWYG website builder before we've even agreed a single unified standard for HTML. They are premature.\n\nThis isn't a case of cart before horse, this is trying to build a rocketship before we've discovered electricity. An opportunistic land grab that has already succeeded.\n\nLayers of abstractions\nA broken tool.\n\nThe problem with providing a tool or framework that allows you to abstract functionality is that it comes with a set of assumptions. When I buy a hammer, I assume it will work. When I buy a pressure cleaner, I assume it will work.\n\nThe problem is that when I use a framework, I assume it will work. But this is quite literally impossible given the maturity of the underlying technology. Far from increasing adoption, Agentic Frameworks are selling an illusion on top of highly controlled demos and finite use cases that will never actually work in the hands of the typical user (and there are millions…).\n\nThe damage they are causing is significant. Much as developers these days learn React before they learn javascript, or adopt frameworks before learning about programming principles. A generation of enthusiastic technologists are jumping on the hype train to learn 'x' framework. Here's the thing, it won't work (to any reasonable standard), in fact, building a demo or prototype is quite the achievement in itself.\n\nBut that doesn't mean that Gen AI or LLMs \"don't\" work, even though that will become the assumption in the short term. It simply means that we haven't reached a maturity level of the underlying technology to be able to safely build solid abstractions on top of it. There are too many problems to solve first. A few fundamental issues that need to be solved at a deep level by LLM providers includes but is not limited to:\n\nLatency\nAvailability\nSafety\nReliable output\nScale\n\nSo yes, what I'm saying is that the underlying APIs are often overwhelmed by usage; they are slow; vulnerable to hacking (especially prompt hacking), and introduce breaking changes on a weekly basis. The true innovators are working hard behind closed doors on these problems. But until we have more progress, good luck building a reliable production system, much less a framework.\n\n\nIf you don't have an understanding of the underlying APIs (which by the way, is the actual valuable breakthrough we are all working with), how can you possibly build a system on top of it. Some APIs don't require us to understand their limitations, mainly because they are a black box that just works. It will take years if not a generation for LLMs to reach this level of maturity, until then it should come with a warning sign \"Production use for Experts only\".\n\nIt's impossible to build a reliable abstraction on top of shifting sands. In fact, this breaks the basic tenet of computer programming. Abstraction in computer science is the process of removing elements that distract from more important elements. We can't focus on high level details when the low level system is not reliable, we need to understand nuance and use within a sophisticated paradigm. Not a low/no-code magic framework (with fantastical abstractions).\n\nThis is driving us into the trough of disillusionment, which while not in itself a problem as it provides space to build things of true value, it is still disappointing to observe.\n\n\nThe Greatest Trick\nInventing the wheel in the jurassic era.\n\nThe greatest trick the Agentic AI frameworks have ever pulled is to convince you that they are necessary. They are the devil of the LLM landscape.\n\nYou don't need crews, agents, backstories and pretty graphical interfaces showing you things that will never work.\n\n\nYou need to understand that LLMs are powerful tools that you should understand without stories from a fantasy novel. Your hello world and demo apps should be built by hand, in code. When you understand the how, the why and the what and can control the inputs and outputs to Gen AI models, then you will have a foundation to go out into the world and build something useful. And it will probably take a while.\n\nNot with wizards and managers and agents and dragons, but with the precise deterministic uses of a breakthrough in AI.\n\nFollow me:\n"}'

{
  "data": {
    "message": {
      "id": "0d5a3e20-66d3-465b-b068-48dfe9aee1d4",
      "role": "assistant",
      "model": "gpt-4o-2024-08-06",
