I’ve been quiet for over a month.
Some of it was personal. Anyone who follows me knows this spring has been hard. I’m not going to make this post about that.
But part of the silence had a different reason.
I was reading.
Specifically, I was reading Max Tegmark’s Life 3.0: Being Human in the Age of Artificial Intelligence. And I couldn’t write while I was reading it — because every time I tried, I kept going back to the book.
Tegmark is an MIT physicist and co-founder of the Future of Life Institute — the organization that, in a meaningful way, put AI safety on the map as a serious research priority. The book is his attempt to answer what he calls the most important question his generation faces:
What kind of future do we want — before AI makes that choice for us?
I read that sentence and felt it land somewhere specific. Not philosophical. Professional.
Because I work in higher education. And higher education is not asking that question. Not seriously. Not yet.
He Gave Me a Taxonomy. I Couldn’t Unsee It.
Tegmark introduces three stages of life.
Life 1.0 is biological. Your hardware and your software — everything you are — is shaped by evolution. You get what you’re born with. You cannot update it within a lifetime.
Life 2.0 is cultural. Your biology is still fixed, but your software — your beliefs, your skills, your knowledge — can be redesigned within a single lifetime. That’s us. That’s been us for roughly 100,000 years.
Life 3.0 is technological. Both hardware and software can be redesigned. Life that can rewrite its own operating system. That’s what we’re building toward. Whether we’re ready or not.
I put the book down after that page.
Not because I needed a break. Because I needed to think about what I do for a living.
I work in higher education. I have for sixteen years. And when I held Tegmark’s framework up against the institution I know — the one most of us work inside — I saw something I couldn’t unsee.
We are a Life 2.0 institution trying to prepare people for a Life 3.0 world.
And we are running out of time to notice.
The University We Have Is Not the University We Need
Here is the honest version of the history.
University 1.0 made sense when knowledge was scarce. The institution existed to warehouse it and distribute it to those with access. The lecture was the delivery mechanism. The degree signaled that you absorbed what was stored. This was legitimate for centuries. It worked because knowledge itself was the scarce resource.
University 2.0 — where most of us work right now — exists to structure exposure to knowledge and certify completion of that exposure. The degree still signals something. But if we’re honest with each other, it increasingly signals persistence more than capability. We all know this. We just don’t say it at faculty senate meetings.
University 3.0 is what we need now. Not eventually. Not after the next strategic planning cycle. Now.
Because here is what has changed: knowledge is no longer scarce.
It is generatable. On demand. At scale. By machines.
The AAAI 2025 Presidential Panel documented what this actually means in practice: AI systems have achieved substantial capacity for multi-step reasoning, synthesis, and domain-specific analysis. The cognitive tasks we have historically used to certify that someone learned something — the essay, the exam, the research paper — are machine-replicable.
So what exactly are we certifying anymore?
That question has no comfortable answer inside the current model. And the longer we avoid it, the more irrelevant we become.
In a Life 3.0 world, a university cannot exist to transmit knowledge. It must exist to develop judgment — the capacity to evaluate, challenge, decide, and take responsibility when the information is never complete and the stakes are real.
That is a fundamentally different job. It requires a fundamentally different institution.
So I Designed One. I’m calling it Telos University.
Telos is the Greek word for ultimate purpose — the thing a thing is designed toward. In Aristotelian philosophy, it is the reason a thing exists.
A university whose entire design is organized around its intended outcome — judgment, not knowledge transfer — deserves a name that is its own argument.
Here is what it looks like.
No Majors. Six Problems.
Telos has no majors.
It has six Problem Domains — each one a persistent, complex, real-world challenge that cannot be solved by a single discipline, a single technology, or a single generation.
This is not an aesthetic choice. The AAAI 2025 panel found that 95% of the AI research community endorses multidisciplinary research as essential. Siloed expertise is not just limiting — it is a structural liability. EAAI’s 2026 proceedings documented that AI education works best when embedded in a real problem context, not delivered as abstract content.
So what do students at Telos actually work on?
