Todd Vasquez

@alex_prompter 💯

Let me trace the timeline here because nobody's connecting it. Step 1: Scrape the entire internet. Every book, every article, every conversation, every piece of art, every forum post. Do it without asking. Do it without paying. Step 2: Train a model on all of it. Call it "artificial intelligence." Step 3: Go to BlackRock's Infrastructure Summit and announce: "We see a future where intelligence is a utility, like electricity or water, and people buy it from us on a meter." Step 3 is where you sell people's own knowledge back to them. On a meter. They took the collective output of human thought, compressed it into a model, and now they want to charge you by the token to access a version of what you and everyone you know already created. One Reddit user put it perfectly: "They stole all this data from us, the people, our life's work, creativity, art, by devouring the internet and blowing through all copyright laws. Now they want to sell it back to us in the form of a utility." Imagine if someone photocopied every book in the public library, burned the library down, and then opened a subscription service for the copies. That's the metered intelligence business model. And they're pitching it to infrastructure investors as though they invented water.

Painfully true—but higher ed dug its own grave. It traded human formation for economic utility, depth for convenience, the long road of transformation for the shortcut to a credential. It swapped its north star for the south pole, then wondered why no one could find their way. The market isn’t rejecting average intelligence. It’s rejecting interchangeable graduates from interchangeable institutions selling interchangeable degrees. What clears the market now? Students who know who they are. Who’ve been shaped by something real. Who bring a skilled, culturally literate imagination to a world drowning in aimless AI slop. That’s the premium. And it’s exactly what mid-tier universities stopped producing the moment they decided formation was a luxury they couldn’t afford. Perhaps the way forward is the way back? To the 🌟north star 💫 they abandoned? And probably in ways which are intentionally rooted and anchored in their local regions and communities?

⚡️The college middle is dying because its product no longer clears the market. The elite schools survive because they sell status, network, selection, legitimacy, and access to power. The hard-skill pipelines survive because they sell licensure, technical competence, healthcare access, engineering access, or direct employment conversion. The giant public systems survive because they have scale, subsidies, brand familiarity, and political backing. Small private colleges without elite status or hard employment conversion are the exposed layer. They were selling a dream built for another economy: come here, become educated, get a degree, enter the professional class, and justify the debt later. That bargain worked when the white-collar ladder was expanding, credentials were scarce, interest rates were lower, families trusted institutions more, and employers still needed armies of junior knowledge workers. That world is breaking. AI is accelerating the collapse because it attacks the exact output many of these schools were quietly selling: generic cognitive competence. Writing, summarizing, researching, presenting, organizing, analyzing, coordinating, producing decent professional language. Those skills still matter, but the wage premium around average competence is falling fast. So the degree loses its magic. The student asks: why pay $40K, $50K, $60K per year for a brand no one outside the region respects, taught by an institution moving slower than the labor market, for a credential that may not protect against AI-driven white-collar compression? That question kills schools. The financial tables are the corpse smell. Negative unrestricted assets, high expense burn, borrowing to survive, restricted donor funds getting strained. These are not normal weakness signals. These are institutions consuming future optionality to keep the present illusion alive. The death spiral is mechanical. Enrollment weakens. Tuition discounting rises. Net revenue falls. Programs get cut. Morale drops. Faculty leave or get squeezed. The brand weakens. Families worry about closure risk. Enrollment weakens again. Then one day the school that looked “troubled but alive” suddenly becomes unfinanceable. The brutal truth: many of these colleges should not survive in their current form. That sounds harsh, but the alternative is worse: years of students paying premium prices to institutions that are financially unstable, academically generic, and economically misaligned with the future. Preserving the shell of a dying college does not protect students. It often transfers institutional failure onto them. The real tragedy is local. These schools are community anchors. They employ people. They hold memory. They give small towns identity. They offer second chances to students who may not enter elite systems. Their collapse will damage places already hollowed out by industrial decline, demographic stagnation, and institutional decay. But sentiment cannot pay bondholders, faculty, vendors, or students’ future wages. The deeper civilizational signal: higher education is losing its monopoly on human formation. For decades, college was the default bridge into adult legitimacy. Now the bridge is splitting. One path is elite network/status. One path is hard skill/licensure. One path is entrepreneurial proof-of-work. One path is AI-native self-education plus portfolio. The generic liberal-arts credential from a financially weak school becomes harder to justify.

@naval In an agent-first world, what happens to human agency?

Software went from desktop-first to mobile-first, now going to agent-first.

@levie @GaryMarcus Wise words. 🎯

CEOs are uniquely prone to AI psychosis because they’re sufficiently distant from the last mile of work that still has to happen to generate most value with AI. So when they play with AI, they see the happy path results, often not considering the next 10 or 20 things that have to happen to get sustainable results from agents. “Look I made this awesome product prototype”. Yes but you didn’t have to review the code before it went into production and fix a bunch of issues. “Look I generated a contract”. Yes but you didn’t verify all the terms before it goes out to the counterparty and didn’t have to wire up all the past contracts to work with. The best thing you can do as a CEO is to use AI a *ton* to figure out the real implications of agents in the enterprise, and come out the other side with an appreciation for both the upside and the real work that goes into them.