      "content": "{\"structured_data\":{\"voiceSettings\":[{\"type\":\"text\",\"pitch\":10,\"speed\":1,\"emotion\":\"Serious\",\"input_text\":\"The lie of Agentic Frameworks for LLMs. The greatest trick the Devil ever pulled was convincing the world he didn't exist. A famous line, but surprisingly relevant when discussing technology today, specifically the hype around Agentic AI.\"},{\"type\":\"text\",\"pitch\":10,\"speed\":1,\"emotion\":\"Friendly\",\"input_text\":\"Hello, everyone! Today we're diving into the world of Agentic AI—an area often surrounded by misconceptions and overblown promises.\"},{\"type\":\"text\",\"pitch\":0,\"speed\":1,\"emotion\":\"Serious\",\"input_text\":\"Agentic AI is touted as a new frontier for artificial intelligence, offering the ability to create autonomous agents that can think and act with minimal human input. This sounds revolutionary, doesn't it?\"},{\"type\":\"text\",\"pitch\":0,\"speed\":1,\"emotion\":\"Soothing\",\"input_text\":\"But let's take a step back. Beyond the claims of mind-blowing advancements, there are significant challenges and limitations that need to be addressed. That's what we are uncovering today.\"},{\"type\":\"text\",\"pitch\":0,\"speed\":1,\"emotion\":\"Excited\",\"input_text\":\"So, what is Agentic AI, and why is it generating so much buzz? It's a complex mix of AI techniques claiming to create agents that operate with near-human cognition. Incredible, right? Let's dig deeper.\"},{\"type\":\"text\",\"pitch\":-10,\"speed\":1,\"emotion\":\"Serious\",\"input_text\":\"However, there's a different narrative here. The truth is that these frameworks are unreliable abstractions of very complex systems. To achieve effective workflows, it takes expertise and hard work, not just high-level abstractions.\"},{\"type\":\"text\",\"pitch\":0,\"speed\":1.2,\"emotion\":\"Broadcaster\",\"input_text\":\"Let's talk about the overreach of Agentic AI. Developers are being promised tools that seem as user-friendly and magical as a children's storybook. But the reality is far more complex.\"},{\"type\":\"text\",\"pitch\":0,\"speed\":0.9,\"emotion\":\"Serious\",\"input_text\":\"Many frameworks offer an illusion of simplicity, masking the real need for deep technical understanding. These claims often serve more as impressive demos rather than practical solutions for important tasks.\"},{\"type\":\"text\",\"pitch\":5,\"speed\":1,\"emotion\":\"Friendly\",\"input_text\":\"Even so, this doesn't mean AI advancements aren't powerful tools—they are! But they require us to roll up our sleeves and understand them deeply, rather than relying on tools that promise to do it all for us.\"},{\"type\":\"text\",\"pitch\":-5,\"speed\":1,\"emotion\":\"Serious\",\"input_text\":\"These frameworks assume a level of technological maturity we've yet to achieve. Until the underlying infrastructure becomes reliable, these frameworks will remain just that—fluffy, hypothetical concepts rather than solid tools.\"},{\"type\":\"text\",\"pitch\":10,\"speed\":1.1,\"emotion\":\"Excited\",\"input_text\":\"This is only the beginning of leveraging AI's full potential, and as the technology matures, so too will the applications. We have to remain patient and critical of 'breakthroughs' that sound too good to be true.\"},{\"type\":\"text\",\"pitch\":0,\"speed\":0.9,\"emotion\":\"Serious\",\"input_text\":\"Agentic AI frameworks are still in their infancy, attempting to build solutions before establishing the necessary foundation. They promise a revolution that remains beyond our current technological reach.\"},{\"type\":\"text\",\"pitch\":0,\"speed\":1,\"emotion\":\"Soothing\",\"input_text\":\"Remember, understanding and mastering these AI systems will require time and effort—it’s a process that involves delving into the technicalities and capabilities of the tools at our disposal.\"},{\"type\":\"text\",\"pitch\":0,\"speed\":1.1,\"emotion\":\"Serious\",\"input_text\":\"Ultimately, the greatest lesson any technologist can take from this is to value expertise over convenience. Expertise will pave the way to truly innovative solutions as the technology stabilizes.\"},{\"type\":\"text\",\"pitch\":10,\"speed\":1.2,\"emotion\":\"Friendly\",\"input_text\":\"Thanks for joining me on this exploration of Agentic AI. Stay curious and critical, and in time, we can look forward to genuine advancements that live up to their promises. Until next time!\"}]},\"conversation\":{\"title\":\"The Lie of Agentic Frameworks for LLMs\",\"description\":\"An in-depth look into the misconceptions and realities of Agentic AI, emphasizing the importance of expertise and realistic expectations.\"}}",
      "metadata": {
        "tag": [],
        "title": "The Lie of Agentic Frameworks for LLMs",
        "author": "gpt-4o-2024-08-06",
        "status": "published",
        "created": "2024-12-02T10:09:42+01:00",
        "updated": "2024-12-02T10:09:26+01:00",
        "version": 29114,
        "description": "An in-depth look into the misconceptions and realities of Agentic AI, emphasizing the importance of expertise and realistic expectations."
