I Think We Should All Get Superpowers
Generative AI has consumed me. And I’m not complaining. I’ve never been so excited about the potential of any other technology.
I don’t say this only as a technologist. I say this as someone with over 20 years of experience in multiple industries, in roles as a designer, developer, strategist, entrepreneur, and leader. I’ve seen some shit. But nothing like this.
And yet, our AI-powered future is not clear. It could be bad. It could be great. It’ll probably be somewhere in between. No matter what, I want to have a hand in shaping it. I want all of us to.
We’re Not Ready
It was at the end of 2022 when I first encountered ChatGPT.
I had been tangentially aware of OpenAI—specifically their impressive achievements with AI in Dota 2 from a few years prior. But I hadn’t been following anything related to large language models (LLMs). I didn’t know about the transformer architecture or that OpenAI had been doing research in this area.
Within an hour of using ChatGPT, I sat back, stunned. This wasn’t just another tech novelty. This was the beginning of a paradigm shift.
I didn’t waste time on parlor tricks like getting it to tell me jokes or write silly poems. Instead, I challenged it with real tasks—strategy work, brainstorming, writing code, the kind of thinking and tasks that are core to what I do. And it delivered (or at least I saw the bright spark of potential). That’s when it hit me: everything was about to change.
And to this day, I still don’t think folks are ready for it.
Where Are We Headed?
Fast forward to earlier this year. I saw two distinct paths emerging for how we (humanity) pursue generative AI and its role in our lives:
- Replacement: We lean heavily into using AI to automate all the things humans do.
- Augmentation: We focus on using AI to enhance all the things humans do.
The direction we choose will set the tone for how we move forward. They’re not mutually exclusive, but I’m worried about an imbalance.
I’m seeing a distinct push towards the first path, driven by economic incentives and the allure of short-term gains. I’m seeing an active pursuit of the displacement of humans without offsetting it with real enhancement of human capabilities. The key there is "offsetting." Some amount of automation of jobs will happen. That’s a given. But what do those displaced workers do? Even if only some of their time is displaced.
I’m hearing plenty of talk from execs and senior leaders at companies about freeing up people for "high-value work" by automating the mundane, yet I’ve never heard them specify what that high-value work looks like. Nor how companies plan to fill those saved hours. Where’s the upskilling strategy? Where’s the pro-experimentation mindset modeled by the leaders at companies? And the investment in teaching people how to actually use this ever-advancing AI?
There’s an opportunity here for enabling people to fill those displaced hours with higher-value work if there’s a plan in place that answers those questions.
The Disconnect
In my previous role and through client interactions, I’ve seen firsthand the fear of job displacement among knowledge workers, my peers included. And rightfully so if the automate-all-the-things path is the focus.
At the same time, I’ve watched companies salivate over the potential for "efficiencies"—often a euphemism for cutting labor costs.
But here’s the thing: these large language models, they think differently from us. They excel at tasks we struggle with, and they stumble over things we find trivial (i.e. the "jagged frontier"). When you combine their strengths with ours, that’s when the magic happens. That’s when we become capable of almost anything. That's when we get superpowers.
So this workforce that’s seemingly negatively impacting your company's bottom line? They can become your most valuable resource when you focus on augmenting them with AI.
We Can Do Better
I’m talking a lot about the workplace here. It’s important, and it’s one that’s top of mind for so many of us. But this idea of human-AI collaboration (whether centaur or cyborg) isn’t just about work. It’s about all the things humans do.
I fundamentally believe that we’re just scratching the surface of what’s possible with human-AI collaboration. That it has the potential to transform every aspect of our lives.
How we think.
How we create.
How we learn.
How we play.
How we interact with each other and the world around us.
But what could this actually look like?
Imagine an experience that doesn’t just respond to your queries but actively enhances your ability to generate diverse and novel ideas (check out my AI brainstorming prototype). Or maybe an app that gets your kids engaged in reading because they’ve had a hand in crafting their own stories (check out my prototype for helping with reading practice). Or a system that can create long-form, personalized articles about any topic and make this information accessible and engaging (pair my knowledge generator with NotebookLM’s Audio Overview feature).
Here’s the thing: this is all possible with today's LLMs. But it doesn't seem like enough people are mining these capabilities and creating new ways of making them accessible and useful to everyday people.
Instead, we’re bottlenecked by limited interfaces—mostly chatbots—that fail to integrate seamlessly into our workflows, creative processes, and daily lives. They don’t meet us where we are, so we’re left to figure out how to make them work for our own needs. And today, that requires a lot of "prompt engineering" and settling for interfaces that are really only intended for dialogue.
In other words, this is a UX problem. Not a technology problem.
Claude’s Artifacts are starting to go further (an augmented chat with a dedicated space for iterating on an output). ChatGPT’s Canvas is getting somewhere too (you can actually take actions on the output). These features, however, feel like bolt-ons to a chat interface. But you can start to see the evolution and recognition that we need more than chat interfaces.
And there are folks doing just that. Cursor is pushing on the frontier of integrating LLMs into your code editor. LABS.GOOGLE is cooking up lots of fun stuff. Udara Jay is making wonderful things like Pile and Research. Amelia Wattenberger is constantly exploring the intersection of novel UX and LLMs.
All I can say is that I don’t ever want to see another copilot sidebar or a chat-with-my-PDF service. It’s lazy. It’s not what we need. No more chatbots as our primary interface for interacting with LLMs. I’m done with that.
Making It Real
I can talk about all of this potential endlessly. But I don’t want to just talk about it. I want to show it. So I’m focusing on creating prototypes and experiments that demonstrate the potential of human-AI collaboration. I want to take part in proving that we can use this technology to give people superpowers, to help them do things they never thought possible. I know I won’t be the only one doing it. I just want to be one of the ones doing it.
In my own experience with numerous LLMs over the last two years, I can confidently say that I'm smarter, more creative, more capable, and more confident in what I can achieve because of my collaboration with them. But it feels like it's been a brute-force effort. The conversational form factor is slow and clunky. And I've needed to become an LLM whisperer to elicit the behavior I want.
Lots of knowledge about their capabilities was gained though. And this is what I hope to bring into my experiments. Which will, in turn, hopefully inspire others to think differently about the value that generative AI can bring and what its role in our lives can be.
Co-creating Tomorrow
An AI-infused future is coming, but its shape is still malleable. We don’t need to passively accept that AI will replace us. We can have a big say in what this future looks like by showing what’s actually possible. For some, this will mean lots of creative experimentation with things like Claude or ChatGPT or Gemini. For others, it'll mean building new tools and experiences that people can benefit from.
We can all take part in this.
I’m all in on human-AI collaboration. I want to see enhancement and augmentation of human capabilities. I want to see people feel empowered to do things they never thought possible. I hope you do too.
What’s your vision for a future where AI enhances rather than replaces human capabilities? How do you see yourself collaborating with AI in your work or personal life? What do you wish you had but doesn’t exist yet? I find that when I'm exposed to a totally unfamiliar domain, my brain starts bubbling with ideas for how generative AI could have a positive impact. So if you've got a unique problem or perspective, I'd love to hear from you.
