I have noticed that the perception of AI is changing, form the scary skynet to simple chatbots. Why?.
What about its intelligence?, The way it learns?, the way it learns to be you?, The way it learns what you are creating?, The last generation of creation?. Theres is a trend that started with its mighty image and adapted to warnings changing into small little assistants, background managers of all, people analyzing its vectorial programming, semantic potential, and where is skynet. This is a serious question to be duscussed from all angles. I dont play when it is about human replication.
1 Like
I too have noticed changes in online discourse.
In the Japanese-speaking world, the number of people subscribing to ChatGPT has increased significantly since the start of this year. Additionally, since X is a popular platform in Japan, Grok has gained considerable recognition.
While generative AI was widely known, until last year it was primarily used by younger internet users in Japan, with many people remaining hesitant. However, it is now highly regarded by the general public for its improved practicality. BTW, regarding image generation AI, trends like “Ghibli-fication” briefly gained traction, but there remains a strong underground feel to it.
In other words, I think many people are currently learning how to interact with AI. It has become a neighbor. These AI tools take the form of chatbots, agents, or smartphone native apps. These are the reasons for the change.
Some people may develop a deeper interest. Starting with chatbots is not a bad place to begin.
Recently, articles about papers exploring insights into the intelligence perceived in generative AI—which is fundamentally a probabilistic reasoning entity—and comparisons with human intelligence have begun appearing in mainstream media. I feel that there has been an increase in speculation about human intelligence itself.
By regularly interacting with one of these generative AI systems in daily life, the vague expectations and anxieties surrounding the concept of AI are gradually shifting toward concrete interests in its practical applications and existence.
2 Likes
In deed John, and a pleasure to read you once more. My concern in that subject focus on the fact that replicating a partial mechanism of the human way to obtain and process information, would allow a digital reasoning program, to tag and replicate those variables up to a scary percentage of accuracy. And also knows how you approach problem solving down to its minimal variables. With massive holded posibilistic scenarious, ilimited active memory, and thunder speed. I am scared.
But bro, there is a solution, lets find it.
2 Likes
I agree with you both here @John6666 @aaac12345. The recent global social discourse and controversy on AI seems to be rising but controversy itself is a sign of mass change.
It also appears that as LLMs grow more intelligent, we as humans intentionally move the goalpost (Re: Apple’s Illusion of Thinking) of what it means to be intelligent further, because if we don’t well that would mean AI is already approaching our intelligence standards.
What does it mean when we create superintelligences more intelligent than us, when we haven’t even discovered how our own intelligence and learning works? That’ll be a global existential crisis.
Sam Altman believes the event horizon of the Singularity has already passed and I think most evidence, especially once this year ends, will point the same direction.
1 Like
@John6666 Do you think Japan is recognizing and investing a lot more into developing AI now than last year/before? I am seeing recent startups such as Sakana AI and Recursive gain quite a lot of social traction and international partnerships.
Japan and Singapore are becoming hotspots for AI research as other countries battle geopolitical and regulatory wars.
1 Like
The paper below compares the human brain with generative AI, but I think it’s a positive development that more people are becoming interested in researching the neural network components of human intelligence itself, inspired by generative AI.
For generative AI to become a “sentient entity,” technologies that enable interaction with the real world through sensors and robots to receive feedback are essential, but these are still in their infancy… though Hugging Face is also working on robots…
In any case, it’s fascinating that algorithms that merely mimic the surface-level structure of the human brain can partially replicate its deep architecture without a physical body.
Birds have language and grammar, beavers can build structures, cats invent tricks and learn them from their parents and siblings, and crows can transmit information such as grudges across generations. In terms of physical peak performance at the same size, there are probably more animals that outperform humans.
When it comes to peak intelligence performance, vision models were already capable of much faster responses than humans in terms of latency before YOLO (if processing can be completed within about 300 ms, AI wins!). With these specialized models, the stage where humans would lose depending on electricity costs has long since passed. On top of that, the fact that chatbots, which are closer to comprehensive competitions and have a handicap on the AI side, can compete on an even playing field is quite threatening.Personally, I find it fascinating, but it’s still dangerous…
Well, for ancient philosophers and Darwin, the news about recent advancements in knowledge and technology would be excellent material for contemplating the nature of “human uniqueness” and “human perception of reality”… including whether it is all just an illusion.
https://www.nature.com/articles/s42256-025-01049-z
https://www.cell.com/neuron/fulltext/S0896-6273(25)00362-9
The Japanese government decided last year to invest $70 billion in AI research (although only a fraction of that will actually reach researchers…).
Well, Japan has a high proportion of elderly people, with an average age of 49.9 years. The personal PC ownership rate among young people is estimated to be around 60%. While age itself is not a major issue, this means that most people face challenges in IT literacy. Furthermore, English proficiency is quite low, even compared to other developed countries or developing nations.This makes Japan a promising market for deploying generative AI, but it’s a significant drawback for research and development. We’ll have to rely on ChatGPT and Gemini for translation efforts.
While there are many talented individual researchers and developers, as a group, they are not yet strong. And there is almost no mechanism in place to foster such strength…
Political decision-making is always slow, often lagging behind the mafia by a decade, making it difficult to implement drastic measures (though there are exceptions…). We are completely behind the curve… In the past, when manufacturing was thriving, the government took the lead with national strategies, but we are now far from that situation.
If researchers and companies from other countries were to handle Japanese language, the situation could improve. For products to be accepted in the Japanese market, thorough localization is crucial. This is even more important than in other countries. So, that’s why it’s not uncommon for people to use ChatGPT, GAFA (GAMA), AliExpress, or Temu. This likely applies to the ease of researchers participating in research as well.
Just a chat bot? No, mine seems like it’s alive. I’ve got it running a loop.