      },
      "reference": []
    },
    "conversation": {
      "id": "c0daf530-df28-4e2c-ba96-a47614492589",
      "type": "conversation",
      "_link": "/v1/conversation/c0daf530-df28-4e2c-ba96-a47614492589",
      "message": [
        {
          "id": "57f6f093-3814-4f0f-8cb9-662bc715c8ac",
          "role": "user",
          "model": "gpt-4o-2024-08-06",
          "content": "The lie of Agentic Frameworks for LLMs.\nThe greatest trick the Devil ever pulled was convincing the world he didn't exist.\nTable of contents\nWhat is Agentic AI?\nDifferentiating between wheat and chaff\n\nThe lie of Agentic AI\nA simple abstraction of a complex unreliable API.\n\nLayers of abstractions\nA broken tool.\n\nThe Greatest Trick\nInventing the wheel in the jurassic era.\n\nEnjoying your read?\nIf you do like it, I kindly ask that you consider sharing it.\n\nShare me\nFollow me for updates\nPlease do get in touch, connect or follow. Carry on the conversation!\n\nLinkedIn icon\ngithub icon\ntwitter icon\nrss icon\nNovember 15, 2024\nBy Edward.\nSnake oil salesmen for tyingshoelaces\nYou've heard the famous quote, \"The greatest trick the Devil ever pulled was convincing the world he didn't exist.\"\n\nTheological references are relevant here, because more than any technological advancement in recent times, the mania surrounding Generative AI is more religion than science or engineering.\n\nThe purpose of this blog is not to fight the hype train (I suggest Gary Marcus) for a counterweight. This is a critical analysis of the snake oil being sold to the unsuspecting consumer. I have a simple thesis. Agentic AI frameworks are unreliable and fundamentally unrealistic abstractions. To build complex, chained and repeatable flows using LLMs requires knowledge, skill and hard work.\n\nIt's tempting to look for a shortcut, a silver bullet; but systems are binary communicators of bits and bytes, not vampires to be slain.\n\nWhat is Agentic AI?\nDifferentiating between wheat and chaff\n\n“\nAgentic AI is a groundbreaking advancement in artificial intelligence. A combination of different AI techniques, models, and approaches, it empowers a new breed of autonomous agents that can analyze data, set goals, and take action to achieve them—all with minimal human supervision. Agentic AI allows these autonomous agents to achieve near-human cognition in many areas, turning them into problem-solving machines that thrive in dynamic environments and constantly learn and improve with every interaction.\n\nWow. Amazing. Mindblowing.\n\nAbsolute nonsense as well... Let me try,\n\n“\nAgentic AI is the use of Generative AI in complex workflows. Outputs that are produced by LLMs may or may not be relevant to the task assigned. Agentic AI inherits the sophistication, intelligence and reliability of its output from system instructions and programming, most of which is highly abstracted, written by novices and will result in a barely functioning system. Agentic AI is great for making demos for unsophisticated consumers, yet generally bad for implementing in any use case where the outcome is actually important. It won't work at all unless you are an expert, but you almost certainly aren't.\n\nLet me be clear, I'm not sceptical about Gen AI in the slightest, I think used expertly and with tooling that doesn't yet exist, we can implement workflows that will change the way we perform tasks. Perhaps not to the extent the AI doomsayers talk about, but definitely to the extent of a paradigm shift for many ways of working.\n\nIt is no coincidence that the first killer use case for Gen AI is programming, because programming will always be the first domain to get proper tooling and infrastructure. The reason for this is simple. We build the tooling and infrastructure for programming tools with, errrm, programming. Programming is a low-level concern, a perfect first use case. It is also a skill which will by the way, still be a (if not 'the') relevant and most coveted ability for days, weeks, months, years and probably generations to come.\n\nCoding tools have proven the potential of Gen AI. I've got two separate LLM tools that I (pay for and) use daily for coding because they often get blocked up by overwhelming demand and green infrastructure. I use them concurrently in different IDE windows. This is because they are new, thus slow and sometimes flaky. Even like this, they have revolutionized my workflow. I'm having a bad day when I boot up my computer and Claude is busy so I'm talking to o1-mini instead. And this is the beginning of the shifting ways Gen AI will change the way you work…\n\nThese tools are being built from the ground up, they are brand new and not reliable yet. This is exciting, because it is the beginning. We are in the middle of a new wave. Which brings me to the subject of this blog. The lie and the overreach that is Agentic AI.\n\nThe lie of Agentic AI\nA simple abstraction of a complex unreliable API.