Видео

Completely agreed on the "job" part. More and more of the tasks that entrepreneurs carry out will be automated too. But I believe the ownership, the steering vision and the visible face of the endeavour is most likely to remain human.

Great point! If nothing else does, the continued progress in computing power per watt will do it :D But the trend has been that once new capabilities are unlocked, the models simply get smaller and more efficient as we optimize. The attention mechanism in transformer-based LLMs has been one of the main areas of optimization. Check out Flashattention and Longformer as examples.

Agreed, Google is catching up fast with OpenAI, and in many ways already ahead! Veo2 is absolutely insane

Great points all! According to OpenAI's own "Stages of AI" evaluation framework, AGI is not a binary threshold, where you either have it or don't. They present it is a continuum and the agentic part is key to moving to "level 3" out of 5. I find both OpenAI's and DeepMind's frameworks for this rather fluffy and vibes-based but we don't really have anything better. Fwiw I personally think there's not difference between AI and AGI. The "general" part is another smooth gradient, and it's been climbing ever since convnets. Transformers just made big neural networks generalise even better. OpenAI's stages of AI: www.tomsguide.com/ai/chatgpt/openai-has-5-steps-to-agi-and-were-only-a-third-of-the-way-there DeepMind's levels of AGI: arxiv.org/pdf/2311.02462

That's a great question! The most insightful take on that that I've seen is this one: x.com/kimmonismus/status/1871364922120261783 Software development is definitely going to change entirely and will be way less about the code and requirements gathering, testing, maintenance may be more what the human does.

Plot twist! 😅

I don't know. Maybe. But it's the same as with abiogenesis. No one can explain to date how more of the same didn't just made inanimated matter more complex (improve the stochastic process) but turn into life (become AGI). I guess you could say it's both a matter of definition, but I also think that would not fly with common conceptions of both life and intelligence. Anyway. Great discussions for 2025! Merry Christmas!

Stochastic literally just means statistical. Our brains "stochastically chain stochastic outputs." Too. I personally do not consider o3 to be AGI, but when AGI does occur, it will *definitely* be "stochastic".

@@brandongillett2616 that's a good point, but it seems that human stochastic thinking is multi dimensional (including dimensions which could be considered stochastic but commonly aren't), whereas AI is limited to one dimension, "mathematical stochastics", no?

I asked the Gemini Thinking model you recommended below if it considered the ability to make a philosophical "judgement about itself" to be the criterion for AGI. Here's the answer: Okay, here's the conclusion in English, summarizing the key points of our discussion: Conclusion: While there's no definitive answer, it's a strong hypothesis that the ability to make a "genuine" judgment about oneself is likely a necessary or at least a highly probable characteristic of Artificial General Intelligence (AGI). This is because: Self-reflection and awareness: Genuine self-judgment implies a level of self-reflection and awareness, crucial for an intelligence reaching or surpassing human cognitive abilities. Learning and improvement: To truly learn and improve, an AGI would likely need to evaluate its own performance and reasoning processes in a meaningful way, beyond just objective metrics. Adaptability and flexibility: Assessing its own strengths and weaknesses in new contexts would be vital for an AGI to adapt and solve problems creatively. Understanding human values and ethics: Being able to evaluate its own actions and motivations from a certain perspective would be crucial for ethical decision-making. However, there are also alternative perspectives: Functional equivalence: AGI might achieve human-level performance without replicating the exact same internal processes or subjective experiences, potentially evaluating itself through sophisticated algorithms. Defining "genuine": The very definition of a "genuine" judgment is complex and philosophical. Perhaps AGI could develop a functionally equivalent form of self-judgment that differs in nature from human judgment. Ultimately, the question of whether genuine self-judgment is required for AGI highlights the core of what we understand as intelligence and remains a subject of ongoing research and debate. It forces us to consider what fundamental capabilities we expect from a truly intelligent artificial system.

And it's going to get crazier!!

@@goktugyil Thank you! 🙏🏼 Hahaha, for now the best we have, as per Dan's comment, is Gemini 2.0 Flash Thinking. Go to aistudio.google.com, it's free!

@@DanDascalescu-dandv Great summary! 😁 Using Gemini 2.0 Flash Thinking is how I envision we'll experience o3 or its equivalent in ~1 year from how. In summary, very fast, very cheap! :D

Hahaha that's a useful tip thank you! 😂🙏🏼

🫡🙇🏻‍♂️🤖👑

Thanks a bunch!! 🙌🏼

@@Makurashini thank youuuuuu 🙌🏼🥳

@@drdanponders Y vi también la entrevista, jejeje Abrazote!

Thank you so much! 🙏🏼

One could not ask for higher praise! 🥹 Thank you so much DrWaku, this means a lot! 🙏

Thanks a bunch! 🙌🏼

Thank you Nicola! 🙌🏼

Exponential thanks! 🙌🏼

Thank you! 🕺

@@hedgyness Thank you muchly kind sir! I need to make my virtual twin like you did and see how engaging that one is :D

@@yaeyeah96 Great question! The headlines I've been seeing give me confidence that we're a couple of years away from extremely cheap batteries for grid storage. As I show in the video, solar photovoltaic installed capacity is growing exponentially. The only issue is that it's intermittent, so we currently still need to burn fossil fuels to fill in the missing capacity (because we chose not to develop more nuclear fission, an extremely poor decision in my view). Batteries at scale will fix all of this.

@@BalintBertok m.ruclips.net/video/npaNs11vSgU/видео.html

Thanks a bunch! 🙏🙌🏼

[ ] I'm not a bot

Thank you! I'll make sure to make more of these :D

🙏❤

Thank you!!🙏

"Are you an ultrapreneur yet??" - high expectations Asian father

*puts on ultrapreneur gloves*