- How AI systems interact with law, governance, civil infrastructure, and public trust
- AI in clinical decision-making, biomedical discovery, mental health, and health equity
- Climate modeling, resource allocation under uncertainty, and ecological systems
- Misinformation, AI factuality, knowledge production, and information infrastructure
- Workforce transition, adult reskilling, and human-AI team design
- Threat modeling, AI in security, and institutional resilience
Each domain runs as a living research-and-application unit. Real partners. Real problems. Real data. Students are not simulating anything. They are contributing to ongoing work with stakes outside the classroom.
That distinction matters more than it sounds. Simulation builds confidence in controlled environments. The world is not a controlled environment.
No Bachelor’s Degree. No GPA. No Excuses.
Tegmark asks directly: What career advice should we give today’s kids?
His answer centers on flexibility. On human skills that are hard to automate. On the capacity to keep learning throughout a lifetime, not just during a four-year window between eighteen and twenty-two.
That is not a credential architecture. It is a capability architecture.
Telos builds credentials around demonstrated capability — tiered, domain-specific, evidence-backed.
Foundation. You can work with AI systems in your domain. You understand their mechanisms, their failure modes, and their ethical dimensions. You can interrogate an AI output and explain your evaluation to someone who has no technical background. 60 Contribution Points. Two completed Studio projects. Roughly 14 to 24 months — by evidence, not by calendar.
Practitioner. You can lead human-AI collaborative work. You have navigated a genuinely contested problem — one where the right answer was not obvious, stakeholders disagreed, and your reasoning was tested under real pressure. 160 Contribution Points. Five Studio projects. One domain partner engagement.
Navigator. You have done Practitioner-level work and brought someone else through it. You have taught, supervised, and managed a team inside an AI-augmented environment. 280 Contribution Points. Documented mentorship. At least one published contribution to a domain problem. This is the entry point for faculty roles at Telos — and for senior leadership in partner organizations.
These credentials are not sequential in time. They are sequential by demonstrated evidence.
The portfolio moves you forward. Nothing else does.
You cannot buy your way to the next tier. You cannot wait for your way there. You have to show your thinking, including where it failed, and demonstrate that the failure taught you something.
How Learning Actually Happens: The Studio
No lecture halls.
The EAAI research is consistent across sixteen years of proceedings: the structures that actually produce capable graduates are project-based, iterative, and socially embedded. Passive content delivery does not produce judgment. It produces test performance. Those are not the same thing.
Studios are the core unit at Telos.
A team of 8 to 12 students. One Navigator. One or more domain partners. One real problem. Every student has a differentiated role based on their current competency level — Foundation students contribute to defined sub-tasks, Practitioner students lead workstreams, Navigator students manage the Studio itself and mentor those below them.
A Studio semester runs a deliberate arc:
Weeks 1–2: Problem Scoping. What is the actual question? What do we already know? What would a good answer look like — and how would we know if we found it?
Weeks 3–8: Deep Work. Research, analysis, prototyping, iteration. AI tools are used throughout — and documented throughout. Every tool decision gets recorded.
Weeks 9–11: Challenge Phase. This is the part most universities would never do. The Navigator and domain partner deliberately destabilize the work. New constraints appear. Contradictory data arrives. A key stakeholder changes their position. Students have to demonstrate adaptive judgment — not just competency in a stable environment, but the ability to think clearly when the ground shifts.
Week 12: Portfolio Submission. Not a final product. A documented record of the entire process — what was tried, what failed, what was learned, and what would be done differently.
The AAAI-26 keynote by Ashok Goel named exactly this structure as the emerging standard for AI-era workforce preparation: problem-anchored, iterative, team-based, with AI as a tool students must learn to supervise — not just operate.
The graduates most at risk in a Life 3.0 world are those who only learned to perform under stable, predictable conditions. Every exam. Every rubric. Every controlled assignment. A lifetime of training for an environment that no longer exists.
Telos builds instability into the curriculum on purpose.
Because that is what the world is actually like.
The Transcript Is Dead. Long Live the Living Portfolio.
Here is a question nobody in higher education wants to answer out loud:
What does a GPA actually tell an employer?