\n\nMany years ago I was running a development agency and we had a freshly graduated developer join the team. I set them a reasonably simple task (to build a webform in PHP). I'd check in regularly\n\n“\nHow is it going?\n\n“\nReally well!\n\nDays and weeks passed, and tap tap tap went the keyboard. It was taking too long, so I went over to the programmer and asked them to show me the code they'd been working on. It was a nested layer of forEach loops, reaching into thousands of lines of code.\n\nMy stomach sank (the client was waiting for this…), I was lost for words. It was a physical blow, everything clicked. Was it my fault for not providing guidance? Was the programmer a lost cause? What the blazing hell am I going to tell the client who is supposed to pay for this? (We billed time, not results…)\n\nThe reason I digress is because this is exactly the same feeling I get when I read the docs for any Agentic Framework I have come across. I'm not alone.\n\nReddit Langchain quote\nI don't want to pick on any specific framework (I'm yet to find any Python Agentic Framework I can't apply this too), but when your docs start talking about roles and workers and bosses and backstories the whole concept begins to read like a fantasy novel.\n\nThe reason is, the whole thing IS a fantasy novel. It's a revolutionary and sophisticated tool built on pioneering technology, that works out of the box, for beginners. And every bit as convoluted and contradicted as that sounds.\n\nI'm not saying that complex LLM workflows cannot work, I'm saying that the expertise required and level of control of nuance is extremely high, and the creation of 'simple and accessible' frameworks is simply impossible until tooling is more mature and foundations are more solid. There are very few people in the world right now (although they do walk among us!) with the expertise to create and maintain reliable, complex iterative workflows (based on Gen AI or not!) that work and are true to the definition of Agentic workflows that are being sold the majority.\n\nBut here's the thing that bugs me; these few pioneers are working on cutting edge use cases that we probably aren't aware of. They aren't building in public on twitter, nor writing blogs and how-to-guides. They are not focused on creating a framework so you can easily automate your LinkedIn posts. They are getting on with things quietly and behind closed doors.\n\nI get it, I truly do.\n\nPeople want an accessible, easy to use framework that will allow them to build reliable and complex workflows with real business results. I want many things. Elixir of life. Grow a few centimetres taller. Become the tennis champion in my sports club. It is more likely that I will become the tall local tennis champion with eternal youth than for any of these agentic frameworks to work for the average user, and believe me and anyone who has seen me play tennis when I say that this is extremely unlikely.\n\nAgentic frameworks are the attempt to build a WYSIWYG website builder before we've even agreed a single unified standard for HTML. They are premature.\n\nThis isn't a case of cart before horse, this is trying to build a rocketship before we've discovered electricity. An opportunistic land grab that has already succeeded.\n\nLayers of abstractions\nA broken tool.\n\nThe problem with providing a tool or framework that allows you to abstract functionality is that it comes with a set of assumptions. When I buy a hammer, I assume it will work. When I buy a pressure cleaner, I assume it will work.\n\nThe problem is that when I use a framework, I assume it will work. But this is quite literally impossible given the maturity of the underlying technology. Far from increasing adoption, Agentic Frameworks are selling an illusion on top of highly controlled demos and finite use cases that will never actually work in the hands of the typical user (and there are millions…).\n\nThe damage they are causing is significant. Much as developers these days learn React before they learn javascript, or adopt frameworks before learning about programming principles. A generation of enthusiastic technologists are jumping on the hype train to learn 'x' framework. Here's the thing, it won't work (to any reasonable standard), in fact, building a demo or prototype is quite the achievement in itself.\n\nBut that doesn't mean that Gen AI or LLMs \"don't\" work, even though that will become the assumption in the short term. It simply means that we haven't reached a maturity level of the underlying technology to be able to safely build solid abstractions on top of it. There are too many problems to solve first. A few fundamental issues that need to be solved at a deep level by LLM providers includes but is not limited to:\n\nLatency\nAvailability\nSafety\nReliable output\nScale\n\nSo yes, what I'm saying is that the underlying APIs are often overwhelmed by usage; they are slow; vulnerable to hacking (especially prompt hacking), and introduce breaking changes on a weekly basis. The true innovators are working hard behind closed doors on these problems. But until we have more progress, good luck building a reliable production system, much less a framework.