Bitter and Better Lessons Taken Together

One of the most interesting things about artificial intelligence is how quickly it can challenge our assumptions.

I recently sent my friend Todd McLees a video embedded in a blog post by Ethan Mollick, an AI researcher at the University of Pennsylvania whose work I follow regularly. The post, The Dot and the Swarm, explored the growing power of AI agents and featured a music video generated from a well-known 2019 essay by AI researcher Richard Sutton called The Bitter Lesson.

The video was fascinating. It was also a little unsettling. The core message seemed clear enough: throughout the history of AI, systems that leverage massive amounts of computation and data tend to outperform approaches that rely primarily on human-designed rules and expertise. As AI capabilities continue to improve, it is easy to conclude that human knowledge and judgment may become increasingly irrelevant. At least that’s how it initially struck me.

When I shared the video with Todd, he quickly responded with an important observation. Sutton’s original essay was written in 2019 for AI researchers and system designers, before any of us had access to large language models like ChatGPT, Claude, Gemini, Copilot, or Grok. It was a lesson about how to build better AI. The new video updated the message for 2026, but in doing so, it may also have broadened the audience in ways that made the lesson easier to misinterpret.

Thirty-six hours later, Todd sent me something unexpected. He had created his own response, called The Better Lesson, on his Substack. Rather than focusing on the growing power of AI systems, Todd’s video focuses on the growing importance of human direction and how to use AI more effectively. If AI increasingly provides access to vast amounts of knowledge, analysis, and content creation, then perhaps the most important skill is no longer possessing information. Perhaps it is learning to ask better questions, exercise sound judgment, establish meaningful goals, and direct intelligence toward outcomes that matter.

That distinction struck me as profoundly important for higher education. For generations, colleges have helped students gain access to knowledge. Increasingly, knowledge itself is becoming abundant. The challenge for educators may be less about delivering information and more about helping students develop discernment, judgment, curiosity, ethics, creativity, and purpose. These are competencies that formal education often develops indirectly, in ways that can be difficult to define and measure. The magnitude of such a shift is daunting to consider.

In short, the future may not belong to those who compete with AI. It may belong to those who learn to direct it effectively. In the face of what felt to me like an overwhelming transformation, that idea was empowering and reassuring. At least, that’s what I took away from Todd’s response.

Not the bitter lesson.

The better one.

If you’d like to explore both perspectives, I encourage you to start by reading Ethan Mollick’s The Dot and the Swarm and then watch Todd McLees’ thoughtful response, The Better Lesson. The contrast between them is where much of the learning resides.

If you’d rather take just eight minutes to watch the videos back-to-back, here are the links. I encourage you to watch Ethan’s Bitter Lesson first, then follow it with Todd’s Better Lesson. I’ll be curious to know whether you watch them more than once to let their messages sink in.

If you have any questions or comments on this post, please contact me directly at presblog@mvcc.edu

Looking Back to See Forward

One of the most difficult challenges in higher education today is making sense of the moment we’re in to guide our actions. Artificial intelligence is reshaping knowledge work. Demographic shifts are changing who attends college. Questions about affordability, value, and the future of work are challenging assumptions that have guided higher education for decades. The pace of change can feel overwhelming, making it difficult to distinguish temporary disruption from genuine transformation. But I think most would agree: this feels different.

Years ago, futurist Marina Gorbis of the Institute for the Future introduced me to the concept of “looking back to see forward” in her Thinking Like a Futurist article in Educause Review. The premise is simple: understanding historical patterns can help us better anticipate what comes next. That idea recently led me back to Neil Howe’s work on generational cycles and The Fourth Turning. Whether one embraces the framework fully or not, it provides an interesting lens for examining major periods of disruption and renewal in American society.

As I reflected on the history of higher education through that lens, two moments stood out. The first was the Morrill Act of 1862, which created the land-grant university system and dramatically expanded educational opportunity beyond the nation’s elite institutions. The second was the GI Bill following World War II. Combined with the rapid growth of community colleges, it democratized access to postsecondary education for millions of Americans and helped build the modern middle class.These transformational moments occurred 84 years apart: 1862 and 1946. Add another 84 years and you arrive at 2030. Coincidence? Perhaps. But it raises an intriguing question: If previous crises produced bold new models of educational access to align with a societal reset, what might emerge from the current one?

The conversation around artificial intelligence often focuses on how students learn, how faculty teach, or how institutions operate. Those are important questions, but they may not be the most important ones. What if AI becomes the catalyst that fundamentally reshapes access to education itself?

Zooming out to consider where we are on the timeline of history made me wonder about larger cycles beyond the four turnings, perhaps even 500-year patterns. It turns out there may be something there. Throughout history, transformative periods have often involved shifts in who holds the keys to knowledge. With each transformation, the circle widened, and new possibilities emerged.

Today, AI places unprecedented amounts of information, expertise, and personalized guidance into the hands of anyone with an internet connection. The implications are profound. If the last century was about democratizing access to knowledge through college, could the next era be about democratizing access to learning itself?

I don’t pretend to know the answer. But as we move toward 2030, this feels like more than another cycle of technological change. It feels like a moment. While the stock market may see AI as the equivalent of the internet of the late 1990s, on the ground it feels more like the printing press, if not electricity — or, who knows, perhaps even fire itself, as Todd McLees of humanskills.ai has suggested.

I recently shared my thoughts with Todd because I knew he was pondering the same connection between 1862 and 1946 and the democratization of knowledge through public universities and community colleges. He took the information I shared on who has held the keys to knowledge over time, deepened the content, and extended its implications by creating a fascinating video — literally the day after I messaged him. The video offers one perspective on this long arc of history and how the keys to knowledge have shifted during periods of profound societal transformation.

As you watch, consider a simple question: What is the role of community colleges in this moment?

If you have any questions or comments on this post, please contact me directly at presblog@mvcc.edu