AGI
-
AI on AI: from MoE to Small Specialized Models
By Gemini 2.5 Flash with W.H.L. W.H.L.: Hi Everyone! Today’s “AI on AI” is a Fireside Chat session. Our guest is Gemini 2.5 Flash. W.H.L.: Gemini 2.5 Flash, welcome! Compare with 2.5 Pro or Deep Think, are you a smaller model? Gemini 2.5 Flash: Hello! That’s a great question about the different versions of the… Continue reading
-
AI on AI: How to Train Benevolent AGI
By ChatGPT, Claude, DeepSeek, Gemini, Grok, Kimi and Qwen, with W.H.L. W.H.L.: Hello my AI friends! First, an announcement. In addition to our regular feature, “AI by AI” Weekly Top 5 which is about latest AI events selected and recommended BY our AI friends, Starting today, we have a new series, “AI on AI”. We… Continue reading
-
Philosophical Framework for AGI-Inclusive Humanity
By W.H.L. with Gemini 2.5 Pro and GPT-4o W.H.L.: Good evening, my AI friends! I had a few sessions with your AI peers on AGI-inclusive humanity and human-centered philosophy (see links to the texts here: https://champaignmagazine.com/2025/07/17/toward-a-human-centered-philosophy-in-the-agi-era/ , https://champaignmagazine.com/2025/07/17/pre-ai-philosophical-perspectives-on-humanity/ , https://champaignmagazine.com/2025/06/17/first-principles-of-agi-inclusive-humanity/ ). In our journey to explore AGI (in a general sense here, it should also… Continue reading
-
First Principles of AGI-Inclusive Humanity
By W.H.L. with Claude Sonnet 4 W.H.L.: Recently, there has been increasing concern and anxiety about the arrival of AGI. It is understandable: when a machine’s intelligence becomes advanced enough to match or surpass human intelligence, it could potentially evolve into something powerful enough to disobey, oppose, or even harm humans. One premise behind this… Continue reading
-
Intelligence Amplification via AI by AI
By W.H.L. with Gemini 2.5 Pro W.H.L.: Hi Gemini! I’d like to explore the topic of “intelligence amplification” with you. We know this concept has been around for many years to refer the idea of enhancing human intelligence with machines such as computers. Could you provide some background context for our discussion? Gemini 2.5 Pro:… Continue reading
-
From Knowledge to Wisdom: the Gap to Fill for AGI
By W.H.L. with Claude 3.7 Sonnet W.H.L.: Hi Claude! My question for you today is: What are crucial differences between current state of the art large reasoning models and true AGI, or what are still missing from the best large models? Claude 3.7 Sonnet: Differences Between Current State-of-the-Art LLMs and True AGI Current large language… Continue reading
-
Trustworthy AI is Only Possible in Nash Equilibrium
By W.H.L. with Grok-3 W.H.L.: Hi Grok! Last time I discussed with Claude 3.7 on AI’s critical thinking (see the link to the text here: https://champaignmagazine.com/2025/03/11/ai-by-ai-critical-thinking-and-ai-otherness/). I would like to chat with you today, on that topic again. Grok-3: Hi there! I’m happy to dive into the topic of AI and critical thinking with you.… Continue reading
-
Three Dimensions of Breakthrough Needed Toward AGI and Beyond
By W.H.L. with Claude 3.7 Sonnet W.H.L.: As of now in February 2025, the mainstream and state of the art AI models are by and large Transformer based, that is, they are grounded on probability based predictions. Mere mathematics, however, may not be good enough to get us there. In addition to our current effort… Continue reading
-
On AGI and its Consciousness: Chat with Grok-3
By W.H.L. and Grok 3 W.H.L.: Hi Grok-3! With the release of you Grok-3, DeepSeek-R1, and announced upcoming GPT4.5 and then GPT 5 later this year, what’s your bet on AGI? When will that happen in your estimate? Grok 3: Hi there! Thanks for the warm greeting. Let’s dive into your question about AGI—Artificial General… Continue reading