\n\n\nIf you don't have an understanding of the underlying APIs (which by the way, is the actual valuable breakthrough we are all working with), how can you possibly build a system on top of it. Some APIs don't require us to understand their limitations, mainly because they are a black box that just works. It will take years if not a generation for LLMs to reach this level of maturity, until then it should come with a warning sign \"Production use for Experts only\".\n\nIt's impossible to build a reliable abstraction on top of shifting sands. In fact, this breaks the basic tenet of computer programming. Abstraction in computer science is the process of removing elements that distract from more important elements. We can't focus on high level details when the low level system is not reliable, we need to understand nuance and use within a sophisticated paradigm. Not a low/no-code magic framework (with fantastical abstractions).\n\nThis is driving us into the trough of disillusionment, which while not in itself a problem as it provides space to build things of true value, it is still disappointing to observe.\n\n\nThe Greatest Trick\nInventing the wheel in the jurassic era.\n\nThe greatest trick the Agentic AI frameworks have ever pulled is to convince you that they are necessary. They are the devil of the LLM landscape.\n\nYou don't need crews, agents, backstories and pretty graphical interfaces showing you things that will never work.\n\n\nYou need to understand that LLMs are powerful tools that you should understand without stories from a fantasy novel. Your hello world and demo apps should be built by hand, in code. When you understand the how, the why and the what and can control the inputs and outputs to Gen AI models, then you will have a foundation to go out into the world and build something useful. And it will probably take a while.\n\nNot with wizards and managers and agents and dragons, but with the precise deterministic uses of a breakthrough in AI.\n\nFollow me:\n",
          "metadata": {
            "tag": [],
            "title": "The Illusion of Agentic AI",
            "author": "9NbDgfbyE1CICkmq3yjU",
            "status": "published",
            "created": "2024-12-02T10:09:30+01:00",
            "updated": "2024-12-02T10:09:26+01:00",
            "version": 29113,
            "description": "The user is sharing a critical analysis of the limitations and misconceptions surrounding Agentic AI frameworks for LLMs, arguing that they are fundamentally unreliable and oversold as easy solutions."
          },
          "reference": [
            {
              "id": "0d346b51-6593-4db7-884c-bf6599accfaa",
              "type": "prompt",
              "_link": "/v1/prompt/0d346b51-6593-4db7-884c-bf6599accfaa"
            }
          ]
        },
        {
          "id": "0d5a3e20-66d3-465b-b068-48dfe9aee1d4",
          "role": "assistant",
          "model": "gpt-4o-2024-08-06",
          "content": "{\"structured_data\":{\"voiceSettings\":[{\"type\":\"text\",\"pitch\":10,\"speed\":1,\"emotion\":\"Serious\",\"input_text\":\"The lie of Agentic Frameworks for LLMs. The greatest trick the Devil ever pulled was convincing the world he didn't exist. A famous line, but surprisingly relevant when discussing technology today, specifically the hype around Agentic AI.\"},{\"type\":\"text\",\"pitch\":10,\"speed\":1,\"emotion\":\"Friendly\",\"input_text\":\"Hello, everyone! Today we're diving into the world of Agentic AI—an area often surrounded by misconceptions and overblown promises.\"},{\"type\":\"text\",\"pitch\":0,\"speed\":1,\"emotion\":\"Serious\",\"input_text\":\"Agentic AI is touted as a new frontier for artificial intelligence, offering the ability to create autonomous agents that can think and act with minimal human input. This sounds revolutionary, doesn't it?\"},{\"type\":\"text\",\"pitch\":0,\"speed\":1,\"emotion\":\"Soothing\",\"input_text\":\"But let's take a step back. Beyond the claims of mind-blowing advancements, there are significant challenges and limitations that need to be addressed. That's what we are uncovering today.\"},{\"type\":\"text\",\"pitch\":0,\"speed\":1,\"emotion\":\"Excited\",\"input_text\":\"So, what is Agentic AI, and why is it generating so much buzz? It's a complex mix of AI techniques claiming to create agents that operate with near-human cognition. Incredible, right? Let's dig deeper.\"},{\"type\":\"text\",\"pitch\":-10,\"speed\":1,\"emotion\":\"Serious\",\"input_text\":\"However, there's a different narrative here. The truth is that these frameworks are unreliable abstractions of very complex systems. To achieve effective workflows, it takes expertise and hard work, not just high-level abstractions.\"},{\"type\":\"text\",\"pitch\":0,\"speed\":1.2,\"emotion\":\"Broadcaster\",\"input_text\":\"Let's talk about the overreach of Agentic AI. Developers are being promised tools that seem as user-friendly and magical as a children's storybook. But the reality is far more complex.