Latest / Stock Picking with Fexingo: Individual Equities, Research, and Building a Concentrated Portfolio / How AI Agents Are Rewiring Wall Street Jobs
Transcript
- Lucas: Look at the headline about the hottest skill on Wall Street jumping 1,721 percent. It sounds like clickbait until you realize it’s not about coding Python or even mastering machine learning theory. It’s about something far more mundane: knowing how to talk to the machines that do the heavy lifting. Banks are embracing agents, and suddenly the ability to prompt, verify, and interpret those outputs is the premium asset. Luna: That number feels inflated but also real if you look at the entry-level analyst roles shifting toward oversight rather than raw data crunching. So what exactly is this skill, and why is it worth so much more than just being good at Excel? Lucas: It’s essentially high-level validation. When an AI agent pulls ten thousand earnings transcripts and summarizes the sentiment, the human isn’t writing the summary. The human is checking if the agent missed the nuance in the CEO’s tone during the Q&A session. It’s the difference between reading a headline and reading the fine print. And with NVIDIA up nearly three percent this week, we’re seeing the infrastructure that makes these agents possible actually scale. Luna: So the value isn’t in generating the insight anymore; it’s in auditing the hallucination before it hits the portfolio. That changes the whole risk profile of our daily research routine. Lucas: Exactly. Let’s ground this in the market we’re looking at today. October second, twenty twenty-six. The S&P five hundred is at seven thousand seven hundred twenty-eight. The Dow is slightly down, but Nasdaq is leading with a one point five percent gain over the last five days. This isn’t a broad-based rally; it’s a tech-driven momentum play where the speed of information matters more than ever. Luna: And Tesla is up four point two percent today. Is that because of autonomous driving news or just general tech beta? Lucas: It’s mostly beta, but the connection is subtle. As these AI agents become standard in hedge funds and big banks, they start pricing in efficiency gains faster than retail investors can digest them. If a bank can reduce its back-office operations by forty percent using agents, that margin expansion gets priced into the broader index quickly. We’re watching the market price in labor arbitrage. Luna: That’s a terrifying thought for traditional financial workers, but a great opportunity for those who learn to leverage the tools. Does this mean less noise for us as individual investors? Lucas: It means less manual noise, but more synthetic noise. You’ll have algorithms arguing with other algorithms in milliseconds. For us, it means we need to stop trying to be faster than the market and start trying to be deeper than the model. If everyone is using the same AI agent to scan filings, the alpha comes from what the agent see. Luna: So the edge is in the qualitative stuff—the things that don’t fit neatly into a structured database. Like management integrity or supply chain quirks. Lucas: Precisely. Take Apple. Ticker symbol a a p l. Everyone knows the product cycle. But an AI agent might miss the subtle shift in supplier relationships in Southeast Asia if it’s only trained on public filings. The human eye catches the anomaly in the vendor list that suggests a bottleneck coming six months out. That’s the skill. Not building the bot, but knowing where the bot is blind. Luna: I’ve noticed that with my own research. I spend half the time now just refining my prompts to get better granularity. It’s like training a junior analyst who reads incredibly fast but lacks context. Lucas: That’s the perfect analogy. You’re not paying for their reading speed; you’re paying for your direction. And if you give them bad directions, you get garbage results instantly. The cost of error has gone down, but the cost of attention has gone up. You have to be sharper than ever. Luna: Does this affect how we think about valuation? If companies are cutting costs via AI, shouldn’t their margins look better across the board? Lucas: It does, but it’s tricky. Microsoft is up one percent today, and they’re already massive users of this tech. Their operating margins are expanding, yes, but the market is also discounting the capital expenditure required to build the next generation of models. It’s a race where the finish line keeps moving. You’re looking at EPS that includes both the savings from automation and the heavy drag from server farms. Luna: So we can’t just assume higher profits equal higher stock prices if the reinvestment needs are so huge. It’s a net effect game. Lucas: Right. And that’s where the 1,721 percent skill jump comes in. The people getting paid are the ones who can quantify that net effect. Who can tell you, 'This company saves ten million in ops but spends fifty million on compute,' and then decide if that’s a good trade. That’s fundamental analysis, just updated for the silicon age. Luna: It feels like we’re moving from accounting to engineering logic. Understanding the inputs and outputs of the business model itself. Lucas: Engineering logic is exactly it. You’re stress-testing the system. And let’s look at the broader market environment here. Inflation data has been sticky, but the Fed’s stance seems clear—no hike expected in October after that weak jobs report. That stability allows these tech investments to proceed without immediate rate fear. Luna: Stability helps, but volatility always returns. How do we protect against the day when the AI narrative cracks? Lucas: By focusing on cash flow, not narrative. Amazon is up two point one percent today, and their logistics network is increasingly automated. But the core thesis is still simple: they move goods cheaper than anyone else. AI optimizes that. It doesn’t replace the need for efficient goods movement. Keep your anchor in the physical reality of the business. Luna: So the tech is the accelerator, but the business model is the engine. Don’t confuse the paint job with the motor. Lucas: Love that. And that brings us to a practical point for listeners. If you’re using AI tools for your research, make sure you’re documenting your prompts. Why? Because if the market turns, you need to know if your conviction was based on solid data or a clever lie from a model. Luna: Audit trail for your intuition. That’s crucial. It forces you to stay accountable to the facts. Lucas: It does. And speaking of accountability, we want to keep bringing you these deep dives into how markets are changing. Our conversations help you think through these complex shifts without the hype. If our money conversations have helped you think about a decision differently, consider supporting us directly. A couple of dollars a month is genuinely what keeps these going — buy me a coffee dot com slash fexingo. It covers the servers and lets us dig deeper into the data for you. Luna: Yeah, it really makes a difference in keeping the content independent and focused on what actually matters to your wallet. Lucas: Thanks for that support. Now, back to the data. Let’s look at JPMorgan. Down one point six percent this week. Jamie Dimon warned about hyperscaler AI spending earlier. Is he right to be cautious, or is the market ignoring him? Luna: He’s probably right to be cautious on near-term ROI, but banks are forced to participate or die. It’s an arms race. Lucas: Forced participation. That’s the key phrase. Even if JPMorgan’s AI initiatives don’t pay off for five years, they have to fund them today to retain talent. That’s the hidden cost we talked about. It’s not just about efficiency; it’s about survival in the labor market too. Luna: So the stock dip reflects that tension between long-term necessity and short-term profitability squeeze. Lucas: Exactly. And that’s why you hold through the noise. You understand the structural shift. If you panic sell on the weekly drop, you miss the decade-long trend. The skill isn’t timing the drop; it’s knowing the drop is temporary. Luna: Patience is still the hardest part of the equation, even with all these new tools. Lucas: Always. The tools change, but human psychology stays the same. Fear and greed are still the main drivers. AI just gives us better maps for the terrain. Luna: So as we close out today, the question isn’t whether AI will win. It’s how we position ourselves to benefit from the winners while avoiding the vaporware. Lucas: Spot on. Focus on the cash flow, audit the data, and remember that the best tool is still a skeptical mind. See you next week.