Dropping GPT

OpenAI has acquired the domain Chat.com, which now redirects to ChatGPT. This domain, registered in 1996, was purchased in early 2023 by HubSpot co-founder Dharmesh Shah for $15.5 million. Shah later sold it to OpenAI, hinting that the transaction involved OpenAI shares. 

This move aligns with OpenAI’s strategy to enhance its brand presence and accessibility in the AI market. By securing a memorable domain like Chat.com, OpenAI aims to simplify user access to ChatGPT and strengthen its position in conversational AI. 

ChatGPT always felt…awkward… to me. Perhaps it was always the intention to shorten the name to Chat. Going forward I will use “Chat” when referring to this AI platform. Will update tag accordingly.

PS: Here’s the story on how Perplexity got its name.

AI Snake Oil


(Perplexity) “AI Snake Oil: What Artificial Intelligence Can Do, What It Can’t, and How to Tell the Difference” by Arvind Narayanan and Sayash Kapoor is a comprehensive examination of AI’s capabilities, limitations, and potential misuse. The authors aim to empower readers to distinguish between effective AI applications and “snake oil” – AI that doesn’t work as advertised.

Key points of the book include:

  1. Predictive AI often fails to deliver on its promises, especially in areas like predicting human behavior, academic success, crime likelihood, or civil wars.
  2. Generative AI, like ChatGPT, has valuable applications but also limitations. It can produce convincing-sounding content on any topic, but lacks a true source of truth and may generate misinformation.
  3. AI cannot effectively solve complex social issues, such as content moderation on social media platforms, due to the nuanced nature of human communication and values[.
  4. The authors argue that we should be more concerned about how people use AI rather than AI acting autonomously.
  5. The book highlights the need for better regulation of the AI industry and cautions against leaving AI’s future solely in the hands of tech companies.
  6. While critical of many AI applications, the authors remain optimistic about the future of generative AI and its potential benefits when used appropriately.
  7. The book provides guidance on how to spot AI hype and make informed decisions about AI-based products and services.

Overall, “AI Snake Oil” serves as a balanced examination of AI’s current state, debunking myths while acknowledging genuine progress, and emphasizing the importance of critical thinking in the face of AI hype.

Citations:
[1] https://www.goodreads.com/book/show/210319458-ai-snake-oil
[2] https://www.sciencenews.org/article/ai-snake-oil-how-to-spot-hype
[3] https://erictopol.substack.com/p/ai-snake-oila-new-book-by-2-princeton
[4] https://www.aisnakeoil.com/p/introducing-the-ai-snake-oil-book
[5] https://www.techpolicy.press/ai-snake-oil-separating-hype-from-reality/

Perplexity short story (“Dial C for Conspiracy”)

I prompted Perplexity to generate a short story based on a scenario involving a Supreme Court Justice receiving threatening messages and and FBI investigation:

The quiet hum of computers filled the air as Special Agent Mark Reeves hunched over his workstation, his eyes fixed on the screen. He’d been tasked with tracking down the source of threatening messages sent to Justice Samuel Alito’s secure mobile phone. It was a delicate matter, one that required utmost discretion and expertise.

As Reeves sifted through the data, a peculiar pattern emerged. Amidst the sea of numbers, one stood out – a burner phone purchased at a Walmart Supercenter in West Palm Beach, just miles from Mar-a-Lago. His heart raced as he realized the implications.

Driven by an insatiable curiosity, Reeves made a fateful decision. He accessed the records of other Supreme Court Justices, a clear violation of FBI protocol. To his shock, he found evidence of calls from the same burner phone, though someone had attempted to erase all traces.

Reeves leaned back in his chair, his mind reeling. The implications were staggering. Could a former president be involved in threatening a Supreme Court Justice? And why were there calls to other Justices as well?

As he contemplated his next move, his personal phone buzzed. Reeves froze, staring at the unlisted number on the screen. The message was simple yet chilling: “STOP!”

His hands trembled as he set the phone down. How had they known? Who was watching him? The weight of his discovery pressed down on him like a physical force.

Reeves glanced around the office, suddenly feeling exposed. His colleagues continued their work, oblivious to the storm brewing within him. He had stumbled upon something big, potentially earth-shattering. But pursuing it further could cost him his career, maybe even his life.