\"},{\"type\":\"text\",\"pitch\":0,\"speed\":0.9,\"emotion\":\"Serious\",\"input_text\":\"Many frameworks offer an illusion of simplicity, masking the real need for deep technical understanding. These claims often serve more as impressive demos rather than practical solutions for important tasks.\"},{\"type\":\"text\",\"pitch\":5,\"speed\":1,\"emotion\":\"Friendly\",\"input_text\":\"Even so, this doesn't mean AI advancements aren't powerful tools—they are! But they require us to roll up our sleeves and understand them deeply, rather than relying on tools that promise to do it all for us.\"},{\"type\":\"text\",\"pitch\":-5,\"speed\":1,\"emotion\":\"Serious\",\"input_text\":\"These frameworks assume a level of technological maturity we've yet to achieve. Until the underlying infrastructure becomes reliable, these frameworks will remain just that—fluffy, hypothetical concepts rather than solid tools.\"},{\"type\":\"text\",\"pitch\":10,\"speed\":1.1,\"emotion\":\"Excited\",\"input_text\":\"This is only the beginning of leveraging AI's full potential, and as the technology matures, so too will the applications. We have to remain patient and critical of 'breakthroughs' that sound too good to be true.\"},{\"type\":\"text\",\"pitch\":0,\"speed\":0.9,\"emotion\":\"Serious\",\"input_text\":\"Agentic AI frameworks are still in their infancy, attempting to build solutions before establishing the necessary foundation. They promise a revolution that remains beyond our current technological reach.\"},{\"type\":\"text\",\"pitch\":0,\"speed\":1,\"emotion\":\"Soothing\",\"input_text\":\"Remember, understanding and mastering these AI systems will require time and effort—it’s a process that involves delving into the technicalities and capabilities of the tools at our disposal.\"},{\"type\":\"text\",\"pitch\":0,\"speed\":1.1,\"emotion\":\"Serious\",\"input_text\":\"Ultimately, the greatest lesson any technologist can take from this is to value expertise over convenience. Expertise will pave the way to truly innovative solutions as the technology stabilizes.\"},{\"type\":\"text\",\"pitch\":10,\"speed\":1.2,\"emotion\":\"Friendly\",\"input_text\":\"Thanks for joining me on this exploration of Agentic AI. Stay curious and critical, and in time, we can look forward to genuine advancements that live up to their promises. Until next time!\"}]},\"conversation\":{\"title\":\"The Lie of Agentic Frameworks for LLMs\",\"description\":\"An in-depth look into the misconceptions and realities of Agentic AI, emphasizing the importance of expertise and realistic expectations.\"}}",
          "metadata": {
            "tag": [],
            "title": "The Lie of Agentic Frameworks for LLMs",
            "author": "9NbDgfbyE1CICkmq3yjU",
            "status": "published",
            "created": "2024-12-02T10:09:42+01:00",
            "updated": "2024-12-02T10:09:26+01:00",
            "version": 29114,
            "description": "An in-depth look into the misconceptions and realities of Agentic AI, emphasizing the importance of expertise and realistic expectations."
          },
          "reference": [
            {
              "id": "0d346b51-6593-4db7-884c-bf6599accfaa",
              "type": "prompt",
              "_link": "/v1/prompt/0d346b51-6593-4db7-884c-bf6599accfaa"
            }
          ]
        }
      ],
      "metadata": {
        "tag": [
          "ai_generated"
        ],
        "title": "The Illusion of Agentic AI",
        "status": "published",
        "created": "2024-12-02T10:09:26+01:00",
        "updated": "2024-12-02T10:09:26+01:00",
        "version": 29112,
        "description": "The user is sharing a critical analysis of the limitations and misconceptions surrounding Agentic AI frameworks for LLMs, arguing that they are fundamentally unreliable and oversold as easy solutions."
      }
    },
    "structured_data": {
      "voiceSettings": [
        {
          "type": "text",
          "pitch": 10,
          "speed": 1,
          "emotion": "Serious",
          "input_text": "The lie of Agentic Frameworks for LLMs. The greatest trick the Devil ever pulled was convincing the world he didn't exist. A famous line, but surprisingly relevant when discussing technology today, specifically the hype around Agentic AI."
        },
        {
          "type": "text",
          "pitch": 10,
          "speed": 1,
          "emotion": "Friendly",
          "input_text": "Hello, everyone! Today we're diving into the world of Agentic AI—an area often surrounded by misconceptions and overblown promises."
        },
        {
          "type": "text",
          "pitch": 0,
          "speed": 1,
          "emotion": "Serious",
          "input_text": "Agentic AI is touted as a new frontier for artificial intelligence, offering the ability to create autonomous agents that can think and act with minimal human input. This sounds revolutionary, doesn't it?"