It tells them how consistently a student performed inside a controlled, predictable environment designed to measure performance in controlled, predictable environments.
That is the wrong signal. And we have known it for years.
At Telos, every student has a Living Portfolio — a dynamic, versioned, professionally accessible record of their learning history. Not a grade sheet. Not a resume. A documented cognitive history.
Every portfolio entry contains five things:
- The problem — what it was, why it mattered, who was affected
- The approach — what the student tried, including what failed and why
- The AI record — what tools were used, what outputs were produced, how they were evaluated, where they were accepted, where they were rejected, and the reasoning behind each decision
- The outcome — what was produced, what changed, what would be done differently
- The assessment — a Navigator evaluation tied to specific competency indicators, written in plain language, not a number
The AAAI 2025 panel argued that as AI produces increasingly polished outputs, evaluation must shift to process — what the thinking reveals, not what the product demonstrates. EAAI 2025 and 2026 both found that students who maintain explicit records of their AI use develop significantly stronger AI literacy than those who use tools without reflection.
An employer looking at a Telos portfolio sees not a 3.8 GPA but a documented account of how this person thinks when the answer is not obvious, when the data is incomplete, when a stakeholder pushes back, and when an earlier assumption turns out to be wrong.
That is what employers actually want.
They have just never been given it before.
Contribution Points: Because Time Is Not a Credential
Credit hours measure time spent in a room.
Telos does not credential time.
Contribution Points accumulate by evidence of capability:
| Activity | CPs | Note |
|---|---|---|
| Studio project (full cycle) | 12–20 | Range reflects complexity and leadership demonstrated |
| Theoretical Foundation Module | 2–5 | Must be followed by applied demonstration — theory alone credentials nothing |
| Domain Partner Engagement | 3–8 | External partner signs off on demonstrated capability |
| Portfolio peer review | 1–3 | Assessed for quality, not just completion |
| Mentorship hours (Navigator-tier only) | 2–5 | Documented reflection required |
| Published contribution to domain problem | 5–15 | External review required |
One rule governs all of it: theory alone credentials nothing.
The Theoretical Foundation Modules — covering AI reasoning, factuality, ethics, agents, cognitive science, and the seventeen AAAI panel topics — only generate CPs when the student demonstrates applied understanding in a Studio context.
You cannot game this system by accumulating low-stakes activity. The portfolio moves you forward. Nothing else does.
Telos Does Not Hire Professors. It Hires Navigators.
The AAAI 2025 panel named this problem directly: academia is losing its best AI talent to corporate environments. The faculty who remain operate under incentive structures — publish or perish, committee service, grant cycles — that have almost nothing to do with what students actually need.
Navigators must have navigated a complex, high-stakes, AI-involved problem in the real world. Not just studied one. Written about one. Gotten tenure for analyzing one.
Actually navigated one.
They maintain their own Living Portfolio as a condition of employment. A Navigator who stops building their portfolio loses their role. Because a teacher who has stopped learning has no business teaching people to learn.
The faculty pool at Telos includes clinicians who have implemented AI diagnostic tools and know exactly where they fail. Engineers from autonomous systems teams who have managed safety-critical deployments. Public health researchers who used AI modeling in real outbreak response. Security practitioners who have designed threat frameworks in AI-augmented environments. Policy analysts who have evaluated actual AI governance proposals.
Academic researchers are welcome — but only those with current domain engagement, not just publication records.
Navigators work with a maximum of three Studios at a time. They are evaluated by student portfolio outcomes, domain partner feedback, and peer Navigator review. Compensation is competitive with mid-level corporate roles.
This is non-negotiable. Because we cannot build a Life 3.0 institution by paying Life 1.0 salaries.
No Provost. No Dean. No Department Chairs.
Telos has three operating structures.
Domain Councils — one per domain, composed of three Navigators, two rotating students, two rotating domain partners, and one Alumni Practitioner. They make all decisions about curriculum, problem selection, partner relationships, and Navigator hiring within their domain. No central approval required. No committee process. No waiting.
The Portfolio Authority — five rotating Navigators responsible for the integrity of the credentialing system. They set CP standards, conduct annual audits, and handle credential appeals.
