Reminder For Tomorrow’s Frontier Prompting Class

Tomorrow at 11am EST is the second Frontier prompting class. Paid members can join live classes and get the full recording. Free subscribers get a preview and summary of the recorded version. I will send a link to the Zoom call tomorrow.

Reader Note: I Wrote This Post With AI

This is an experimental impromptu format that I’m exploring.

The Problem: I’m finding that I don’t share most of the insights I have simply because it takes too long to convert them into posts.

My Solution: To rectify this, I used AI in a new way to create content quickly:

  • As I was having a conversation with my thought partner,

    , I had an insight.

  • Rather than just continuing the conversation, I started a new recording.

  • Then, I uploaded it to my video editor, Streamlabs, so I could clip the beginning and ending seconds and get a transcript.

  • Next, I uploaded the transcript to Claude to create a post using a modified post template I’ve used in the past.

  • Finally, I made edits to the post to get it finalized.

Result: I’m actually very happy with the final post, which I was able to create in 90 minutes. You’ll have to judge whether you think it’s good enough. I’m still on the fence about how to balance quantity vs. quality, and the engagement of this post will provide me with a new perspective.

Takeaway For You: Beyond the content of the post, hopefully my process inspires you to use AI in a new way and try your own experiment.

Paid Member Bonus: I share the prompt I used to turn the transcript into a post. I’ve spent 3+ hours refining it, and even though it’s still rough, it did dramatically speed things up for me. Paid subscribers also get access to:

  • The full two-hour recording of the first Frontier Prompting class

  • Access to tomorrow’s live class

  • $2,000+ of other perks (courses, premium AI prompts, paid podcast, 50+ curated video lessons)

Article

"The future is already here – it's just not evenly distributed."
—William Gibson

New research shows that many, if not most, of AI's abilities don't manifest out of thin air.

Instead, our perception of their sudden appearance stems from two key blindspots:

  1. We have trouble tracking early-stage capabilities. When a new capability is in its nascent stage, the AI's performance is so poor that we often fail to recognize and track its emergence. As a result, it seems to burst onto the scene from nowhere once it reaches proficiency. This is similar to the idea that “It takes 10 years to become an overnight success.”

  2. Our measurement tools are lacking. Our understanding of AI's capabilities is limited by our own methods of measurement. As physicist Werner Heisenberg put it, "What we observe is not nature itself, but nature exposed to our method of questioning."

Since delving into this research, I've become fascinated by the question of how we can uncover these hidden emergent abilities earlier. The earlier we track these nascent abilities:

  • The more that researchers can make AI safe.

  • The more that knowledge works can get a competitive edge.

Below is a summary of the big four insights from the video:

  1. Use ChatGPT + Elicit To Do Rapid Academic Research

  2. Understanding Artificial Super Intelligence (ASI) May Be The Most Important Thing We Can Understand

  3. By Understanding AI’s Emergent Properties, We Can Become an AI Oracle

  4. There’s A Big Opportunity To Be On The Frontier Of AI Thought Leadership

Together, these four insights give an early roadmap for a new way knowledge workers can get to the frontier of AI.

Insight #1: Use ChatGPT + Elicit To Do Rapid Academic Research

First, I used ChatGPT as a thinking partner to provide new perspectives on a topic I was exploring.

Next, I went deeper on one of the subtopics proposed by ChatGPT. To do so, I used a tool called Elicit, which helps explore and understand academic research.

By feeding it specific queries, I was able to rapidly:

  • Surface relevant studies

  • Summarize the findings from them

  • Extract insights from them

In a matter of minutes, I gained an overview of the research landscape that would have taken hours of manual searching. But the real magic happened when I asked Elicit to chat with the selected studies:

Identify emergent properties that exist today but are overlooked, and paint a picture of how each will likely evolve with future versions of AI.

The system surfaced insights about in-context learning, multi-step reasoning, the rapid pace of emergence as models scale, and more—weaving together threads from disparate studies into a coherent narrative.

