July 12, 2025
Sir, I refer loosely/closely to “Humans must remain at the heart of the AI story” by Marc Benioff, FT July 11, 2025.
About bank regulations, an obsession of mine, I dialogued with the recently introduced to society Grok 4, SuperGrok.
In reference to whether a hypothetical “Super-Duper-Grok” could outperform the decentralized, messy wisdom of millions in the market, without amplifying systemic risks it stated:
“The market, as an aggregate of millions of individual decisions, is a remarkable mechanism for pricing risk and allocating capital. It’s not perfect—prone to bubbles, panics, and mispricings, its strength lies in its diversity of perspectives, incentives, and local knowledge. Interest rates act as signals, reflecting collective judgments about risk and reward.
Now, imagine a “Super-Duper-Grok”—an AI with vastly superior data processing, predictive modeling, and decision-making capabilities. Could it outdo the market? In theory, it might:
But here’s the catch: centralizing risk management in a single AI, no matter how advanced, introduces new vulnerabilities:
Single Point of Failure: If Super-Duper-Grok miscalculates or is hacked, the fallout could be catastrophic.
Model Risk: Even a super-AI relies on assumptions and data inputs. If those are biased or incomplete (e.g., missing “unknown unknowns”), it could misjudge risks.
Loss of Human Judgment: Markets incorporate intuition, skepticism, and ethical considerations—qualities an AI might lack. An AI, however sophisticated, might over-optimize for measurable metrics, ignoring intangibles like cultural values or long-term societal resilience.”
Sir, if only those besserwisser bank regulators in the Basel Committee had possessed such humility. Then they would not have decreed those risk weighted bank capital requirements which so dangerously and weakening distort the allocation of credit.
Sir, it would seem that without stupid human intelligence to feed it, artificial intelligence could also become stupid. Thanks God, it seems that we will still be needed. Let's make sure we fulfill our part.
July 03, 2025
What if a Basel Committee had regulated the payouts of casinos?
Sir, in “The risks of funding states via casinos” FT, July 3, Martin Wolf writes: “In the run-up to the GFC, the dominant form of lending was to the private sector, particularly in the form of mortgages.”
The dominant form of lending was indeed mortgages but that obscures the truth of what really happened.
Mortgages to the subprime sector were packaged into securities which, if obtaining an AAA to AA rating could, thanks to 2004 Basel II, be held by US investment banks and European banks against only 1.6 percent in capital, meaning they could leverage 62.5 times with these.
As should have been expected, the temptation to package MBS sausages with the worst ingredient, which maximizes profits when then able to sell these as made with pure tenderloin, proved irresistible.
Sir, how long will FT keep being obsessed with minimizing the distortions of the Basel Committee’s risk weighted bank capital requirements?
Casinos? Ask Mr. Wolf whether he believes there would be any casinos left if its regulators ordered these to make higher payouts on safer bets, e.g., red or black than on riskier ones, e.g., a number?
Subscribe to:
Posts (Atom)