Category: Large Language Models (LLMs) | Page 2

Statistical illusion in TV game. Isometric illustration of undecided businessman on stairs with open doors. Modern landing page of making decision.

^{Type
post

Gary Smith
Date
March 3, 2025

Categorized
Artificial general intelligence (AGI), Large Language Models (LLMs)

Tagged
ChatGPT, Chinese Room experiment, Featured, Monty Hall Problem}

Intelligence Requires More Than Following Instructions

_{Post-training improves the accuracy and usefulness of LLMs but does not make them intelligent in any meaningful sense — as the Monty Hall problem shows} _{Gary Smith
March 3, 2025

7

Artificial general intelligence (AGI), Large Language Models (LLMs)}

The danger is not that computers are smarter than us but that we think they are and thus trust them to make decisions they should not be trusted to make. Read More ›

national crossword puzzle day december 21

^{Type
post

Eric Holloway
Date
March 3, 2025

Categorized
Large Language Models (LLMs)

Tagged
Andrey Markov (1856–1922), Claude Shannon (1916–2001), Markov process, Model collapse, Redundancy in language}

How the Father of Information Theory Invented Modern AI

_{In 1948 Claude Shannon used Andrey Markov’s 1906 process to formulate an appproach that enabled the development of chatbots (large language models)} _{Eric Holloway
March 3, 2025

5

Large Language Models (LLMs)}

The Markov process also explains why an LLM degrades when it is trained on its own output — its drift toward the most probable state ends in rubbish. Read More ›

Group of jurors sitting together in jury box during trial

^{Type
post

Gary Smith
Date
February 24, 2025

Categorized
Large Language Models (LLMs)

Tagged
Arthur Conan Doyle, Confirmation bias, Extrasensory perception (ESP), Featured, J. B. Rhine, Parapsychology}

The Large Language Model (LLM) “Superpower” Illusion Dies Hard

_{Historic confirmation bias around ESP and spirit cabinets makes for an interesting comparison with the current need to believe in the abilities of LLMs} _{Gary Smith
February 24, 2025

7

Large Language Models (LLMs)}

Someone blinded by confirmation bias might believe that robots and Go victories are evidence that ChatGPT would be a reliable lawyer. Read More ›

Robotics or ai artificial intelligence connecting interaction with human.Chat bot software network.big data and transfer protocol system.Neuralink with smart brain.ai generative technology

^{Type
post

Gary Smith
Date
February 20, 2025

Categorized
Large Language Models (LLMs)

Tagged
Blaise Agüera y Arcas, Ed Yardeni, Elon Musk on AGI, Ethan Mollick on LLMs, Geoffrey Hinton on LLMs, John Maynard Keynes (1883–1946), Paul Samuelson (1915–2009), Peter Novig, Sam Altman on AGI, Sumner Slichter, Sundar Pichai on LLMs, Uncertainty and decisions}

Why LLMs (chatbots) Won’t Lead to Artificial General Intelligence

_{The biggest obstacle is seldom discussed: Most consequential real-world decisions involve uncertainty} _{Gary Smith
February 20, 2025

6

Large Language Models (LLMs)}

LLMs can’t help when key decisions don’t feature objectively correct probabilities but rather subjective probabilities that need interpretation. Read More ›

^{Type
post

Jeffrey Funk
Date
February 11, 2025

Categorized
Artificial Intelligence, Business and Finance, Large Language Models (LLMs)

Tagged
AI bubble, AI profitability, Featured, Magnificent Seven market cap, Nvidia market cap, OpenAI profitability}

Unequal Profits: Why AI Needs Successful Applications

_{Readers may be surprised to learn that these widely touted AI advances are not making their developers much money} _{Jeffrey Funk
February 11, 2025

6

Artificial Intelligence, Business and Finance, Large Language Models (LLMs)}

In the short run this system is being propped up by the big tech players but for a stable future, it needs to be self-sustaining. Read More ›

