Ram Ahluwalia says artificial general intelligence has cleared exactly one bar so far: the CAPTCHA test, which he says he personally fails "five out of eight times."
Nvidia's Jensen Huang has all but declared AGI arrived, running on Nvidia's own Blackwell chips. Ahluwalia reads it the other way: a chipmaker crediting itself before AMD can claim the same thing, on the back of an OpenAI model that just shipped.
"These open weight models are acid for margins. Really caustic."
Ahluwalia is CEO and co-founder of Lumida Wealth Management, a New York advisory built around alternative assets, and founded the credit-analytics firm PeerIQ before it was acquired by Cross River Bank. He studied economics and philosophy at Columbia, which does most of the work in this one.
I listened to the full episode so you can skip it. 15 minutes of audio, 8 minutes of reading.
Here are the 8 takeaways that matter.
👤 Speaker: Ram Ahluwalia, CEO and co-founder of Lumida Wealth Management, who founded the credit-analytics firm PeerIQ before its sale to Cross River Bank
📰 Published: 9 September 2026 on YouTube
🔴 YouTube | 🟣 Apple Podcasts | 🔗 Show notes | ⏱️ 15 min | ✅ Time saved: 7 min
Key Takeaways
He calls Jensen Huang's AGI declaration "purely political," not analysis
Says Nvidia's incentive is to take the victory lap before AMD can claim the same win, and to credit OpenAI, one of its biggest chip customers
Open-weight models are "acid for margins" for Anthropic and OpenAI, he says
Expects a crowded field of AI leaders in three years, not a two-company race: "Search engine wars all over again, right?"
His own daily use of GPT-6 Astra convinces him AGI isn't here
Calls it a "token hog" with "careless oversight" that won't suggest faster ways to deploy an app even when asked
The real test for AGI is pragmatism, not consciousness, he argues
Credits the idea to philosopher William James: a system counts as AGI when it's reliably useful, not when it seems self-aware
He blames "bad philosophy" for corroding modern thought, from French existentialism to Vienna School logical positivism
Names Karl Popper and Friedrich Hayek as the correctives, and says George Soros studied Popper closely
"AGI just passed the CAPTCHA test," he says — a bar he claims to personally fail more than half the time
His bar for the real thing: an AI that a Mark Zuckerberg figure could hand a market and a budget to, unsupervised
1. Jensen's Political AGI Lap
Ahluwalia opened by dismissing Jensen Huang's recent comments claiming AGI has effectively arrived.
"That's it. Purely political." He argues Nvidia's incentive is to take credit for the milestone before AMD can claim the same thing, running the workload on Nvidia's own Blackwell servers
He said Huang also needs to "pat OpenAI on the back" since it's one of Nvidia's largest chip customers, even as OpenAI's own open-weight competitors undercut the business case for buying more chips
"These open weight models are acid for margins. Really caustic." He expects a crowded field rather than a two-company race three years out: "It'll be many people on the top of Mount Everest, not two" — comparable, he said, to the early search-engine wars, with specialization by niche likely
2. Astra Isn't AGI, It's Slow
Turning to his own use of OpenAI's newly released GPT-6 Astra, Ahluwalia said the tool undercuts the AGI claim on contact.
"Slow as molasses." Describing Astra, he said: "It's a token hog. It has careless oversight." As an example, it won't suggest automatic cloud deployment to speed up an app build even when asked directly for the most effective approach
"I call it Forrest Gump. That's what it is." — then corrected himself, noting the fictional Forrest Gump actually ran fast, so the comparison undersells how slow Astra is
He expects the workflow gaps to get fixed but said that's beside the point: "But it's a far cry from AGI. We're not there yet. We're so not there yet."
He named memory and compute as the real constraints, mentioned a fireside chat in New York on September 16 with Michael Preck, and said Lumida investors would get an invite
3. The Practical Test For AGI
Asked by himself what a measurable AGI test would even look like, Ahluwalia turned to philosophy rather than a benchmark.
"But the test for AGI is the practical test. It's a very practical test." He credited the framing to William James and the school of pragmatism, which he called a distinctly American philosophy
He had the AI chatbot Grok explain James's own answer: an idea is true if it proves useful in practice, so a system counts as AGI once it reliably delivers human-level results on real tasks — measured against benchmarks like the Turing test for conversation and ARC for abstraction and reasoning, with no single test sufficient on its own
On what makes it distinct: "what's distinct about the American approach... is this focus on is it useful? Did it create value? It's very entrepreneurial" — a cultural trait he called a source of competitive advantage
4. Bad Philosophy, Bad Outcomes
Ahluwalia argued that a country's philosophical tradition has real economic consequences, contrasting the American approach with two European ones he called failures.
