Subconscious AI’s Post

The "Illusion of Knowledge" and a Hypothetical Economy of Algorithms I recognize it: I'm absolutely fascinated by this apparent access to knowledge, and we are just getting started. However, after some time, I've come to believe that generative artificial intelligences give us with an "illusion of knowledge." We operate as if we have this vast knowledge, but we do not. There is no real assimilation. To be considered knowledge, it must be connected and allow us to make both accidental and intentional connections. In this sense, the use of generative artificial intelligences creates a dependency by providing contributions that are, let's say, powerful but not refutable, not necessarily true, and not solid, as they lack valid logic beyond mathematics and often lack true premises. There are many human tasks that can be supported by these vaporous, creative, and almost random connections. But, ultimately, if we want these systems to be truly useful, we will need certainty; we will need trust. Beyond this tipping point, if we look at human beings from a cultural and biological perspective, we can see them as a kind of algorithm—a complex one, but an algorithm nonetheless. From this perspective, we can imagine a combination of biological human and artificial algorithms working together, eventually leading to an economy of algorithms. This is a bold proposition, given the increasing importance of algorithms in the modern economy. Understanding algorithms as building blocks or "atoms" that make up more than just artificial intelligence opens the door to more fluid collaboration. The question remains: Will humans be able to evolve—cognitively, culturally, and perhaps even biologically—under this new balance of algorithmic interactions?

Artificial Intelligence as our future prediction tool?

Computer scientists working on LLMs should look to social science methodology. We work with stochastic, black-box systems (yes, humans), too, & smart folks have developed a ton of methods for analyzing this sort of data in rigorous ways. AI is not people, but the methods will help. Also social scientists working on LLMs could borrow a few ideas from computer scientists as well - ablation tests, for example, are a clever way of examining mechanisms.

“Another flaw in the human character is that everybody wants to build and nobody wants to maintain.” - Kurt Vonnegut. I see AI ethics as a problem of maintainence. The Hype Cycle around AI naturally privileges a surge in development. But ethical considerations — bias, security, privacy, misinformation etc — are what requires vigilance and maintenance. For the most part, I respect what the companies behind these frontier large-language models (ChatGPT, Claude, Gemini) are attempting to do to stay moral. Navigating this Byzantine haze warrants some degree of empathy. Despite the many lawsuits levied at OpenAI; despite the Scarlett Johannson “Sky” voice debacle; despite Perplexity and Forbes’ latest dispute; despite the miasma of resentment between the artistic community and Stable Diffusion; despite piracy and illegal scraping and pernicious Instagram settings and James Blake and Suno and Udio and AI art competitions and anti-AI rallies and Air Canada refund policies and Musk v. Altman and Clearview v. Social Media… and all the rest of it. Despite that, I’ve been pleasantly surprised with the way the AI community have attempted to address ethical issues. Anthropic’s principle-led “Constitutional AI” charter exemplifies a human-centric approach, acknowledging complexities and motives behind queries. Ask Claude if the Earth is flat and you’ll see what I mean. Gemini and ChatGPT, while criticised for lack of transparency in training data, strive to mitigate misinformation through robust content filters, disclaimers, and context. They excel in practical security with encryption, access controls, and regular audits. Both remain compliant with GDPR and CCPA regulations. What they lack in bias-mitigation they make up for in practical security. Governing this space is intrinsically challenging, with our regulatory desires clashing with the rapid pace of innovation. While no one wants to die at the hands of The Singularity, overly restrictive laws could stifle the exploration of the world at our fingertips. When the CEO of OpenAI himself says, “my worst fears are that we cause significant harm to the world,” yet continues pushing for GPT-5, it underscores the complexity of balancing innovation with ethical responsibility. Yes, it’s somewhat alarming that the “Superalignment” team at OpenAI imploded, and that the company has since been haemorrhaging safety-conscious employees. But at the same time the company famously promises to defend its enterprise customers from all copyright battles, no matter the cost, akin to Airbnb insuring its hosts up to $1m. It’s confusing a confusing smoosh of priorities. But I enjoy these contradictions. They’re human. And as Ethan Mollick’s post shows, we should be looking at the entire AI endeavour through a social scientific lens anyway — it’s not just the machines that are awkward, black boxes of weirdly-behaving machinery, but the humans behind them too.