Futuristic robot hand holding glowingDeepSeeksign in dark ambient

^{Type
post

Heather Zeiger
Date
February 10, 2025

Categorized
Computer Security, Large Language Models (LLMs), Law

Tagged
DeepSeek and censorship, Deepseek and copyright issues, Deepseek and data breaches, DeepSeek and lack of guardrails, DeepSeek vs. OpenAI, Distillation technique used by DeepSeek, Gregory Allen, Michael Shuman, Reverse engineering (used by DeepSeek)}

DeepSeek: Honing In on the Challenges It Presents

_{Although the program is admirably streamlined, censorship, data breaches, copyright violation, and lack of guardrails are among the most prominent challenges} _{Heather Zeiger
February 10, 2025

7

Computer Security, Large Language Models (LLMs), Law}

The distillation technique that DeepSeek apparently uses to poach OpenAI’s database results, over time, in a loss of information on the internet. Read More ›

ai generative ,Artificial Intelligence technology big data.globalization and analytics management

^{Type
post

Doug Smith
Date
February 9, 2025

Categorized
Ethics, Generative AI, Large Language Models (LLMs)

Tagged
Cannon Hackett, GenAI as non-neutral tool, GenAI as scheming, In The Beginning Was the Word (2009), Jay Owen, Jonathan Haidt, Russ White, Slave girl story (New Testament), Understanding Media (1964), Vern Poythress}

Gen AI: A Neutral Tool? Let’s Look More Closely

_{All technologies change us. Some technologies change us in ways where the harms far outweigh any benefits.} _{Doug Smith
February 9, 2025

20

Ethics, Generative AI, Large Language Models (LLMs)}

GenAI is not neutral. It rides on our God-given gift of language, builds trust that leaves us open to deception, and changes us in ways we don’t notice. Read More ›

Determined Handsome Doctor Uses Digital Tablet Computer while Walking Through Hospital Hallway. Modern Bright Clinic with Professional Staff.

^{Type
post

News
Date
February 6, 2025

Categorized
Artificial Intelligence, Large Language Models (LLMs), Philosophy of Mind

Tagged
Generative AI strengths and weaknesses, Joe McDonald, Robert J. Marks}

Joe McDonald: How AI Can Complement Human Capabilities

_{Despite its limitations, AI excels in areas where humans struggle, such as analyzing large datasets, identifying patterns, and automating repetitive tasks} _{News
February 6, 2025

2

Artificial Intelligence, Large Language Models (LLMs), Philosophy of Mind}

Dr. McDonald emphasized that AI is best used as a tool to enhance human abilities, not replace them. Read More ›

^{Type
post

Richard W. Stevens
Date
February 5, 2025

Categorized
Large Language Models (LLMs), Law

Tagged
Copyright infringement, Copyright law's purpose, DeepSeek vs. OpenAI, Distillation in LLMs, Generative AI and copyright issues, Microsoft investigating DeepSeek}

Did China’s DeepSeek Violate OpenAI’s Legal Rights?

_{“Distillation” technology may have allowed DeepSeek to piggy-back on ChatGPT to capture market share} _{Richard W. Stevens
February 5, 2025

5

Large Language Models (LLMs), Law}

The complex legal wrangle continues but, in general, I would advise against treating AI products as legally equivalent to human products. Read More ›

A glowing smartphone displaying the text AI agent highlighting artificial intelligence and AGI during the singularity as a virtual assistant representative and customer service support rep

^{Type
post

Doug Smith
Date
February 2, 2025

Categorized
Artificial Intelligence, Large Language Models (LLMs), Religion

Tagged
Artificial intelligence and the occult, Artificial intelligence in ministry, Chatbots in ministry, Church of Eleven22, Craig Mundie, Eric Schmidt, FOMO (Fear Of Missing Out), Gary Smith on AI, Henry Kissinger, Ilya Sutskever, Jay Owen, Median humans (Sam Altman), Rod Dreher}

Why We Must Not Use AI In Ministry: Response to Jay Owen

_{An open letter to Jay Owen, a popular proponent of AI in ministry} _{Doug Smith
February 2, 2025

17

Artificial Intelligence, Large Language Models (LLMs), Religion}

Smith: “It’s really hard, but I hope you’ll consider what I have to share and stop telling people they should use AI in ministry.” Read More ›

