B23 digest: 2025-03-30 to 2025-04-01 (2 posts)#

What this batch is#

There are two posts, both about his keynote “What does the future look like?” at the 2025 Yidan Prize Conference. ASU’s Mary Lou Fulton College co-hosted the conference.

• “Reimagining learning and education in an age of AI” (2025-03-30, reimagining-education-in-an-age-of-ai) writes up the keynote. It is almost entirely his own prose. The exceptions are two slide quotations (Dario Amodei and Mark Daley), prompt links to Perplexity and ChatGPT summaries (with no AI text in the post), and slide images whose text is not in the corpus. Relevance: high.
• “When AI Takes the Wheel” (2025-04-01, when-ai-takes-the-wheel) announces a Modem Futura podcast episode with Sean Leahy about the same keynote and a “post-scarcity” future. Its episode summaries were generated by ChatGPT o1-Pro. It is skipped per the user’s instruction not to use Modem Futura podcasts.

The batch therefore rests on one post: a clear statement of how he framed AI and education in early 2025.

Main ideas#

1. A named framework: transformative technologies change “what we do, and who we are”. He calls this “a conceptual model that I find myself using increasingly frequently”. Every technology moves us along both axes. AI may push us “far further” along them than earlier transitions. In a footnote he adds a third axis, “where we live”, and places AI “in a category of its own” there as well. Amodei’s “compressed 21st century” stands for the first axis (acceleration of what we can do). Daley’s claim that thinking machines cut “to the heart of who we are” stands for the second (identity).

2. AI as possibly a new category of technology, argued under uncertainty. He sets out three trajectories: a hype bubble and “yet another AI “winter.””; an S-curve ceiling, which he thinks “even more likely”; and exponential growth, “more speculative” but with “some credence”. He then argues that the mere possibility of transformative growth “demands new ways of thinking”. If AI turns out to be the general accelerant and identity-challenger that he, Amodei and Daley think it could be, it belongs “in a fundamentally different category to every previous technology that we’ve developed as a species.” His reasoning has a risk-analytic shape: take a low-probability, high-impact trajectory seriously enough to act now, whatever the modal forecast. The claim is conditional.

3. A value-creation model of why education matters. He asks educators why learning and education matter. His answer is that humans have an evolved ability to imagine better futures and “problem solve” their way to them, creating “value”. He defines value broadly: health and wealth, but also dignity, equity, justice, kindness, care for the planet, and “awe and wonder”. Learning enhances this ability, and education, as the formalisation of learning, amplifies it. So the provocation is this: if AI “can problem-solve faster and better than any one person, or any collective of people”, where does education’s value lie? He calls the prospect an “existential crisis” for education. He means a crisis of education’s purpose, not existential risk to humanity.

4. From defending human distinctiveness to learning to be human with AI. The first of his three provocations asks what it means to be human when AI can “emulate much of what makes us “us.””. He hedges that this may be “blind emulation with nothing more than unthinking digital processes behind it”. The third reverses the direction. Rather than asking how AI threatens who we are, he asks “what happens when we learn how to work with AI to make us more than we are”, and how we “learn how to be human in an age of AI”. Education’s task becomes centring the human while designing learning that uses “both human and artificial intelligence”. A footnote goes further. It calls human learning “a profound evolutionary self-improvement algorithm” and asks whether machines might also learn to problem-solve towards “futures they too can imagine”.

5. “Intelligence is free” and the scarcity model. He takes up Daley’s claim with caveats: “controversial, and not entirely accurate” given the material costs of AI. He still argues that on-demand AI may soon be available to anyone with a phone, “effectively making it free to them”, and may surpass human educators. This would overturn a centuries-old scarcity model in which “elite establishments” hoard intelligence for those who can pay. He presents this as democratising, but calls it “deeply contentious”. He does not raise here the question of who controls that “free” intelligence.

6. Steer, don’t stop; humility; the risk of standing still. He closes with three perspectives: “We cannot stop the march of innovation, but we can guide and steer it”, because “Technology is not deterministic”; flourishing requires “the humility to question assumptions and embrace change”; and the possibilities are “boundless” if we stay “tethered to what makes us human”. The final risk he names is the risk of inertia: “sacrificing what could be on the altar of a blind devotion to what is.”

Concepts appearing#

• What we do / who we are / where we live
• The value-creation model of learning and education
• The “compressed 21st century” (Amodei)
• “Intelligence is free” versus the scarcity model of intelligence (Daley)
• Three AI trajectories: winter, S-curve, exponential
• AI as a general accelerant across technology domains
• Emulation or simulation of human attributes, and “blind emulation”
• Learning to be human in an age of AI
• An “existential crisis” in education
• Human learning as an evolutionary self-improvement algorithm
• Technological non-determinism combined with the inevitability of innovation (“guide and steer”)
• Humility, and change as the natural state
• Meaningful value versus “wealth-creation machines”

What is new or changed#

• The what-we-do / who-we-are model is named as a framework he now uses routinely. Earlier batch notes do not record it in this form. The “who we are” axis links his long interest in being human to his analysis of technology transitions.
• A stronger claim about AI’s categorical novelty. Earlier (for example the 2024 technology-transitions work) AI was one of several boundary-extending technologies. Here, conditionally, it is in a category apart from “every previous technology”.
• An explicit trajectory judgement. He leans towards an S-curve ceiling while arguing that tail scenarios still demand preparation. This is a measured position between hype and dismissal.
• Continuity. “Guide and steer” restates his 2024 assumption that innovation cannot be switched off but can be channelled. Humility and embracing change echo the “Embrace” quadrant of his technology-transitions model. His focus on education and the university’s purpose continues his higher-education writing. The admission that he is “week by week … less sure what the future holds” signals growing uncertainty, not a new position.
• What is absent: there is no hazard-based AI risk and no governance or regulation. Nor does he discuss manipulation, cognitive offloading or the power of AI companies. Amodei is quoted respectfully and without critique.

Most important posts in this batch#

1. 2025-03-30 reimagining-education-in-an-age-of-ai. This is the only substantive post. It is essential for his what-we-do / who-we-are framework, his value-creation account of education, his hedged claim that AI is categorically new, and his stance of steering rather than stopping.