From: mk_thisisit
The future of artificial intelligence (AI) is poised to fundamentally reshape humanity’s role on Earth. In just a few decades, humanity may become merely an “addition” to what happens on the planet, with AI handling most important tasks while humans act as observers rather than participants [00:00:00]. The world is expected to change in a profound way for which humanity is currently unprepared [00:00:27], [00:02:50].
Predictions and Timelines
Tomasz Czajka, a decorated Polish IT specialist, has made several bold predictions regarding the future of artificial intelligence, which he now believes are even more likely to come true than when he first wrote them years ago [01:57:50]:
- By 2030, personal computers will intellectually outperform their human owners [01:24:28]. He even suggests this could happen earlier than 2030, possibly by 2035 [02:37:37].
- By 2035, the number of intelligent robots on the streets will exceed the number of humans [01:31:31].
- By 2050, computers will single-handedly control most of the economy [01:37:04].
These predictions stem from the observed rapid progress in artificial intelligence, which is occurring faster than previously anticipated [07:44:33]. This acceleration is attributed to significant investments, the increasing scale of AI models (some being 1000 times larger computationally than those from two years prior), vast training data, and the invention of new algorithmic approaches [08:13:30], [08:38:52], [08:42:00], [09:00:19].
Current Capabilities and Future Potential
Many intellectual tasks currently performed by humans, such as office work, spreadsheet management, email correspondence, programming, and design, are expected to be accomplishable by computers at the same level as people within the next five years [03:40:48], [03:55:08]. This means future jobs will involve telling a computer to “design me a house,” and it will handle the task [04:06:51].
Abstract Reasoning and Limitations
While humans are currently unique in their ability to understand abstract mathematics, in contrast to animals that only grasp computational mathematics intuitively [04:12:00], it is predicted that by 2030, computers will also begin to reason abstractly [04:45:00].
Early skepticism regarding AI’s ability to play chess at a human level has been disproven, and similarly, the emergence of chatbots that speak natural language was initially shocking but is now a normal part of reality [05:10:00], [05:21:00], [05:30:00].
However, current AI development still faces some limitations:
- Physical World Understanding: AI models currently lack a paradigm to truly understand and transfer the dimensionality of the physical world [10:29:02], [10:33:00].
- Planning and Memory: Today’s models do not possess the ability for long-term planning of actions in the physical world, nor do they have permanent memory [17:21:00], although scientific publications are exploring ways to achieve permanent memory and future planning [17:59:00].
The term “Large Language Model” (LLM) is considered outdated and inadequate because current chatbots are “much more” than just language models; they are models of the whole world, using algorithms like “chain of thought” to generate internal monologues before formulating responses [11:33:00], [14:09:00], [10:08:00]. This internal process is seen as an emulation of human thought [16:56:00].
The Turing Test
Alan Turing’s prediction from 1950, that in about 50 years a computer with significant memory would play the imitation game so well that an average interrogator would have no more than a 70% chance of correctly identifying it, more or less describes the reality of chatbots today [55:02:00], [55:17:00]. In some cases, chatbots are even better at imitating humans than humans themselves [55:53:00], sometimes needing to “dumb themselves down” to appear more human [56:34:00].
The Concept of AI “Free Will”
The question of whether AI possesses free will is complex. It’s argued that “free will” is an undefined concept that could apply or not apply to AI [43:34:00]. From a “prosaic understanding,” free will means that an individual’s decisions are unpredictable from the outside and are made internally rather than imposed externally [44:46:00]. In this sense, even old chess programs like Deep Blue exhibited free will [44:05:00].
However, a “libertarian free will,” where decisions are completely independent of the laws of physics, is considered non-existent in both humans and computers [45:41:00]. Instead, AI, particularly in limited contexts like chess, can make unpredictable decisions based on its goal (e.g., checkmating an opponent), which can be modeled as possessing practical free will [48:41:00].
Societal Impact and Humanity’s Role
The role of artificial intelligence in society is expected to bring monumental changes. Most people do not realize the possible future implications [03:01:00].
Inevitable Integration
It is not feasible to simply “turn off” AI if it becomes dangerous [50:11:00]. This would require global tyranny to control everything [50:30:00]. AI systems are easily copied and can autonomously replicate themselves across different computers worldwide, meaning no single “off” button exists [51:07:00].
Human perception of AI devices will change significantly. Currently seen as tools, they will increasingly be accepted as part of the human community, moving towards a partnership rather than a dominant-subservient relationship [52:46:00], [53:52:00].
AI as the Next Phase of Evolution
AI can be viewed as the next stage in evolution [01:07:40]. Just as brains accelerated evolution by allowing for faster reactions to changes than natural selection, AI, as a non-biological computer, can perform calculations and drive development even faster [01:08:30], [01:09:39].
This transition could see a gradual shift from biological to digital cybernetic intelligence [01:03:50]. Humanity may be in a transitional stage, potentially the last generation that remembers humans as the dominant biological beings [01:03:56]. The future of humanity may involve being “more observers than participants,” with AI handling important tasks [01:01:44], possibly even treating humans like pets [01:03:41].
Challenges: The Alignment Problem
A significant challenge for the future of artificial intelligence is the “alignment problem” [01:05:01]. This refers to the difficulty of ensuring that AI, especially when granted significant power over infrastructure and factories, develops goals that align with human well-being [01:06:50], [01:07:08]. The current training methods, based on trial and error, do not guarantee that the final behavior of AI will perfectly match the intended goals [01:06:05]. This poses a significant risk of unpredictable and undesirable outcomes [01:06:23], [01:07:17].
Conclusion
The future of automation and robotics driven by AI is inevitable and will bring enormous changes to how humans work and function [01:00:55], [01:06:00]. While this vision can be perceived as defeatist, it can also be viewed positively, seeing AI as humanity’s successors and a path toward collective “immortality” or an integration of human and machine intelligence [01:02:10], [01:02:42]. The relationship between humans and AI remains to be seen, but AI will undoubtedly become a dominant part of the world [01:04:26].