The Operations Core — twelve people for a 600-student institution. Facilities, technology, legal compliance, financial operations, and student support. They do not make academic decisions. They do not vote on curriculum. Their job is to keep the institution functional so that Navigators can focus on students.
The institution is governed by the people doing the work.
Not by people managing the people doing the work.
Every administrator who has ever sat in a meeting about a meeting will understand exactly what this is designed to fix.
What It Costs — And Why the Model Is Fair
Tegmark is direct about AI’s effect on economic inequality. In a Life 3.0 world, the gap between those who can navigate AI systems and those who cannot will not close on its own. It will widen — unless institutions are deliberately designed to work against it.
Telos uses an Income-Contingent Contribution model.
During enrollment, $150 per week covers technology, portfolio infrastructure, and Studio materials. That is $7,800 per year. Aid is available for students who cannot manage it.
After credentialing, once employed: 4% of annual income above $40,000, for a maximum of 7 years, capped at 1.5x the full cost of the credential tier.
| Credential | Full Cost | Maximum Contribution | Monthly at $65K |
|---|---|---|---|
| Foundation | $35,000 | $52,500 | $83 |
| Practitioner | $80,000 | $120,000 | $83 |
| Navigator | $130,000 | $195,000 | $83 |
If you never earn above $40,000, you contribute nothing beyond the participation fee.
No interest. No default risk. No debt collection. No phone calls from a collections agency ten years after you graduated from a program that didn’t get you the job it promised.
Telos is funded by graduate income-contingent contributions, domain partner contracts — organizations pay to have their real problems worked on by Studio teams — targeted philanthropic grants, and government-funded research. This model requires patient capital for the first 8 to 10 years before the graduate contribution pipeline matures.
That is a financing challenge. It is not a reason to abandon the model.
Three Questions You’re Already Asking
What about accreditation?
Current frameworks are built for credit hours. They do not recognize Telos’s structure. This is real and it is not small. The path runs through regional accreditors who have precedent for alternative frameworks — SACSCOC and HLC have both worked with competency-based models — and through federal pathways the Department of Education has used. It takes five to seven years. It is not insurmountable. And it is absolutely not a reason to design the institution around compliance requirements instead of learning outcomes.
We have been doing that for too long already.
Who is the first student?
Someone who cannot afford a traditional university, has already demonstrated the capacity to navigate a real problem without institutional scaffolding, and needs a credential structure that actually reflects what they can do. This student exists in enormous numbers. Tegmark would call them the people most at risk from a Life 3.0 transition managed without them in the room.
Telos is built for this student first. Not as an afterthought. First.
What if the model becomes obsolete?
Then it adapts. Domain Councils can revise domains and CP standards without a faculty senate process. The Living Portfolio evolves as the environment evolves. The institution is not designed to last forever unchanged.
It is designed to remain responsive.
That is, in fact, the entire point.
What the Book Actually Changed
Here is what Life 3.0 did that nothing else had done quite as cleanly: it gave me the right level of abstraction.
I had been thinking about AI and higher education at the level of tools. Chatbots in the classroom. AI policy statements. Plagiarism detection. Generative AI in writing assignments. These are real problems.
They are also the wrong level.
Tegmark forced the question higher. Not how do we manage AI in the classroom — but what kind of human being does a Life 3.0 world require, and are we building institutions capable of developing that person?
The AAAI research gave me the specificity to answer it.
Together, they gave me the design brief for this institution.
Telos doesn’t exist yet.
But the question Tegmark asks — what kind of future do we want, before AI makes that choice for us — applies to higher education as directly as it applies to AI governance, labor markets, or geopolitics.
We can design University 3.0 deliberately.
Or we can watch it emerge by default from the wreckage of University 2.0.
I know which one I’m working toward.
What would you build differently? I want to hear it.
Sources: Max Tegmark, Life 3.0 (2017). AAAI 2025 Presidential Panel Report. EAAI 2025 and 2026 proceedings. AAAI-26 keynote, Ashok Goel. AAAI 2024 Diversity in AI Education symposium.
Leave a Reply