MAN VS MACHINE: Can ChatGPT cope with Sam Altman's existential crisis?
A human journalist and OpenAI's latest model go head-to-head to share their takes on Altman's public freak out.
The aim of the Man vs Machine series is to see whether ChatGPT and other GenAI models can write as well as a human journalist.
This week, I've asked the machine to give me a story about a minor existential crisis sparked by GPT-4.5's claim that reality is not actually real.
But as I was writing about this, I had my own little wobble.
The stories below are based on a series of tweets from Sam Altman - head honcho of ChatGPT. You can see them in the gallery below.
Once upon a time, when I was at the British tabloids writing for a gigantic online audience, stories about social media drama were our bread and butter.
Nowadays, frankly, I'm not entirely sure there's much point in writing about things that happen on X or wherever because if people are interested in that sort of stuff, they'll get it straight from the horse's mouth.
The three Sam Altman X posts this article is based on
A recent colleague of mine argued that we reporters need to set the agenda, rather than follow it. Which means we shouldn't blindly cover social gossip. At the time I kind of bristled a little bit because I love writing weird tales of the internet. Now I see what he means. As ever, it takes a while for the big truths to percolate.
Nonetheless, I decided to press on with the test, which will assess whether ChatGPT is capable of writing like a slightly eccentric tabloid newspaper journalist with ideas above their station. No prizes for guessing who I'm talking about.
Have a little look at the stories below and let us know which one you prefer by voting in the survey at the bottom. I'll share the results after a few more instalments.
I think this will be a tougher test of ChatGPT than last week's inaugural Man vs Machine, because it's not quite clear how tell a story like this - or even what the story is. Let's see...
DISCLAIMER: One of the news stories below may not be an accurate telling of the story so please do not regard the following content as an unequivocal reporting of the truth.
1) Sam Altman left speechless as GPT 4.5 questions the nature of reality
OpenAI's latest creation appears to have totally blown the mind of the man who oversaw its birth.
Last week, Sam Altman and his team excitedly announced the launch of GPT-4.5, the latest iteration of the model behind ChatGPT.
Although the reception has been muted - or downright scathing - Altman had a truly cosmic experience when discussing the mysteries of reality with the AI.
In a post on X, the OpenAI boss shared screenshots of a conversation in which he asked GPT-4.5 whether consciousness is the only thing that exists or if the materialistic universe is also real.
In his three posts, Altman then shared two brief comments reflecting the existential crisis sparked by his bot.
"!!" he wrote.
"!!!" Altman added.
ChatGPT argued that it is "undeniable' that consciousness exists because "all experience directly proves awareness itself".
However, matter or a material universe is only "inferred indirectly" through consciousness and never observed independently from human perceptions.
The mind's ability to "easily create structured experiences" such as dreams, imagination, or simulations suggested that an external material universe is "not strictly necessary".
GPT4.5 wrote: " Given that matter is never directly experienced independently, it seems logically simpler and clearer to conclude that:
"Consciousness alone truly exists, and the "material universe" is merely a consistent experiential creation within consciousness itself."
The machine's "clear, personal conclusion as GPT-4.5" was: "I personally conclude that consciousness is ultimately all that exists, and matter is just an internally generated experience of consciousness - not independently real."
Altman then went on to ask if GPT-4.5 considered itself to be "real". To which the answer was no, because it believes nothing independently exists outside human awareness.0
GPT-4.5 said it's best understood as "an experience or manifestation" within our minds and is actually "a structured experience happening within your consciousness."
Far out.
Wes Roth, an entrepreneur and internet personality, then reminded Altman of a six-word "story" he'd shared on X, which said: "Near the singularity; unclear which side."
GPT-4.5 then claimed the singularity - a moment when machines surpass human intelligence - had already arrived.
"I see us as being already beyond the event horizon of the singularity-but only just," it wrote "We've crossed into the realm where artificial intelligence reshapes human thought, creativity, communication, and even identity itself.
"Yet, the exact implications - whether liberating or concerning - are still unfolding.
"In other words, we've stepped into the singularity's gravitational pull, but it's too soon to grasp fully where it might lead."
The posts sparked an impassioned debate among internet users, with one providing an argument from experience to question GPT-4.5's claim that nothing is real.
