AGI (Artificial General Intelligence) is very near, domain-specific AGI is already here
|
Home - EndCapitalism.org By Mehmet Kurtkaya 11.6.2025 X /Twitter @mkurtkaya Mehmet Kurtkaya - Telegram Yesterday, while checking my notes from 15 years ago, which I had not looked at in many years, I noticed an old quote of mine that I had tagged as an aphorism: "We are the fruits of time, and time is our interpretation of life." In the past, I was taking notes on my philosophical thoughts, which I had not done in years. I had forgotten about this and other quotes of mine long ago. That long separation makes one indifferent, as if it were someone else's quote, and one begins thinking about it anew. I had the idea to ask DeepSeek AI for an evaluation. Specifically, I asked DeepSeek this: `I will give you a quote and you'll tell me if it is deep or play of words nonsense. We are the fruits of time and time is our interpretation of life` I chose the DeepThink version by pressing the button and saw its chain of thoughts but did not read them, except for a few lines from each. What followed is a conversation that proves beyond doubt that AGI is here, by the simple fact that few humans can hold such a conversation, and I invite anyone interested in philosophy or machine intelligence to review it. You don't need to read the entire text; you may skip the inner thoughts of the AI but observe the flow of dialogue. I noticed that DeepSeek was making flattering comments about my ideas, perhaps too much, but that's the only thing I can criticize. Look at the flow of ideas, the bringing up of concepts from history and the world of physics, and how it ties them together while still remembering my original quote and checking back against new information, both what it generated and my additional ideas, and testing them for consistency. It did a job that professors of physics or philosophy should review. They should examine, first and foremost, the consistency rather than the validity of the AI's work. Not only that, anyone working on AI should have a look. I am sure there are many exciting dialogues that people have with their AI models, but this goes directly to human existence and the universe around us. It may go deeper; people may explore other aspects of human existence, but it may be one of the shortest yet broadest coverages of the concept of time. I have long said that domain-specific AGI is already here, and AI agents will certainly unleash AGI in short order if we are not already there. Check my tweet: Yes. For months, I have been writing that domain-specific AGI is already here, and while seemingly contradictory, DeepSeek loved the term to describe the current state of affairs in AI. And AGI in the broadest sense is about to be unleashed via AI agents. Here is our short conversation with DeepSeek AI on TimeI also asked Grok about this short article of mine and my conversation with DeepSeek and Grok found it impressive yet didn't call it AGI. Here is Grok's closing remarks: "While the dialogue is impressive, it’s domain-specific (philosophy and science). Full AGI would require generalization across unrelated tasks (e.g., composing music or solving legal disputes). Your claim of “domain-specific AGI” is accurate, but “AGI in the broadest sense” (as per your tweet) may still be emerging, as you suggest with AI agents." I disagree with Grok's refutal of AGI as AI can compose music and suggest legal recourse.
Mehmet Kurtkaya
|