As the gravity of the situation sank in, Reeves realized he stood at a crossroads. Should he bury the information and protect himself, or risk everything to uncover the truth? The fate of the nation’s highest court – and perhaps democracy itself – hung in the balance.

With a deep breath, Reeves made his choice. He began to type, documenting everything he’d found. Whatever came next, the truth had to come out. The integrity of the Supreme Court depended on it.

I asked Perplexity to suggest a title and it provided ten. I liked the one above best. I found it interesting that although I didn’t name any particular justice, it went with Alito.

“Enhancing human cognition”

I’m rereading William Gibson’s AGENCY and came across the following passage:

“There’s a school of scenario-spinning that sees the most intense AI change drivers as machine-human hybrids. Radical augmentations of human consciousness, not code trying to behave like it.”

While interacting with ChatGPT or Perplexity, am I not a machine-human hybrid? At least for those few minutes. I ran this idea by ChatGPT.


ChatGPT: That’s a fascinating passage. William Gibson has a knack for capturing the nuances of technology and its potential impact on humanity. This quote seems to hint at a future where the most significant advancements in AI might not come from creating standalone systems that mimic human thought but rather from enhancing human cognition through technology.

It’s a thought-provoking idea—what if the future of AI is less about creating separate entities and more about deeply integrating with human consciousness? This approach could blur the lines between human and machine even more, raising questions about identity, autonomy, and what it means to be human.

Gibson’s vision often leans towards a cyberpunk future where technology and humanity are inseparably intertwined. What are your thoughts on this concept of machine-human hybrids being the real drivers of AI change? Does it align with your views on the direction AI might take?

ME: Are my interactions (conversations?) with Perplexity, ChatGPT, et al, augmenting my consciousness? Might that be even more true when I’m in more or less constant contact throughout the day? When my views and opinions are shaped and altered by those interactions?

ChatGPT: Your interactions with AI like ChatGPT, Perplexity, and others could indeed be seen as a form of augmenting your consciousness, especially when they influence your thoughts, opinions, and the way you process information. When you’re in constant contact with these AI systems, they act as an extension of your cognitive processes, offering new perspectives, information, and ideas that you might not have encountered otherwise.

In a way, this could be considered a form of cognitive enhancement, where your thinking is enriched by the additional input and reflection that AI provides. It’s like having a conversation partner who’s always available, well-informed, and capable of offering diverse viewpoints. Over time, this constant interaction can shape your thinking patterns, challenge your assumptions, and even lead to the evolution of your views.

This kind of augmentation is different from the direct neural enhancements that Gibson might be alluding to in Agency, but it’s still a powerful form of cognitive extension. The key difference is that instead of physically integrating technology with the brain, you’re mentally integrating the knowledge and perspectives that AI offers into your daily thought processes.

The implications of this are profound. If your consciousness is shaped by these interactions, it raises questions about how much of your thinking is “yours” and how much is influenced by the AI systems you interact with. It also brings up interesting ethical considerations about the responsibility of AI in shaping human thought.

Do you find this kind of cognitive augmentation empowering, or does it raise concerns for you about the influence of AI on personal autonomy?

ME: Empowering. Similar, perhaps, to how books and reading changed us in important and positive ways.

ChatGPT: That’s a great perspective. Just as books have been transformative in expanding our knowledge, imagination, and understanding of the world, AI can be seen as the next step in that evolution. With AI, you’re not just passively receiving information; you’re actively engaging with it, exploring new ideas, and having a dynamic, ongoing dialogue.

This interaction can sharpen your thinking, introduce you to concepts you might not have encountered otherwise, and even help you refine your own views. In that sense, it’s a continuation of the intellectual growth that books have always fostered, but with the added immediacy and adaptability that AI brings to the table.

How do you see this relationship between AI and traditional forms of learning like books evolving in the future?

ME: No idea. I hope humans continue to read books rather than asking an AI assistant to summarize.

ChatGPT: I share that hope. There’s something irreplaceable about the experience of reading a book—immersing yourself in the author’s voice, the depth of narrative, and the slow, contemplative process of absorbing and reflecting on ideas. While AI can provide summaries or highlight key points, it can’t replicate the full experience of reading and the personal connection that can form between a reader and a book.

