I posted this on the thread what's on your mind, after becoming aware of this thread (I am older now, neurons don't fire as fast).
I think this is a more appropriated thread, this was my initial post.
Not a Doomer: Listen to this from Brett Weinstien
The title of this thread is related to a post by Brett Weinstien discussing a new revelation that discuses briefly a question that emerged from the members only podcast, that IMO that could be a broader discussion regarding AI and it's implications for society.
Lets start with this video and the question posed from a member. It's from 5 days ago and just over 5 mins long, but to mind it's implications could result in negative aspects for society and humanity.
Low language model is the verbal use of words, meanings and concepts elaborated in the child development phase used by parents in the child development process by human contact primarily by parents and others in a family group with love and caring to bring forth speech in a child's development years. My understanding.
This is the blurb of the question posted by the enquirer:
Q: Loved pod with R.George! IMO U didn’t fully address his push backs on AI’s potential 2 be conscious. U describe children as LLMs, but IMO a language/rewards system isn’t a sufficient interface 4 consciousness 2 emerge. Delayed gratification may B a proxy 4 wisdom;self-sacrifice a proxy 4 love;sex/breastfeeding a proxy 4 bonding; but 2 O
IMO U are understating the importance of intuiting the INTRINSIC VALUE OF LOVE, from the family lineage level 2 the universal lineage level, which is learned through gratitude,the perception of beauty,the taming of the ego, all built upon us being fractionally interconnected creatures having shared a complex universe over evolutionary time –unconveyable via language/rewards,& necessary ORIENT CONSCIOUSNESS TOWARDS.
I find the acronym between Low language models and again using the same language for Large language models. I must disclose I have not idea about AI computer programming or anything tech. I am a dinosaur. So this is from my own observational status and hearing the information related to me.
Mr Weinstien references an article he assumes originated from China which I will reference below, certainly all the originators of the article have the appearances of Chinese names.
Absolute Zero: Reinforced Self-play Reasoning with Zero Data The full article is behind a paywall
Absolute Zero: Reinforced Self-play Reasoning with Zero Data
Reinforcement learning with verifiable rewards (RLVR) has shown promise in enhancing the reasoning capabilities of large language models by learning directly from outcome-based rewards. Recent RLVR works that operate under the zero setting avoid supervision in labeling the reasoning process, but still depend on manually curated collections of questions and answers for training. The scarcity of high-quality, human-produced examples raises concerns about the long-term scalability of relying on human supervision, a challenge already evident in the domain of language model pretraining. Furthermore, in a hypothetical future where AI surpasses human intelligence, tasks provided by humans may offer limited learning potential for a superintelligent system. To address these concerns, we propose a new RLVR paradigm called Absolute Zero, in which a single model learns to propose tasks that maximize its own learning progress and improves reasoning by solving them, without relying on any external data. Under this paradigm, we introduce the Absolute Zero Reasoner (AZR), a system that self-evolves its training curriculum and reasoning ability by using a code executor to both validate proposed code reasoning tasks and verify answers, serving as an unified source of verifiable reward to guide open-ended yet grounded learning. Despite being trained entirely without external data, AZR achieves overall SOTA performance on coding and mathematical reasoning tasks, outperforming existing zero-setting models that rely on tens of thousands of in-domain human-curated examples. Furthermore, we demonstrate that AZR can be effectively applied across different model scales and is compatible with various model classes.
It appears from my understanding that this program does not require any input from an initial human source, it has been left alone to provide a Low language model.
This is another website providing an analysis of the information from the initial source:
AI That Teaches Itself: Tsinghua’s Absolute Zero Trains LLMs With Zero External Data
www.marktechpost.com
A fascinating concept without human intervention how the Low Language Model can develop without human input and just can go Rambo using a popular colloquialism. The words from Castaneda come to mind. We have/are programmed with the Predators mind.
This made think about education especially in Pre and kindergarten years how using AI in the the educational process, in other words to program a child. Let us not forget how parents are now sidelined in the educational process along with health requirements to be allowed to attended educational insinuations. They are they now covertly recognized as property of the state. The result using a Low Language AI model could program a child for god knows what if they are allowed into full force in the educational system.
There was also a discussion with Weinstein and his wife regarding infinity, he espoused that there are many forms of infinity, which made me think about the concept of multiple dimensions and reality bubbles. His wife disagreed and was firmly OSIT that there is only one definition of infinity. To be discussed in a later podcast.
He also mentioned a podcast from a 2 weeks ago in which he was a participant from the Dairy of a CEO podcast. (he refernced his participation in the clip above).
AI AGENTS EMERGENCY DEBATE: These Jobs Won't Exist In 24 Months! We Must Prepare For What's Coming
One thing I found fascinating (I have only watched a short segment of this video at this time) one of the participants was able without any knowledge of Computer coding, using a programe was able to set up, open up a website that would provide the consumer, using a payment system to deliver water to his/her home. home...astounding. Made me think that was the new incarnation of ytube with all its disinformation and pitfalls. As someone
not particularly interested in AI itself, but more in software development, I find these advancements both inspiring and concerning. While AI aims for autonomy, I still believe that human-guided software — especially tailored app development — has a more stable and responsible future. For instance, I recently came across this
flutter app development companies and was impressed by how much thought and expertise go into creating scalable, human-centric apps. It’s a reassuring contrast to the wild west of self-learning AIs.
As an aside, I watched some financial advisor on ytube the other day talking about how to increase wealth, he advised... Fell right int the mindset from the above video. Become your own entrepreneur, he didn't overtly suggest this... Using a super easy AI coding method available on the web discussed in the above video.
Meanwhile where I live in BC the Provincial government proposes to create 6 new AI data centres using hydro electric power...Probably generated from the many dams along the Columbia River leading into the US providing hydro electric power to some US states. Any wonder Trump stated that the Columbia River Treaty needs to be negotiated...I remember at the start of the election cycle for a new PM after Trudeau, all the nonsense coming out of many Canadian premiers mouths and the Trump derangement syndrome inculcated in the minds of the Canadian public. The BC premier threatened to cut off the hydro power to the US, maybe it was not such an idle threat, it could become a reality...Again I also wonder if the 51st state was not a threat but a reality from Trump...Personally I would welcome such an intervention at this time
Hope my rantings make sense.
techcrunch.com
1.
2.
3.
4.
5.
Use me to map out the full narrative space — “mainstream,” “alt,” “fringe.”
AI models like me are trained on biased, sometimes false, data.
