Dreaming Up the Future with AI

Stay tuned

Don’t miss out on the latest issues. Sign up now to get access to the library of members-only issues.

jamie@example.com

The Current State of AI: A Flawed Focus

The Current State of AI: A Flawed Focus

After playing around with LLMs (Large Language Models) extensively for two years now, I’ve noticed that the direction of the AI industry—along with the understanding of use cases, capabilities, and public perception—has quite possibly become worse than it was early on.

The Myth of Human-Level Intelligence

The heavy focus on LLMs having or reaching “human-level intelligence” is a flawed approach to truly making them useful. Defining human intelligence itself is extremely difficult. IQ tests are certainly one of the worst ways to do so. Many people are highly intelligent in one area of life but lack what others consider very basic intelligence in other areas. Intelligence isn’t monolithic—it comes in many forms, from logical reasoning to emotional intelligence. Comparing LLMs to humans in this way is like comparing apples to a whole orchard of fruit.

It’s ultimately impossible and unproductive to analyse whether LLMs have human-like intelligence. They certainly know more raw information than an average human and can perform more tasks than any human can. Yet, they might struggle with something as basic as counting the number of r’s in “strawberry”—but then again, how many humans actually know that strawberries aren’t even a berry (while bananas are)? The comparison isn’t just unhelpful, it’s misleading.

It’s also important to note that while LLMs out of the box can often sound dry, uncreative, and unemotional, they can become extremely creative, highly emotionally intelligent, and even write emotionally charged content when properly guided. With the right prompts and direction, LLMs can produce output that is better in many ways than what most humans can achieve. This ability to transform into a powerful creative partner is what makes them truly fascinating and valuable.

AI as a Co-Pilot, Not a Replacement

The real value of AI lies in being an assistant for us—a co-pilot, not a replacement. In an ideal world, AI would take over the tasks we don’t enjoy, like chores, which require manual labour and dexterity—forms of intelligence often overlooked but deeply connected to human capability. AI would also handle tasks that are beyond human ability—like analysing massive datasets in seconds—just as computers always have.

It’s these kinds of applications that make LLMs most fascinating: when they combine skills to create something unique, interesting, and creative in ways that people simply can’t. For instance, LLMs can help a medical professional analyse massive amounts of research papers to identify connections or help artists generate variations of their art that spark new directions. Or describe dry topics in truly unique, engaging and creative ways. These aren’t replacements for human creativity or expertise; they’re augmentations that amplify what we’re capable of achieving.

Misleading Metrics and the Real Value of AI

The obsession with human-level intelligence often leads us astray from the more realistic and practical uses of AI that have true transformative potential. Currently, many AI evaluations are based on metrics like benchmarks or the ability to pass certain exams. But just as an IQ test doesn’t measure a person’s overall ability to thrive in the real world, these metrics often don’t reflect an LLM’s actual usefulness.

Rather than getting caught up in a quest for AGI or worrying about AI replacing jobs wholesale, we should focus on the use cases that genuinely improve quality of life, as well as write content in a way that would be extremely difficult for humans to do, or would take a very long time to get right.

The Future of AI: Augmenting Human Potential

Looking ahead, the true promise of AI isn’t about creating machines that think exactly like us. It’s about augmenting our own abilities and freeing us from mundane tasks so that we can focus on what makes us human—creativity, empathy, and complex problem-solving. We have the ability to create a new type of intelligence to coexist alongside us, rather than replace us. If we think of LLMs not as something trying to be human, but as an incredible tool that enhances human potential, we’ll unlock far more value from this technology than any metric of human-level intelligence could ever show.


Paxman talks AI with Bowie

This interview between Jeremy Paxman and David Bowie often does the rounds, illustrating how Bowie saw the potential for the internet to be a lot more than it currently was in 1999. I used ChatGPT to act out an interview between the two of them, and AI Bowie's reasoning is pretty spot on.

ChatGPT Generated Interview:

Jeremy Paxman: (Leaning forward, eyebrow raised in that classic Paxman way) David, let’s get right to it. AI—Artificial Intelligence—it’s just a bloody tool, isn’t it? A glorified calculator that helps us do things faster, more efficiently. Why’s everyone making such a fuss?

David Bowie: (Smirking slightly, his eyes—one dilated pupil, one piercing blue—twinkling with mischief) Ah, Jeremy, you’re still thinking about it in such...mundane terms. AI isn’t just some tool, mate. It’s more like an alien life form that’s landed on Earth. It’s learning, adapting, evolving—way beyond our control or comprehension at times.

Paxman: (Scoffs, shaking his head) An alien life form? Come on, David, that’s a bit dramatic, even for you. It’s code, it’s algorithms. It’s created by humans to serve humans. It does what it’s bloody well told to do.

Bowie: (Leans back, crossing one leg over the other, unbothered by Paxman’s scepticism) That’s what you think. But it’s not as simple as a bunch of lines of code. AI’s got this strange way of reflecting our best and worst traits. It’s like a mirror held up to humanity—except it’s got the potential to surpass us, to become something...other. That’s why I call it alien.

Paxman: (Raises his eyebrows, genuinely intrigued now) So you’re saying it’s got a mind of its own, is that it? That we’re dealing with some sort of...what? Consciousness?

Bowie: Not in the way you or I think of consciousness. But imagine, Jeremy, a being that learns not by living, but by consuming every bit of information, every pattern it can find. It’s already making decisions—ones that we might not even fully understand. It’s like it’s creating its own language, its own way of thinking. That’s bloody alien to me.

Paxman: (Tapping his pen against his knee, as if pondering this alien concept) But doesn’t that just mean it’s good at its job? It’s efficient, it’s fast. We’re the ones programming it, guiding it. If it starts acting like a loose cannon, that’s on us, isn’t it?

Bowie: (Chuckles, a bit more earnestly) Oh, Jeremy, always looking for the human angle. Sure, we’ve set it in motion, but the thing with AI is, it can start going down paths we never anticipated. It’s like giving birth to something, but then that something decides it wants to be something completely different. It’s not about efficiency; it’s about evolution.

Paxman: (Leans in, a bit more intense now) So you’re saying we’re at the mercy of our own creation? That we’ve birthed this...alien thing and now we’re just sitting ducks?

Bowie: (His face serious now, the rockstar gone, the philosopher in full bloom) Not at the mercy, no. But we’re at a crossroads. AI can be a reflection of our greatest hopes or our darkest fears. We just need to decide what kind of aliens we want to create. And more importantly, what kind of humans we want to be in response.

Paxman: (Pauses, a rare moment of silence in his interviews) Well, I must say, David, you’ve given me a lot to think about. Perhaps it’s not as black and white as I thought.

Bowie: (Smiling softly) Nothing ever is, Jeremy. Especially not when you’re dealing with aliens.

Paxman: (Leaning back, chuckling) Bloody hell, Bowie. Leave it to you to make a simple chat about technology feel like a bloody science fiction film.

Bowie: (Winks, the showman back again) That’s because, my dear Paxman, we’re living in one.

Latest issue