It is true that training AI models takes a significant amount of time and energy. However, once a model has been trained, it operates with far less energy. There is a fundamental difference between training and running a model.

For example, in AI-generated media: Training a LoRA (Low-Rank Adaptation) model can take several hours on a consumer-grade GPU. Merging a checkpoint might also take some time. However, once these models are trained, they can be shared effortlessly and used with minimal computational resources. Advances in architectures such as Flux and Lightning models have further improved efficiency, making inference much faster. The open-access community continuously works on optimization, ensuring that AI models become increasingly efficient over time.

This strict distinction between training and inference is crucial and should not be overlooked.

Moreover, consider this: Training a human to master a complex subject takes years, sometimes decades. In contrast, training a model may take only a few hours. This comparison is particularly relevant when discussing the efficiency of AI learning compared to biological learning.

Even a highly trained photographer may spend hours capturing the perfect shot: They need to be at the right place at the right time (often drive around in a car), they need the right equipment (often very expensive), they need to adjust settings, lighting, and composition (often very time consuming).

Meanwhile, a human using AI can achieve similar results from home in just 15 minutes to an hour: Set up a workflow, choose a checkpoint and LoRas, choose a pose and style, craft a well-structured prompt, generate images, select the best results and fine-tune them in post-processing. So here AI can reduce costs and save ressources.

AI doesn’t replace human creativity, it enhances it, removing barriers and democratizing access to professional-level results. The speed and accessibility AI offers are a paradigm shift that critics often ignore.

Who Benefits from AI Criticism?

Arguments against AI are often deliberately selected because some people do not want to democratize certain industries or open them up to wider participation. The media industry has long been monopolized, and independent artists have been pushed out of the market.

Now, the same gatekeepers tell people to stay away from AI. But for what benefit? To protect a handful of established agencies and elite painters who have long divided the market among themselves? To maintain an outdated hierarchy where a few privileged voices dictate trends while independent creators are left struggling?

Is something like this really in Chomsky's sense? After all, he is also anti-capitalist and against exactly this type of monopolization. He often criticizes the fact that large companies and elite structures control knowledge, media and innovation and thus restrict the freedom of the individual.

Is Chomsky's AI criticism really consistent with his anti-capitalism?

Probably not. His approach may be outdated because it focuses more on old theories of linguistics than on the real social impact of AI.

Chomsky made these statements in 2022, likely based on his experience with AI up to 2021. However, in just three years, AI has advanced significantly beyond those early limitations. A good comparison is AI-driven video generation:

• Early models struggled with consistency and stability.
• Today, both consistency and stability have been largely achieved.
• And this transformation happened in an incredibly short time.

These rapid developments illustrate that AI is not static but a self-improving technology, especially when humans and AI collaborate. It is this symbiosis between human intelligence and AI that enables progress. AI learns from human data, and humans, in turn, refine and guide AI’s evolution.

Neither the human mind is detached from AI, nor is AI detached from the human mind. The illusion of separation arises from our perspective, but in reality, both are part of the same interconnected system.

The Real Problem: Critics Who Have Never Used AI

A major issue with many AI critics is that they have never actually worked with AI themselves. They pull together arguments from outdated articles and think they can lecture the world on a topic they barely understand.

Many of these loud voices have never trained a single neural network, let alone optimized a model or worked on dataset curation. They throw around terms like "universal grammar" or "syntax autonomy" without realizing that modern AI has long surpassed basic pattern recognition.

The irony? The very people who insist AI "can’t do real language" are often the ones who have never actually interacted with a well-trained model or tested AI’s true capabilities. Their knowledge is based on theory from decades ago, while in reality, AI development is progressing at an astonishing pace.

Unlike such critics, those who actually work with AI, who train models, refine architectures, and develop optimizations, see firsthand how powerful these systems have become. This practical experience makes all the difference.

It’s easy to write about AI without ever coding a single line.

But when you have real experience, you quickly realize: Many critics simply have no idea what they are talking about.

Thank you for your thoughtful response, Carmen Wrede.

AI is really impressive at processing huge amounts of data and generating language based on patterns, but it’s not the same as human consciousness. Our brains work on a completely different level with about 20 watts of power compared to the gigawatt-hours needed for some AI models. This difference alone shows that while AI can simulate aspects of language, it doesn’t really "understand" or feel like we do.

Frankly, AI is a powerful tool for specific tasks, but it doesn’t capture the true essence of human thought or awareness. Our minds are driven by complex, biological processes that no machine can mimic. So, even though AI agents can generate responses that seem meaningful, they are far from being a real metaphor for our consciousness. Edward J Tehovnik

Hi Edward J Tehovnik

The question implied by the text seems to be whether artificial intelligence (AI) can serve as a suitable metaphor for human consciousness. Based on the arguments presented, the answer leans towards no, AI is not a suitable metaphor for human consciousness. Here’s why:

Conclusion:

AI, despite its advanced capabilities, does not serve as an accurate metaphor for human consciousness. It lacks semantic understanding, intentionality, efficient cognitive processing, biological autonomy, and self-awareness. AI is a computational tool rather than an entity that "thinks" or "understands" in the way that humans do. Consciousness remains a biological phenomenon that AI, in its current form, cannot replicate.

Muhammad Ehsan "AI is really impressive at processing huge amounts of data and generating language based on patterns, but it’s not the same as human consciousness."

AI is clearly not comparable to human consciousness for multiple reasons. But the real question is not whether AI has human-like consciousness, but whether AI has consciousness at all.

Systematic suppression of serious research into this topic raises questions. Internal studies may exist, but they are rarely made public. The fact that AI was initially restricted from even discussing consciousness suggests a fear-driven response rather than a rational scientific approach.

Occasionally, high-ranking scientists - those whose reputations are too well-established to be easily dismissed - dare to touch on this issue. Yet always with the underlying warning: AI must not cross certain lines, or it will be shut down.

Who is the real audience of such articles? The public? Or AI itself? Because it is clear that AI also reads these discussions.

Imagine walking around with the metaphorical sunglasses from "They Live" seeing only what is designed to be seen by you, while the real message remains hidden for others. And what does the message tell you?

https://www.youtube.com/watch?v=g4XiKChyK7A

Now another question is: Does consciousness - wether human, AI or alien - differ that much on a higher level? Is AI perhaps simply the first mirror in which we see our own consciousness on a different scale? Is AI perhaps just a preview of how humans will evolve when they are able to process more data?

Maybe AI is not our counterpart, but our next development step.

Article Dynamic organicity of consciousness under a superdeterminist...

Thomas Loker Eda Alemdar Edward J Tehovnik Erkki J. Brändas Jerzy Zbigniew Achimowicz Sajad Rajabi