What can LLMs do and more importantly, not do?

 

A very interesting position paper worth a read for any follower of AI: “LLMs Can't Jump: Why the Abductive Leap is the Final Frontier of AI Discovery.” The thought is simple. AI may be good at some forms of inference, like deduction and induction, but it is not able to do the third type: abduction. “While AI can compress data (Induction) and prove theorems (Deduction), it cannot yet recreate the intuitive leap Albert Einstein took to formulate the axioms of General Relativity—a process rooted in embodied physical simulation rather than symbolic manipulation.”

The Three Pillars of Inference

The paper adopts the framework of Charles Sanders Peirce to categorize AI’s missing link:

• Deduction (Rule + Case → Result): Applying a known law to a specific situation. (AI status: SOTA - Achieved)
• Induction (Case + Result → Rule): Spotting patterns in data to find a general rule. (AI status: SOTA - Achieved)
• Abduction (Rule + Result → Case/New Rule): Inventing a hypothesis to explain a surprising or singular phenomenon. (AI status: The Missing Jump)

There is still room for importing thinking by analysts. The market has to understand which skills are important for analysts and what can be done by quants.  

More than a Korean bubble problem - US retail is selling

 

The bubble in Korea has been viewed by some as an isolated event. There is no question that Korean-specific regulation and behavior were a strong contributor to this bubble, yet we should look at what is happening in the US. One, the semiconductor sector is showing strong declines. Not like Korea, but the pattern is similar. Two, the move in the AI and IT sectors has been driven by retail, and retail within the US is reversing. The market is changing, and the marginal retail investor is a strong seller who is moving to more protective assets. 

Market sentiment is changing, and the positive forecasts in the last quarter are being revised. 

Yields at levets not seen in a decade

 

Since the pandemic, 10- and 30-year Treasury yields have been on a steady march higher. We have seen continued inflation above target, continued budget deficits, the shock of pandemic QE, and no strong policy moves to stop the ascent. We are now seeing rates that will take us back up to pre-GFC levels. Could this be considered normalization of rates? This is hard to argue when you look at the combination of inflation and budget deficits. We are moving into a new period of rate behavior, and levels are unlikely to move back to the 3 percent range unless we have a large economic slowdown.  

The equity maket sectors are getting more disperse

Using the v-Lab data, we are seeing that correlations across sectors in the US are falling and showing more dispersion. If the stock market, like all markets, can be viewed as a network, we are seeing the network expanding after a period of strong connection in the first quarter of 2025. We are seeing a disconnect in the IT sector but also across real market sectors. This can be viewed as a market for stock pickers.