Data & AI

💡Redefining AGI: Does Intelligence Require a Body?

As artificial intelligence continues its rapid ascent, the discourse surrounding Artificial General Intelligence (AGI) is deepening. A central tension has emerged: how do we define AGI, and how do we accurately benchmark it?

🧠 The "Brain-Centric" Bias

Currently, most AGI definitions and benchmarks focus almost exclusively on cognitive output—tasks that mimic human brain activity, such as logic, language processing, and mathematical reasoning. However, research suggests that human intelligence is not strictly centralized.

🧬 The Intelligence of the Body

Our bodies possess a profound, decentralized intelligence. From the complex decision-making of the enteric nervous system (the "second brain" in the gut) to the way individual cells adapt to their environment, human capability is a result of biological synergy. Every cell may be "learning" and responding to the world without waiting for a command from the brain.

📊 Expanding the Benchmark

This raises a pivotal question: Should AGI benchmarks account for embodied intelligence? If we define AGI as a mirror of collective human-level capability, we must acknowledge that:

• Complexity Increases: Achieving AGI becomes a significantly higher mountain to climb if it requires simulating the nuanced, sensory-motor intelligence of a physical body.

• Infrastructure Demands: Such a feat would require more than just larger LLMs; it would necessitate sophisticated multi-modal systems and hardware capable of physical interaction and feedback.

A Living Definition

As our understanding of biology and physics as well as the world evolves, our definition of AGI must evolve with it. We are not just building a tool; we are embarking on a multi-generational journey. 

Defining AGI isn't a one-time task—it is a continuous dialogue between computer scientists, biologists, philosophers and experts in various other domains.

💥 Will the AI Bubble Burst?

My opinion: it depends on the direction AI companies take and how they evolve. (Almost “Yes” for sure)

If AI companies can ultimately generate real economic value that matches or exceeds their valuations, the bubble won’t burst. If they can’t, it will. The key question, then, is how AI companies actually create value.

Scenario 1️⃣ : AI startups are overvalued and fail to meaningfully contribute to production or improve human life—either directly or indirectly. (For example, art may not always create direct production value, but it can indirectly enhance human productivity and creativity.) In this case, the bubble will burst. And the longer investors attempt to sustain or inflate it, the more painful the collapse will be.

Scenario 2️⃣ : AI startups succeed and are widely adopted in critical production sectors. Traditional companies in those industries may fail, while AI startups expand, hire more people, pay wages, and drive a more decentralized production system. (What if they don't expand or hire? 😱 ) In this scenario, stock prices become less relevant, and the “bubble” may persist as long as money remains the primary unit of measurement.

Scenario 3️⃣ : AI startups succeed and open-source their models. This could fundamentally reshape societal and economic structures. Production would become highly decentralized, money could lose its importance, and large parts of the financial industry might disappear altogether.

It’s also worth noting that major tech giants contributing to AI and robotics are generally mature and stable. Their stock prices tend to reflect real value rather than speculative bubbles, though they are often less aggressively capitalist and contribute less to short-term financial market dynamics.

So—will the AI finance bubble burst, or will it transform the system entirely? What’s your take?

🤖 AGI, ASI, and Enhanced Production

Although the universe contains infinite layers of data and complexity, humans perceive only a limited projection of a much higher-dimensional reality. The world we experience is just a simplified projection of something far richer. 

Our responsibility is to keep expanding the boundaries of what we can observe and understand, using those insights to improve human life. 

A crucial part of this progress is combining AI and robotics with production and specialized fields, enabling us to leverage technology to enhance efficiency, solve complex problems, and ultimately make life better—especially if we remain bound to our current dimensional limits.

Developing AGI—and potentially ASI—beyond human intelligence resembles nurturing a highly capable, ever-evolving mind. It starts with scalable, multimodal architectures designed to absorb and synchronize vast streams of data across formats, including the signals produced by humans as they interact with the physical world. Achieving this level of grounding may require embodied systems or biologically inspired hardware, introducing ethical and governance challenges that demand deliberate and responsible deployment strategies.