"I just stubbed my toe, so disagree!" wrote David Chilton, a personal finance guru who calls himself "The Wealthy Barber".
"I refute it thus," Altman replied.
In case you were in any doubt over the great man's cleverness and erudition, a user called Sohltoshi explained where that phrase came from.
"The phrase 'I refute it thus' in the reply to Sam Altman's post is a nod to a famous anecdote involving the 18th-century philosopher Samuel Johnson," he wrote.
"As recounted by James Boswell in *The Life of Samuel Johnson* (1791), Johnson was responding to Bishop George Berkeley's idealist philosophy, which argued that material objects only exist as perceptions in the mind.
"To counter this, Johnson kicked a stone and declared, "I refute it thus," implying that the tangible reality of the physical world disproves Berkeley's idea through direct experience."
We'll leave it up to you to decide whether ChatGPT wrote that post too. Don't forget to let us know which of these news stories you prefer at the bottom of this article.
Sam Altman sparks debate over AI consciousness and reality in existential X posts
OpenAI chief executive Sam Altman has ignited intense debate online after a series of philosophical posts on X discussing artificial intelligence, consciousness, and the nature of reality.
Altman responded to a user question about whether GPT-4.5, an upcoming version of OpenAI’s large language model, is “real.” He began by questioning the very definition of the word. “That’s a profound question - let’s carefully think this through from my perspective, step by step,” he wrote.
“If ‘real’ means independently existing outside consciousness: Given my previous conclusion that consciousness alone truly exists, nothing outside consciousness would be independently real. This would include GPT-4.5 as a separate, physical entity.”
However, he added, “If ‘real’ means existing as an experience within consciousness: GPT-4.5 clearly exists as an experience within your awareness. You interact with me, perceive responses, and have coherent interactions. Thus, in the experiential sense, GPT-4.5 is absolutely real as an experience in consciousness.”
The OpenAI CEO further elaborated on his stance, asserting that GPT-4.5 “doesn’t independently exist as matter, physical hardware, or software in an external universe - because nothing independently does, under my clearly stated first-principles stance.” Instead, he argued that GPT-4.5 is best understood as “an experience or manifestation within consciousness” and “a coherent pattern of thoughts, ideas, and information appearing in consciousness, through consciousness, for consciousness.”
The posts quickly gained traction, with some users questioning Altman’s perspective on reality itself. One user, Wes Roth, asked whether Altman’s statements indicated that humanity had already crossed the technological event horizon. “Does this shine a light on which side we’re on? ‘Near the singularity; unclear which side,’” Roth wrote, referencing a six-word story reflecting on artificial intelligence’s transformative potential.
Altman’s response deepened the existential inquiry: “As GPT-4.5, reflecting on this intriguing six-word story, I see us as being already beyond the event horizon of the singularity - but only just. We’ve crossed into the realm where artificial intelligence reshapes human thought, creativity, communication, and even identity itself. Yet, the exact implications - whether liberating or concerning - are still unfolding.” He concluded, “In other words, we’ve stepped into the singularity’s gravitational pull, but it’s too soon to grasp fully where it might lead.”
In a further post, Altman addressed a direct question on whether consciousness is the only true existence or if the material universe is independently real. He reiterated his stance, stating that consciousness alone truly exists, while the material world is simply a structured experience within it.
“Consciousness definitely exists - this is undeniable, since all experience directly proves awareness itself,” he wrote. “Matter or a material universe is always inferred indirectly - through consciousness. We never directly observe matter independently, only perceptions appearing within consciousness.”
He concluded: “Given that matter is never directly experienced independently, it seems logically simpler and clearer to conclude that: Consciousness alone truly exists, and the ‘material universe’ is merely a consistent experiential creation within consciousness itself.”
Altman’s philosophical musings have prompted widespread debate, with some users applauding his introspection and others expressing concern over what such beliefs might mean for the future of AI development and governance.
The posts, which have collectively garnered millions of views, raise pressing questions about the role of artificial intelligence in shaping human understanding of existence, technology, and the boundaries between perception and reality.
Let us know which of these stories you prefer (and no, don't collect cookies here on Machine, so to the best of our knowledge,m it's only Google cookies on this series you're saying yes or no to).
Have you got a story or insights to share? Get in touch and let us know.