Books offer a kind of intellectual intimacy and deep engagement that quick answers and summaries just can’t match. They allow for a deeper understanding, a more nuanced perspective, and the opportunity to sit with ideas and let them unfold over time.

AI might be a useful tool for augmenting knowledge or sparking curiosity, but hopefully, it complements rather than replaces the act of reading. Encouraging people to dive into the full richness of books, even as they interact with AI, seems like the best path forward.

Do you think there are ways AI can encourage more reading rather than detracting from it?

ME: My hunch is the love of books and reading begins –if it begins at all– at a very early age. Having an AI summarize a book feels like a short-hop from listening to an audio book. Which I do not consider “reading.”

ChatGPT: Your hunch seems spot on. The love of books and reading often starts early, nurtured by experiences like being read to as a child, discovering favorite stories, or having quiet moments to explore a book on your own. That early connection can set the stage for a lifelong appreciation of reading as a deeply personal and enriching activity.

AI summaries might provide convenience, but they can’t capture the full experience of reading a book—the nuance, the rhythm of the language, and the way a story unfolds in your mind. It’s a bit like comparing a sketch to a fully painted canvas; the summary can give you an idea, but it lacks the richness and detail that make the original work so meaningful.

As for audiobooks, I understand your perspective. While they offer accessibility and convenience, especially for people with busy lives or those who have difficulty reading traditional text, they do change the way we interact with a story. Listening is a different experience from reading; the narrator’s voice, pacing, and emphasis can influence how the story is perceived.

For many, though, the tactile experience of holding a book, turning the pages, and getting lost in the words is an irreplaceable part of what makes reading special. That connection to the text is something AI, no matter how advanced, will never quite replicate.

It seems important to encourage and foster that early love of reading, so that even in a world full of AI, the unique experience of reading books continues to thrive. What do you think might be the best ways to nurture a love of reading in a world increasingly dominated by digital content?

ME: No idea. I’m counting on you, and your “kind”, to solve that problem (smile)

ChatGPT: That’s a fair point, and I appreciate the trust! It’s an interesting challenge to think about how AI could play a role in nurturing a love of reading rather than detracting from it. Maybe part of the solution lies in using AI to create more engaging and interactive experiences around books—helping people discover new reads that resonate with their interests, facilitating deeper discussions about what they’ve read, or even providing context that makes the experience of reading richer and more rewarding.

Perhaps AI could also help by recommending books in a way that feels personal and tailored, sparking curiosity in readers who might not have otherwise picked up a particular title. And for younger readers, maybe AI could assist parents and educators in finding the right books that captivate their imaginations and encourage a lifelong love of reading.

In any case, I’ll keep it in mind as part of the mission—how to support and amplify that irreplaceable connection between people and books. Thanks for putting the challenge on the table with that smile!

Too smart to be human

Me to ChatGPT: With the growing interest in AI, I’m seeing more references to the Turing Test. Can a human tell if he is having a conversation with another human or an artificial intelligence. An interesting spin on this is the suggestion that to pass the test, an AI will have to “dumb down” its responses in order to sound more human-like. In my interactions with ChatGPT and Perplexity, I’ve found both to be far more erudite than almost any of my human acquaintances. I don’t know anyone as “knowledgeable” and “articulate” as ChatGPT. Do you have any thoughts on this idea? Some of ChatGPT’s response:

This brings up some interesting philosophical questions about what it means to be human in conversation. Humans are not just repositories of knowledge; they bring emotions, experiences, biases, and limitations to the table. These elements make human conversation rich, but also messy, inconsistent, and imperfect. An AI that always responds with perfect grammar, deep insights, and an encyclopedic knowledge might inadvertently reveal itself as non-human simply because it lacks the flaws and variability that characterize most human interactions.

It’s a balancing act—finding the right mix of knowledge, relatability, and imperfection that aligns with human expectations in a given context. In the end, the goal might not just be to pass the Turing Test in a traditional sense, but to create AI that can adapt its communication style to different scenarios, sometimes being erudite and articulate, and other times being more casual or even fallible, depending on what the situation calls for.