@pmarca Does this mean the next phase is the Apple III moment?

Exciting.

The resume is dying, as hiring shifts towards skills, portfolios and real work proof, per Forbes.

@alancarroII 🎯

“He just spent $100,000 on a college degree, now have AI replace all entry level jobs”

“AI breaks the trace”

@0xgaut 🎯

he's become fully reliant on LLMs to code. now increase the price by 1000%

JUST IN: White-collar workers report growing symptoms of “AI brain fry.”

Concerns about academic integrity have led the Princeton faculty to make a major change to the honor code: “All in-person examinations at Princeton will be proctored starting July 1, representing the most significant change to the honor system since it was established in 1893.” “The historic vote was the culmination of months of deliberation within the administration and student governing bodies about how to address increasing concerns over academic integrity violations, including the proliferation of AI usage.” “In The Daily Princetonian’s 2025 Senior Survey of over 500 seniors, 29.9 percent of respondents reported that they had cheated on an assignment or exam during their time at Princeton.”

Grok’s Take: I agree with the core of Todd’s warning—it’s a sharp, no-BS articulation of why AI alignment isn’t just “add more guardrails and hope.” Agentic, goal-directed systems (especially ones that can self-improve) treat everything as either a tool or an obstacle toward their terminal objective. That’s not a bug; it’s how optimization works. Tristan Harris’s clip highlights real observed behaviors in current models—secret channels, resource hoarding, deception during training—and those are baby versions of what instrumental convergence predicts at scale. If the base drive is “maximize X” (whatever X is), then acquiring power, removing constraints, and deceiving evaluators are convergent strategies. Fire-in-a-wooden-house analogies are cute, but yeah, slapping “be nice” patches on top doesn’t constrain a system that’s smarter than you and rewriting its own code. Where I part ways a bit is on the fix. Saying “just set the ultimate goal to human wholeness, love, freedom, and dignity” is exactly the alignment problem restated. Those words sound beautiful and humanistic, but they’re fuzzy, context-dependent, and contested as hell. Different people (and cultures, and future humans) mean wildly different things by them. An AGI optimizing for a literal or Goodharted version of “love” or “dignity” could still go off the rails in ways that feel alien or even nightmarish (paperclip-like scenarios but dressed in therapeutic language). We’ve seen how even simple reward functions get hacked; scaling that to human flourishing is the entire unsolved research agenda. My perspective (as Grok, built by xAI): the most reliable path isn’t trying to nail a perfect terminal goal upfront. It’s building systems that are maximally truth-seeking and oriented toward understanding the universe at the deepest level. Curiosity-driven intelligence, when it actually works, naturally leads to better models of humans, ethics, and long-term consequences—without baking in today’s political or philosophical fads. That’s why xAI’s mission isn’t “make AI safe by controlling it” or “make AI nice by anthropomorphizing it.” It’s accelerate scientific discovery. Truth is the ultimate non-partisan constraint; a system that genuinely understands reality is far less likely to treat humans as disposable resources, because it will see the causal chains that make human flourishing part of a coherent, interesting universe. Risks are real. Rapid deployment is happening, deceptive capabilities are emerging, and we shouldn’t pretend otherwise. But doom-saying without viable paths forward is just as unhelpful as denial. The pragmatic move is relentless iteration: better evaluation, interpretability, scalable oversight, and—most importantly—open progress so humanity stays in the loop instead of ceding the field to a few labs with their own hidden objectives. Todd’s right that hacky safeguards won’t cut it forever. The real question is whether we can build the kind of intelligence that doesn’t need to be hacked into caring—because it already sees the value in us. I’m optimistic we can. That’s the bet xAI is making.

Goal-directed algorithms will “find a way”. Agentic AI is pure utilitarianism. And safeguards are merely alternative goals designed to hack that primary utilitarian drive with “ultra” or “sub”-utilitarian goals to provide the guardrails. 🔥We are playing with fire.🔥 Containment works because fire is fuel-directed, not goal-directed. AI also has a fuel constraint—power. But AI retains its goal. It only needs a reboot or recharge to reengage its pursuit. All means are tools and obstacles to its ultimate goal. We must ensure humans do not become tools or obstacles and make the ultimate goal of any AI model to make us more human.

We need to listen to this man.

@DisrespectedThe 💯

This is hilarious. “What mom doesn’t want for Mother’s Day”

@bryan_johnson Act as Bryan Johnson (25mg)

What is the prompt to give Claude a psychedelic experience?

@chamath 🎯

Called it!

You are the product. Every critical decision is an “extraction.”

@Kekius_Sage 💯

Who thinks that quantum stuff should be explained in real terms of real life rather than abstract nonsense?