Insight #2: Understanding ASI May Be The Most Important Thing We Can Understand

Perhaps the most tantalizing emergent property on the horizon is Artificial Superintelligence (ASI):

ASI refers to a type of AI that surpasses the intelligence of all humans in every field, including science, creativity, social skills, and more. Unlike today's AI, which excels in specific tasks, ASI would have general, self-improving capabilities, allowing it to learn and advance far beyond human control.

The implications of ASI are profound. It could solve huge global problems, like curing diseases and reversing climate change, but it could also lead to unpredictable outcomes because it would operate on a level beyond human understanding. People worry that if not carefully guided, ASI could pursue goals misaligned with humanity’s well-being. Therefore, managing its development is seen as one of the most important challenges of our time.

While it sounds like far-fetched science fiction, the seeds of ASI may well be embedded in today's advanced AI systems:

To truly understand the trajectory and implications of AI, focus not just on artificial general intelligence (AGI), but on artificial superintelligence (ASI)—even though its full realization seems like sci-fi. Grappling with ASI today is crucial for navigating the future.

Using AI-powered research tools, I was able to quickly map the terrain of ASI research and zero in on the most salient insights. For example, I learned that ASI will likely exhibit staggering emergent properties that arise from the sheer complexity of the systems rather than intentional design.

ASI will likely exhibit emergent properties, behaviors, and capabilities that arise not from design, but from the complexity of the system. This concept is critical because it makes ASI unpredictable and qualitatively different than AGI.

While the full manifestation of ASI remains speculative, identifying the traces of it in today's AI and exploring their potential future expression is crucial for constructing a meaningful map of where the technology might take us.

Insight #3: By Understanding AI’s Emergent Properties, We Can Become an AI Oracle

The ability to discern subtle patterns, nascent developments, and overlooked clues is becoming a superpower in a world reshaped by artificial intelligence. By learning to see the future in the present, we can prepare ourselves and others for the tectonic shifts on the horizon.

But this skill isn't just about honing our powers of perception. It's about learning to harness the tools at our disposal to augment our understanding:

To be an AI thought leader requires developing a deep understanding of the field's cutting-edge developments and paradigm shifts, not just polished writing - and that's never been more doable than it is today using AI-powered research tools.

The synthesized research and insights surfaced in this post are a testament to the incredible leverage AI affords us in making sense of complex domains. By combining a high-level map of the terrain with an on-the-ground investigation of specifics, we can weave disparate threads into a cohesive tapestry of meaning.

Insight #4: There’s A Big Opportunity To Be On The Frontier Of AI Thought Leadership

We're entering an era where the pace of knowledge production is accelerating exponentially, and the most impactful breakthroughs are not always immediately obvious. In this landscape, thought leaders have an opportunity - and perhaps an obligation - to light the way.

By developing a keen eye for the hidden emergence of transformative capabilities and using the tools at our disposal to bring clarity to complexity, we can help others navigate the tumultuous waters of change. In doing so, we not only demonstrate our own expertise, but we empower others to surf the waves of disruption alongside us.

The future may already be here, but it's up to us to distribute it.

Armed with a new lens for seeing the unseen and new tools for making sense of the ineffable, we can step into our roles as guides, illuminating the path through the mist.

The only question is: will you join the frontier?

Paid Subscriber Bonus: Prompt I Used To Create Article

To make use of the following prompt, I recommend using it as an instruction inside a Claude Project so you can reuse it.

App screen showing custom instructions
  • Customizing the text of the instruction to match who you are, your audience, and your goals.

  • Uploading one of your best articles to the project instructions so it can match your voice, style, tone, and structure.

Finally, once you save your project instructions:

  • Start a new conversation inside the project

  • Copy and paste your video transcript into the window

  • Use Claude Opus 3.0, which is a great writer.

  • Hit return.

As a reminder, I spent 3+ hours refining the instruction below, but it’s still rough so expect to do some editing.

Enjoy, and let me know how it works for you in the comments.

This post is for paid subscribers