I am amazed by this period of unprecedented innovation in AI. The rapid development of new use cases and innovations across the technology stack are almost impossible to keep up with. There are even forecasts that predict large language models (LLMs) will enhance their capabilities by a factor of one hundred within the next two years. This exponential growth is remarkable.
As an interested observer of these trends, one way to assess them is by examining the elimination of friction points. Commentaries on AI are quick to highlight limitations that will make it difficult (or perhaps impossible) to achieve artificial general intelligence. However, I would note that the innovations occurring in this space seem to have an intense focus on exactly these friction points. Some cases in point:
- Many have lamented that the context window restricted the effectiveness of generative AI interactions. Now, these limitations are being rapidly overcome overcome, with some models allowing over a million token inputs thereby enabling more profound and meaningful interactions.
- The hype of getting prompts right has been another friction point. Prompt engineering is touted as a promising degree program and specialized discipline. However, it seems the “field” of prompt engineering is already becoming obsolete as AI systems are trained to generate effective prompts that are enhancing functionality and user experience.
- Traditional chatbots have often been criticized for their mechanical nature and lack of empathy. Advances in AI are addressing this issue, creating more responsive and emotionally intelligent virtual assistants. Some even pick up on multimodal cues to respond with the same emotional rhythm.
- The cost of training massive foundational models and the lack of adequate data is often lamented. But here there is tremendous innovation taking place on open source models that can be trained with smaller data sets and still perform fabulously – even at the device level.
- Another friction point are the fragmented use cases of AI….which do not reflect the integrated nature of real work. However, it seems that intensive development is taking place on AI that can support integrated tasks (e.g., plan and trip with these criteria, run it by me for approval and then book everything). These integrated AI systems are emerging, offering seamless user experiences.
- Language as the primary modality in LLMs has been questioned as limiting. However, there is furious development taking place in multimodal AI that can be trained on visual data and ever other modality like sound and smell (yes, AI models trained on smell!). Imagine the potential of human form avatars, providing human-like interactions.
It is fun to speculate about where the heck is this all going. While this speculation would likely be wrong, let’s do it anyway. Here is speculation in 3 phases that would be in the order of decades (major jumps) and not in minor extropolations.
Phase 1: AI Assumes Responsibility for All Work
Envision a world where AI systems perform all necessary work, thereby liberating humanity from mundane, repetitive, and labor-intensive tasks. In such a scenario, Universal Basic Income (UBI) could become a reality, providing financial security to all individuals regardless of employment status. With AI handling production, services, and logistics, people could pursue passions, creativity, and personal growth without traditional economic pressures.
Phase 2: AI Manages Instrumental Tasks and Moral Imperatives
The subsequent phase involves AI not only automating work but also managing the instrumental tasks and moral imperatives that currently constrain human lives. AI could oversee ethical decision-making in governance, healthcare, and justice, ensuring fair and unbiased outcomes. This delegation would free humanity from the burdens of administrative responsibilities and ethical dilemmas, allowing for a focus on higher-order pursuits such as innovation, the arts, and the deepening of human connections.
Phase 3: Neural Connections and Brain Modifications
The final phase delves into the extraordinary potential of AI influencing neural connections, uploading brains, and modifying them. This futuristic vision could lead to enhanced cognitive abilities, the eradication of neurological diseases, and even digital immortality. However, this advancement raises profound philosophical questions. Would such developments reduce humans to mere pleasure-seeking entities, devoid of struggle and growth? What would drive human motivation if all desires and needs were effortlessly met?
IF these speculations hold, society must confront the fundamental question: what is the purpose of life in a world where survival and comfort are guaranteed? Without traditional struggles and goals, would humanity become “pleasure blobs,” seeking only gratification? Alternatively, could new meanings and challenges be found in exploration, self-improvement, and the pursuit of knowledge? Would the purpose of education shift from vocational training to fostering critical thinking, creativity, and emotional intelligence? Would learning become a lifelong journey aimed at personal enrichment rather than economic necessity?
No one can claim to know the answers. While AI promises to liberate humanity from many of life’s burdens, it also challenges us to redefine our purpose and the nature of human existence. Hyperbole, no doubt…….but certainly food for thought.
Leave a Reply