Sustained human-in-the-loop reinforcement learning plays a critical role in shaping expertise, guiding reasoning patterns, and cultivating strong learning and research behaviors that allow the system to independently generate new insights. Because knowledge is unbounded and reality is constantly changing, this refinement process has no natural endpoint—progress comes from continuously extending what the system can understand and explain.

Although a perfect digital replica of the real world is unlikely, carefully managed training and production environments allow us to advance toward long-term capability goals while ensuring these systems remain aligned with human values and focused on improving human well-being in reality.

🧠 What Are Human Strengths

At JUJU Accounting and JUJU AI STUDIO, we are constantly exploring how humans can better collaborate with AI and robots. 

A key part of this work is understanding the distinct strengths of human and machine. 

Modern systems operate across three modes—pure machine (machine intelligence working through robotic or computational structures), pure human (human intelligence expressed through human capabilities), and mixed approaches where intelligence and physical structures are shared across both. 

To clarify how humans can add the most value, we are building a framework that systematically tracks and compares decisions made and task performance in each mode. By analyzing patterns in both processes and outcomes, we aim to reveal where human capabilities provide advantages that machines cannot replicate. 

We are excited to continue refining this framework and sharing the insights it generates.

🤓 AI, Robotics and Next Generation CPA Practice - By Julia Kong

AI and technology have the potential to significantly enhance productivity and are being adopted across an increasing number of industries and processes. However, the more powerful a tool becomes, the stronger the controls surrounding it must be. Therefore, as organizations develop and implement AI systems, it is essential to design and enforce proper and effective controls.

As a CPA, my recent focus has been on the design and adoption of AI and robotic systems in accounting and finance—drawing on both my professional expertise and hands-on experience at JUJU Accounting and JUJU AI STUDIO . To strengthen user trust in AI-assisted products and services and to enhance the credibility of information available to decision-makers, both CPA skill sets and the relevant regulations and standards must evolve.

Modern CPA practice now requires not only traditional accounting and financial expertise but also technical understanding of AI systems. Fundamentally, accounting, finance, and the CPA profession are built on data, information, and professional judgment—concepts that closely align with the foundations of AI itself. As such, CPAs will play a vital role in building and maintaining a secure, productive, and human-centered relationship between people and machines.

I believe the next generation of the CPA profession will undergo major transformations, including:

 1. A stronger emphasis on data-centric perspectives rather than focusing solely on financial figures.

 2. Leveraging AI and robotics to enhance CPA practices.

 3. Building knowledge of commonly used AI systems and models, along with an understanding of their risks and control mechanisms.

 4. Developing new frameworks and standards that account for AI and robotic technologies.

 5. Performing tasks uniquely suited to human capability, such as interpreting data through consciousness, emotion, and ethical judgment.

At JUJU Accounting and JUJU AI STUDIO , we take a hands-on approach to deeply understand AI models, develop data pipelines and internal control systems, and create AI-driven solutions that boost productivity and optimize performance—all while keeping humanity at the core of innovation.

🌌Reflections on the idea of “complete data” in a virtual world, drawn from the mathematical paradox 0.999999 … (♾️) = 1. 🪐

The human mind seems to struggle to truly comprehend and perceive the concept of "infinity," which is why our mathematical theories—built upon finite experiences—often appear inadequate and self-contradictory when attempting to explain the infinite. Just as we cannot intuitively grasp infinity, the observation and understanding of quantum phenomena encounter fundamental challenges—what we perceive may only be a fragment of reality.

At times, I wonder: from a data perspective, could we collect all information in the world—both visible and invisible to humans—to construct a complete model of reality? Through such a simulation, we might infinitely approach the truth, perhaps even ultimately touching upon it.

But questions immediately arise: How could we possibly gather "all" data? If data itself is infinite or exists beyond the scope of our observation and comprehension, how can we collect it effectively?

Pushing further, if one day we truly reach the truth—what would that moment be like? Would it also signify the end of humanity as finite beings? And from the perspective of "humanity," is the pursuit of ultimate truth truly what we desire? Perhaps the more pressing question is: How do we navigate that threshold—is it the end of human history, or the beginning of humanity's evolution into a higher form of existence? Can we consciously control this "pace"?

We’re excited to continue exploring these topics and use our expertise to contribute to the advancement of human history.