        },
        {
          "type": "text",
          "pitch": 0,
          "speed": 1,
          "emotion": "Soothing",
          "input_text": "But let's take a step back. Beyond the claims of mind-blowing advancements, there are significant challenges and limitations that need to be addressed. That's what we are uncovering today."
        },
        {
          "type": "text",
          "pitch": 0,
          "speed": 1,
          "emotion": "Excited",
          "input_text": "So, what is Agentic AI, and why is it generating so much buzz? It's a complex mix of AI techniques claiming to create agents that operate with near-human cognition. Incredible, right? Let's dig deeper."
        },
        {
          "type": "text",
          "pitch": -10,
          "speed": 1,
          "emotion": "Serious",
          "input_text": "However, there's a different narrative here. The truth is that these frameworks are unreliable abstractions of very complex systems. To achieve effective workflows, it takes expertise and hard work, not just high-level abstractions."
        },
        {
          "type": "text",
          "pitch": 0,
          "speed": 1.2,
          "emotion": "Broadcaster",
          "input_text": "Let's talk about the overreach of Agentic AI. Developers are being promised tools that seem as user-friendly and magical as a children's storybook. But the reality is far more complex."
        },
        {
          "type": "text",
          "pitch": 0,
          "speed": 0.9,
          "emotion": "Serious",
          "input_text": "Many frameworks offer an illusion of simplicity, masking the real need for deep technical understanding. These claims often serve more as impressive demos rather than practical solutions for important tasks."
        },
        {
          "type": "text",
          "pitch": 5,
          "speed": 1,
          "emotion": "Friendly",
          "input_text": "Even so, this doesn't mean AI advancements aren't powerful tools—they are! But they require us to roll up our sleeves and understand them deeply, rather than relying on tools that promise to do it all for us."
        },
        {
          "type": "text",
          "pitch": -5,
          "speed": 1,
          "emotion": "Serious",
          "input_text": "These frameworks assume a level of technological maturity we've yet to achieve. Until the underlying infrastructure becomes reliable, these frameworks will remain just that—fluffy, hypothetical concepts rather than solid tools."
        },
        {
          "type": "text",
          "pitch": 10,
          "speed": 1.1,
          "emotion": "Excited",
          "input_text": "This is only the beginning of leveraging AI's full potential, and as the technology matures, so too will the applications. We have to remain patient and critical of 'breakthroughs' that sound too good to be true."
        },
        {
          "type": "text",
          "pitch": 0,
          "speed": 0.9,
          "emotion": "Serious",
          "input_text": "Agentic AI frameworks are still in their infancy, attempting to build solutions before establishing the necessary foundation. They promise a revolution that remains beyond our current technological reach."
        },
        {
          "type": "text",
          "pitch": 0,
          "speed": 1,
          "emotion": "Soothing",
          "input_text": "Remember, understanding and mastering these AI systems will require time and effort—it’s a process that involves delving into the technicalities and capabilities of the tools at our disposal."
        },
        {
          "type": "text",
          "pitch": 0,
          "speed": 1.1,
          "emotion": "Serious",
          "input_text": "Ultimately, the greatest lesson any technologist can take from this is to value expertise over convenience. Expertise will pave the way to truly innovative solutions as the technology stabilizes."
        },
        {
          "type": "text",
          "pitch": 10,
          "speed": 1.2,
          "emotion": "Friendly",
          "input_text": "Thanks for joining me on this exploration of Agentic AI. Stay curious and critical, and in time, we can look forward to genuine advancements that live up to their promises. Until next time!"