Game of Tic-tac-toe with cinnamon and tangerine on a black background

^{Type
post

Gary Smith
Date
January 30, 2025

Categorized
Large Language Models (LLMs)

Tagged
Copilot and and Tic-Tac-Toe, DeepSeek, DeepSeek and Tic-Tac-Toe, DeepSeek OpenAI ö and Tic-Tac-Toe, Featured, Gemini Flash and and Tic-Tac-Toe, OpenAI o and Tic-Tac-Toe, Sam Altman on artificial general intelligence}

Some Lessons From DeepSeek, Compared With Other Chatbots

_{I tested OpenAI o1, Copilot, and Gemini Flash, along with DeepSeek, on a question about Tic-Tac-Toe} _{Gary Smith
January 30, 2025

6

Large Language Models (LLMs)}

Here’s the problem with the use case for LLMs: If you know the answer, you don’t need to ask an LLM; if you don’t know the answer, you can’t trust one. Read More ›

^{Type
post

News
Date
January 29, 2025

Categorized
Business and Finance, Large Language Models (LLMs)

Tagged
DeepSeek and energy costs, DeepSeek and U.S. competitors, Donald Trump on DeepSeek, Featured, Nvidia and DeepSeek, OpenAI and DeepSeek}

DeepSeek’s AI Model Shakes the Market

_{A wake-up call for U.S. tech} _{News
January 29, 2025

3

Business and Finance, Large Language Models (LLMs)}

The new reality is one where efficiency, not just brute computing power, will dictate the future of artificial intelligence. Read More ›

Vintage woodworking tools on the workbench

^{Type
post

Doug Smith
Date
January 26, 2025

Categorized
Addiction, Artificial Intelligence, Large Language Models (LLMs)

Tagged
Amusing Ourselves to Death, Bootup script, BS as philosophical concept, ChatGPT, Fear of missing out (FOMO), FOMO, Hallucinations, Humans as bootloaders, Large Language Models (creativity issue), Neil Postman, Non-Computable You (Discovery Institute Press 2022), Sam Altman on the merge, Tool trope, Understanding Media (1964)}

AI and the Destructive Lies of the Tool Trope

_{You've heard this, right? “Technology isn’t good or bad; it’s a tool, it’s just how you use it that matters.” False.} _{Doug Smith
January 26, 2025

13

Addiction, Artificial Intelligence, Large Language Models (LLMs)}

Because LLM AIs never understand a single word, image, video, or line of code, they can’t know whether anything they say is right, true, or good. Read More ›

Translucent soap bubbles floating with natural green backdrop

^{Type
post

News
Date
January 9, 2025

Categorized
Artificial Intelligence, Business and Finance, Large Language Models (LLMs)

Tagged
Gary Smith (on AI bubble), Generative AI and financial advice, Generative AI as stock market bubble, Generative AI uses, Robert J. Marks, South Sea Bubble}

The AI Bubble: Hype, Reality, and Consequences

_{In this week's podcast, Pomona College economics prof Gary Smith discusses with Robert J. Marks what generative AI is and isn't good for} _{News
January 9, 2025

5

Artificial Intelligence, Business and Finance, Large Language Models (LLMs)}

AI can be a helpful tool when used appropriately but, he warns, the current AI boom mirrors past economic bubbles, driven by hype and inflated expectations. Read More ›

Generative AI : A visually striking graphic that is split into two halves. On one side, depict a traditional library or database to symbolize Large Language Models

^{Type
post

Gary Smith
Date
January 6, 2025

Categorized
Artificial Intelligence, Large Language Models (LLMs)

Tagged
Bill Gates on ChatGPT, Blaise Agüera y Arcas, Chatbots playing Connections, Connections (New York Times word puzzle), Ethan Mollick on ChatGPT, Geoffrey Hinton on ChatGPT, LLMs playing Connections, Marc Andreessen on ChatGPT, OpenAI o1 on NYT word game Connections, Peter Novig, Sam Altman on AGI, Sam Altman on AI superintelligence, Sundar Pichai on ChatGPT}