Just completed Dr Li Fei-Fei's "The Worlds I See" over the weekend. Great read for the start of the year! Interesting retrospective on the early days of artificial intelligence and how research on a few fronts (neural networks, computer vision, data collection) converged rather serendipitously, thanks to compute power reaching a milestone and crowdsourcing platforms like Amazon Turk making big data collection humanly possible. The ImageNet competition was almost discontinued as interest waned: "SVM in particular was more robust than I gave it credit for, offering a safe haven for entrants and forestalling the aggressive innovation I dreamt of. For two years running, well-worn algorithms had exhibited only incremental gains in capabilities, while true progress seemed all but absent. Worst of all, participation was already dropping, and precipitously: registrations fell from 150 to 96 in the second year, and the entries themselves fell from 35 to just 15." There's the immigration story of her family's journey from China to US which weaves a personal dimension throughout the book. Despite the hardships - working at a laundry shop, calculator with a malfunctioning TAN key, not being able to find any grad students interested in data tagging, these episodes shape a very hopeful conclusion of her next North Star: "reimagining AI from the ground up as a human-centered practice" https://lnkd.in/gRSKnp_s #goodreads #ai #computervision #biography

Computer scientists working on LLMs should look to social science methodology. We work with stochastic, black-box systems (humans), too, & smart folks have developed a ton of methods for analyzing this sort of data in rigorous ways. AI is not people, but the methods will help.

Does anyone know this coward? He refused to answer or respond to our definiton and a direct ask; he is purposely avoiding so he is NOT wrong, or admits we are correct. This is why people are f*cking awful This is bad for real science when frauds like this clown can't engage in scientific discourse. Thomas G. Dietterich @tdietterich Distinguished Professor (Emeritus), Oregon State Univ.; Former President, Assoc. for the Adv. of Artificial Intelligence; Robust AI & Compute. Sustainability -- quote defining intelligence in those terms is non-sensical and unmeasurable; try ours instead ZEUSFYI: We define generalization in the context of intelligence, as the ability to generate learned differentiation of subsystem components, then manipulate, and build relationships towards greater systems level understanding of the universal construct that governs the reality The purest test for generalization ability; create a construct with systemic rules that define all possible outcomes allowed; greater ability to predict more actions on first try over time; shows greater generalization; with >1 construct; ability to do same; relative to others.

Coming up tonight! In this edition of the Living Earth Youth Dialogues, Jared Morningstar will raise the topic of ecological concerns involved with the rise of generative artificial intelligence. Do AI technologies have the potential to solve major aspects of the climate crisis? What is the relationship between human intelligence, creativity, scientific discoveries, and artificial intelligence? How does AI connect with issues of green energy and humanity’s carbon footprint? We will explore these issues and more in this dialogue, developing a more sophisticated and cautious understanding of artificial intelligence in light of ecological thinking. Learn more & RSVP: https://buff.ly/46CQSZ5

The 2024 DataConnect Conference is an annual, global event dedicated to amplifying diverse voices in data, analytics and AI. Join us in #cbus on July 11th & 12th to hear from folks like Hoa Duong during her talk "GANs for Causal Inference: Harnessing Conditional Independence." Throughout the talk, Hoa will cover: ✨ Generative AI in Causal Inference: The talk explores how Generative AI, specifically Generative Adversarial Networks (GANs), is being adapted to tackle challenges in Causal Inference, with applications in Economics and Political Science. ✨ Development of CausalGANs: Through a year-long research effort, the team is developing CausalGANs, a novel technique that modifies GAN architecture to address the "Missing Counterfactuals" problem in Causal Inference, which is crucial for drawing accurate causal conclusions. ✨ Comprehensive Research Methodology: The research involves establishing a new framework, creating a specific notation system, and performing extensive mathematical proofs. The results are validated through over 200 parallelized experiments on High Power Computing systems for each set of parameters. Don't wait any longer -- grab your ticket today 🎟️ https://bit.ly/4biNq6A

AI and DEI - important to pay close attention.