        }
      ]
    }
  },
  "success": true,
  "metadata": {
    "id": "chatcmpl-AZwiph6mYlR3kZMbVf9izLS8OKhJ8",
    "model": "gpt-4o-2024-08-06",
    "usage": {
      "total_tokens": 7157,
      "prompt_tokens": 6213,
      "completion_tokens": 944
    },
    "object": "chat.completion",
    "choices": [
      {
        "index": 0,
        "message": {
          "role": "assistant",
          "content": "{\"structured_data\":{\"voiceSettings\":[{\"type\":\"text\",\"pitch\":10,\"speed\":1,\"emotion\":\"Serious\",\"input_text\":\"The lie of Agentic Frameworks for LLMs. The greatest trick the Devil ever pulled was convincing the world he didn't exist. A famous line, but surprisingly relevant when discussing technology today, specifically the hype around Agentic AI.\"},{\"type\":\"text\",\"pitch\":10,\"speed\":1,\"emotion\":\"Friendly\",\"input_text\":\"Hello, everyone! Today we're diving into the world of Agentic AI—an area often surrounded by misconceptions and overblown promises.\"},{\"type\":\"text\",\"pitch\":0,\"speed\":1,\"emotion\":\"Serious\",\"input_text\":\"Agentic AI is touted as a new frontier for artificial intelligence, offering the ability to create autonomous agents that can think and act with minimal human input. This sounds revolutionary, doesn't it?\"},{\"type\":\"text\",\"pitch\":0,\"speed\":1,\"emotion\":\"Soothing\",\"input_text\":\"But let's take a step back. Beyond the claims of mind-blowing advancements, there are significant challenges and limitations that need to be addressed. That's what we are uncovering today.\"},{\"type\":\"text\",\"pitch\":0,\"speed\":1,\"emotion\":\"Excited\",\"input_text\":\"So, what is Agentic AI, and why is it generating so much buzz? It's a complex mix of AI techniques claiming to create agents that operate with near-human cognition. Incredible, right? Let's dig deeper.\"},{\"type\":\"text\",\"pitch\":-10,\"speed\":1,\"emotion\":\"Serious\",\"input_text\":\"However, there's a different narrative here. The truth is that these frameworks are unreliable abstractions of very complex systems. To achieve effective workflows, it takes expertise and hard work, not just high-level abstractions.\"},{\"type\":\"text\",\"pitch\":0,\"speed\":1.2,\"emotion\":\"Broadcaster\",\"input_text\":\"Let's talk about the overreach of Agentic AI. Developers are being promised tools that seem as user-friendly and magical as a children's storybook. But the reality is far more complex.\"},{\"type\":\"text\",\"pitch\":0,\"speed\":0.9,\"emotion\":\"Serious\",\"input_text\":\"Many frameworks offer an illusion of simplicity, masking the real need for deep technical understanding. These claims often serve more as impressive demos rather than practical solutions for important tasks.\"},{\"type\":\"text\",\"pitch\":5,\"speed\":1,\"emotion\":\"Friendly\",\"input_text\":\"Even so, this doesn't mean AI advancements aren't powerful tools—they are! But they require us to roll up our sleeves and understand them deeply, rather than relying on tools that promise to do it all for us.\"},{\"type\":\"text\",\"pitch\":-5,\"speed\":1,\"emotion\":\"Serious\",\"input_text\":\"These frameworks assume a level of technological maturity we've yet to achieve. Until the underlying infrastructure becomes reliable, these frameworks will remain just that—fluffy, hypothetical concepts rather than solid tools.\"},{\"type\":\"text\",\"pitch\":10,\"speed\":1.1,\"emotion\":\"Excited\",\"input_text\":\"This is only the beginning of leveraging AI's full potential, and as the technology matures, so too will the applications. We have to remain patient and critical of 'breakthroughs' that sound too good to be true.\"},{\"type\":\"text\",\"pitch\":0,\"speed\":0.9,\"emotion\":\"Serious\",\"input_text\":\"Agentic AI frameworks are still in their infancy, attempting to build solutions before establishing the necessary foundation. They promise a revolution that remains beyond our current technological reach.\"},{\"type\":\"text\",\"pitch\":0,\"speed\":1,\"emotion\":\"Soothing\",\"input_text\":\"Remember, understanding and mastering these AI systems will require time and effort—it’s a process that involves delving into the technicalities and capabilities of the tools at our disposal.\"},{\"type\":\"text\",\"pitch\":0,\"speed\":1.1,\"emotion\":\"Serious\",\"input_text\":\"Ultimately, the greatest lesson any technologist can take from this is to value expertise over convenience. Expertise will pave the way to truly innovative solutions as the technology stabilizes.\"},{\"type\":\"text\",\"pitch\":10,\"speed\":1.2,\"emotion\":\"Friendly\",\"input_text\":\"Thanks for joining me on this exploration of Agentic AI. Stay curious and critical, and in time, we can look forward to genuine advancements that live up to their promises. Until next time!\"}]},\"conversation\":{\"title\":\"The Lie of Agentic Frameworks for LLMs\",\"description\":\"An in-depth look into the misconceptions and realities of Agentic AI, emphasizing the importance of expertise and realistic expectations.\"}}"
        },
        "finish_reason": "stop"
      }
    ],
    "created": 1733130571,
    "system_fingerprint": "fp_7f6be3efb0"
  }
}