"The French had the terrible misfortune of suffering World War I and World War II, and they went into this deep existentialist pit that they haven't recovered from"
On the Vienna School's logical positivism: "Essentially, you put labels on everything, and everything is a matter of syntax. There's no meaning, they're pointers, they're definitions, and everything is syntactical." He linked it to modern moral relativism and criticized the "academic left" for stripping concepts like excellence and the good life out of everyday thinking
He named Friedrich Hayek and Karl Popper as counterweights, citing Popper's The Open Society and its theory of epistemology, and noted that George Soros studied Popper's work closely
5. Deutsch And The Multiverse
Asked to identify a modern standard-bearer for the pragmatic, "technological optimism" tradition, Ahluwalia named physicist David Deutsch.
"David Deutsch is a super badass physicist, really impressive individual." He credited Deutsch's The Beginning of Infinity and his work on the theory of quantum computation
He said Google cited Deutsch's work when it built its quantum computer, and again when it reported its first evidence of quantum computation — concluding, in Ahluwalia's telling, that the result is "logically consistent" with living in a multiverse
6. Plato Versus Aristotle
Ahluwalia traced the same philosophical split back to antiquity, contrasting Plato's abstractions with Aristotle's empiricism.
"Imagine the perfect triangle. Imagine the perfect square. Now imagine the perfect door. Now imagine perfect justice." He described Plato's theory of forms as the start of purely abstract intellectual concepts
He credited Aristotle as the first empiricist — grounded, practical, the founder of biology, taxonomy and the syllogism that fed into jurisprudence — and called him one of the most impactful people in history, one who "eclipsed his teacher, Plato"
7. Posner, Becker, Chicago
Ahluwalia extended his list of philosophers who got it right to two University of Chicago economists.
He named Richard Posner, a former economics-of-law professor who studied antitrust, and Gary Becker, who won a Nobel Prize for applying economic analysis to illegal but voluntary behavior — gambling, prostitution, drug dealing, family formation and divorce
"That was the thread that's led to productive growth in civilizations and it's a big part of American thought process." He said that thread — James, Dewey, Deutsch, Popper, Posner, Becker — is "on the right side" of philosophy, and that other parts of the world don't necessarily share it
8. Real AGI Vs. CAPTCHA
Ahluwalia closed by applying his own "is it useful" standard to the present moment.
"AGI just passed the CAPTCHA test. It just passed the CAPTCHA test." He said the CAPTCHA test — identifying a bus in a grid of images — has effectively replaced the Turing test as the bar being cleared, and joked: "I seem to fail that test five out of eight times." He called that a far lower standard than real AGI
His own definition: "It is spin up this business division on the fly, get after this market opportunity, and let me know how much energy you need." He compared it to a founder like Mark Zuckerberg delegating an entire business unit and simply asking to be told if the team gets blocked
He said memory, compute and natural gas remain the binding constraints on getting there, and pointed to strength in master-limited-partnership energy ETFs as one way to track the data-center buildout, which he called still early
"We need AI agents to watch the other AI agents." He compared the current buildout to onboarding a new hire: more work up front before any output improves
Bonus Insights
He opened with a joke that AI is creating more work, not less: "AI, they should call it the full-time employment program." He said Bernie Sanders "got it completely wrong" on that point
On whether either of them is even conscious: "I don't know if you're conscious. You don't know if I'm conscious. I'm pretty sure I'm an NPC. Pretty sure you're an NPC, too."
Ahluwalia's bottom line is that the industry has cleared a much lower bar than it's claiming, and the standard he'd actually accept — a system trusted to run a business division unsupervised — is still years away, gated by memory, compute and energy rather than by any single model release.
Products, Companies & Tools Mentioned
GPT-6 Astra (OpenAI's newly released model, which Ahluwalia uses daily and calls a "token hog" with "careless oversight" — his main evidence against the AGI claim)
Nvidia (Jensen Huang's Blackwell chips are what Ahluwalia says the AGI "victory lap" was run on)
OpenAI and Anthropic (The two labs Ahluwalia says open-weight models are most "caustic" for on margin)
Grok (The chatbot he queried live to explain William James and pragmatism's answer to what counts as AGI)
Meta (Mark Zuckerberg is Ahluwalia's example of what a real AGI-run business division would look like)
Books & Resources Mentioned
The Beginning of Infinity – David Deutsch (The physicist Ahluwalia calls the modern standard-bearer for pragmatic, technological-optimist philosophy, and whose work Google cited around its quantum-computing claims)
The Open Society and Its Enemies – Karl Popper (Cited for its theory of epistemology; Ahluwalia says George Soros studied Popper's work closely)
If this was worth your time, send it to someone closer to the industry than you are.
Get the latest market chatter as it happens: