AI is advancing at an extraordinary pace, and as knowledge workers, understanding its implications for our roles is crucial. Determining how much time and effort to invest in learning to harness AI in our daily tasks can significantly impact our productivity and job satisfaction.

AI isn't taking jobs so much as it's reshaping the tasks within them. Identifying which aspects of your job you can offload to AI will make you more effective and happier at work, at play, and in life.

What's Your Relationship to AI?

We've created two lenses through which to think about how much you may want to lean into investing in gaining AI skills.

First, we encourage you to reflect on your familiarity and comfort level with AI by considering where you fall on this spectrum: AI Avoidant, AI Cautious, AI Curious or AI Enthusiast.

If you identify more with the AI Avoidant or AI Cautious categories, we recommend a conservative approach to investing in AI skills, opting not to engage deeply until the technology becomes more reliable. However, keep in mind that this approach will come with costs and limitations that could affect your work experience and its relevancy as AI reshapes the tasks within jobs.

Conversely, if you are AI Curious or Enthusiast, we recommend you work to make AI a daily part of your routine -- even if it doesn't always give you consistent results (yet!).

You can download a copy of this slide here.

If you're more comfortable in the AI Avoidant or AI Cautious stages, then you probably want to take a "wait and see" approach to making investments in learning how to use AI in your work until it "just works," because today, you'll often have inconsistent experiences with AI -- sometimes it'll perform magically, and other times you'll hit dead ends. Importantly, though, is to recognize the potential costs and limitations of waiting as outlined above.

What are the machines capable of?

To understand the potential benefits of AI, consider its capabilities today and within 6-18 months across three key categories: Use Confidently, Trust by Verify, and Use Carefully.

You can download a copy of this slide here.

As advancements continue, even tasks that currently require cautious handling will become easier to for machines to manage in the near future. Leopold Aschenbrenner highlighted in his insightful Situational Awareness whitepaper that advancements in AI can be measured in "Orders of Magnitude" (OOM) of improvement, and we just need to "count the OOMs."

"The upshot is pretty simple. GPT-2 to GPT-4—from models that were impressive for sometimes managing to string together a few coherent sentences, to models that ace high-school exams—was not a one-time gain. We are racing through the OOMs extremely rapidly, and the numbers indicate we should expect another ~100,000x effective compute scaleup—resulting in another GPT-2-to-GPT-4-sized qualitative jump—over four years."

At Storytell, we foresee AI evolving from being an exciting novelty for AI Curious & Enthusiast early adopters to an integrated tool within every knowledge worker's routine -- even using Storytell to write code, as engineer Dan Luu shared in this blog post on the front page of HackerNews today.

If the work you do is in the "Use Confidently" bucket and you are AI Avoidant or AI Cautious, you're much more likely to experience the costs and limitations of those stances sooner than if your work is in the "Trust but Verify" or "Use Carefully" buckets.

To help you determine how much to lean into AI capabilities today, here's how we see our Storytell users using AI:

• At Work: Top use cases from our users across job functions‍
• At Play: How our users are showing us they use AI for fun
• In Life: Ways our users are using AI to improve their lives and personal relationships

Dan must be at least AI Curious to try to do something we'd put in the "Use Carefully" AI capabilities bucket. It was interesting that he found it didn't take too much longer to have the machine write the code from scratch than for him to do it himself. He said:

"I don't think it took too much longer to get working code than it would've taken if I just coded up the entire thing by hand with no assistance. I'm going to guess that it took about twice as long, but for all I know it was a comparable amount of time."

That's an incredible statement -- imagine trying to have GPT-2 write code from scratch just a few years ago. At that time, it was magical that GPT-2 could just string a coherent sentence together.

Our goal at Storytell is to make the right AI accessible to humans so they can do their best work within the enterprise. We're building an enterprise-grade platform that make AI safe for enterprises to use with their company data.

Some estimates show that over 80% of the data inside enterprises is unstructured, and Fortune 500 companies often have upwards of 1 billion private assets in this unstructured form (PDFs, Docs, PPTs, videos, audio, chat messages, emails, etc). With Storytell, all that internal knowledge (which we describe as living in "data swamps") transforms from dead pools of data to collaborative intelligence for the employees of that company. And being able to use AI to work with all that unstructured data makes them feel like superheroes.

Determine your right level of AI investment

We recommend reviewing the AI Capabilities framework to determine whether your tasks belong in the “Use Confidently,” “Trust but Verify,” or “Use with Caution” categories.

This assessment can help you inform your stance toward AI and guide your personal investments in its use as a collaborative partner in your work—because the OOM improvements will keep coming!