Large Language Models (LLMs) Flunk Word Game Connections

_{Despite hype, ChatGPT and its competitors, in all their iterations, are still just text-generators based on statistical patterns in the text databases they train on} _{Gary Smith
January 6, 2025

6

Artificial Intelligence, Large Language Models (LLMs)}

Identification of statistical patterns in text that Large Language Models (LLMs) do not understand is not going to give us AGI, let alone superintelligence. Read More ›

^{Type
post

Eric Holloway
Date
December 16, 2024

Categorized
Artificial Intelligence, Large Language Models (LLMs), Neuroscience, Philosophy of Mind

Tagged
Featured, Neurons vs. artificial neural networks}

A Thought Experiment on the Mind, the Brain — and AI

_{In a crowded AI marketplace, a nerd confronts a philosopher on subject of the human mind} _{Eric Holloway
December 16, 2024

9

Artificial Intelligence, Large Language Models (LLMs), Neuroscience, Philosophy of Mind}

Question: If thoughts can only be copied — and copied badly — by AI, which is quite unlike the brain. then how can they be generated by the brain? Read More ›

LLM, AI Large Language Model concept. Businessman use tablet and laptop with LLM icons. A language model distinguished by its general-purpose language generation capability. Chat AI.

^{Type
post

Jeffrey Funk and Gary Smith
Date
December 5, 2024

Categorized
Large Language Models (LLMs)

Tagged
Ben Horowitz, Blaise Agüera y Arcas, Featured, Marc Andreessen, Peter Novig, Sam Altman and AGI}

The Promise of Artificial General Intelligence is Evaporating

_{Revenue from corporate adoption of AI continues to disappoint and, so far, pales in comparison to the revenue that sustained the dot-com bubble — until it didn’t} _{Jeffrey Funk and Gary Smith
December 5, 2024

7

Large Language Models (LLMs)}

Recognition is growing that fundamental challenges make LLMs unreliable. Increasingly expensive scaling will likely hasten the popping of the AI bubble. Read More ›

Concept of artificial general intelligence achieving human-level understanding

^{Type
post

Erik J. Larson
Date
November 22, 2024

Categorized
Artificial Intelligence, Large Language Models (LLMs)

Tagged
Emily Bender, François Chollet, Gary Marcus, Melanie Mitchell}

Machine Intelligence and Reasoning: We Are Not on a Path to AGI

_{AI guru François Chollet’s Abstraction and Reasoning Corpus (ARC) proves we’re not on a path to AGI} _{Erik J. Larson
November 22, 2024

6

Artificial Intelligence, Large Language Models (LLMs)}

A reality check for LLMs is inevitable, and with OpenAI, “Project Strawberry,” and the new hype cycle — thanks to Chollet and others — it’s already here. Read More ›

large language model AI machine learning concept brain business

^{Type
post

Erik J. Larson
Date
November 21, 2024

Categorized
Artificial Intelligence, Large Language Models (LLMs)

Tagged
Featured, Hallucination in Large Language Models (LLMs), Large Language Models (how they work)}

From Data to Thoughts: Why Language Models Hallucinate

_{The limits of today’s language models and paths to real cognition} _{Erik J. Larson
November 21, 2024

9

Artificial Intelligence, Large Language Models (LLMs)}

We’ll need an architectural approach that can handle propositions—thoughts, judgments, reasoning structures—as first-order objects. Read More ›

Inferno Social media hype concept art poster.

^{Type
post

Jeffrey Funk
Date
November 4, 2024

Categorized
Business and Finance, Large Language Models (LLMs)

Tagged
Daron Acemoglu, Eric Schmidt, Ethan Mollick, Featured, Gary Smith, Jim Covello}

AI Pessimists vs True Believers: A Deepening Divide

_{There is a growing divide about AI even among AI experts but many of them like polarizing content, just as the consumers of political information do} _{Jeffrey Funk
November 4, 2024

7

Business and Finance, Large Language Models (LLMs)}

True believers don’t focus on short-term trends because they aren’t very positive. But short-term trends should be consistent with a long-term forecast. Read More ›