CEO David Trainer recently joined Karlo Jo Helms, Founder & CEO of JOTO PR Disruptors on the Disruption/Interruption podcast to pull back the curtain on decades of Wall Street manipulation.
Topics from the discussion include:
- how Wall Street is designed to serve Wall Street, not investors,
- why what is unethical isn’t always unlawful,
- why 96% of Wall Street ratings are buy or hold,
- how New Constructs’ superior research cuts through Wall Street noise,
- how we generate real alpha,
- and much more!
See below for additional details on how New Constructs developed real AI for stock picking.
How We Use Data to Make AI Work
Financial analysis is inherently a deep and complex field, and general-purpose AI models lack the expertise to navigate it effectively.
Before we expect an AI to produce reliable stock picks, we must endow it with the requisite subject matter expertise (SME) to generate alpha.
We endow our Robo-Analyst AI with over 4 billion expertly-verified datapoints, which leverage deep analysis of financial footnotes to calculate a proven-superior measure of earnings, Core Earnings.
Armed with superior data, our AI agent, FinSights, provides investors a real edge, one that cannot be found in any large language models (LLMs).
Small Language Models Over Large Language Models
Unfortunately for users, current large LLMs fail miserably at understanding the meaning and nuances of financial data in filings because they have been trained using low quality financial data. Any stock picking or investment research system built on these models is using bad data, and we all know the timeless saying about models: garbage in – garbage out.
If the financial models used to train the AI can’t generate alpha, AI isn’t going to do any better. LLMs handle a broad range of general tasks reasonably well, but their performance declines sharply when faced with more sophisticated tasks.
There’s no better proof that the quality of training data is paramount to AI’s success than the mounting evidence that small language models (SLM) significantly outperform LLMs.
SLMs focus on a narrower and more specific domain, which enables human experts to ensure the data is accurate and properly organized so that the rules about how to use the data produce reliable results.
For example, the SLM created by New Constructs powers market-beating stock picking. We’re able to empirically prove the accuracy and reliability of our SLM because it is based on:
- Publicly available data in financial filings,
- Systematically gathering data into a clear taxonomy that supports a sophisticated ontology, and
- Metrics and signals whose accuracy are demonstrated empirically by stock market outperformance.
Google Cloud recently invested millions of dollars to build an AI Agent for Investing, called FinSights, to demonstrate the art of the possible when their AI was powered by accurate data. Learn more here.
For additional details, see our report “Thinking Small Drives Big Leaps in AI.”
How We Protect from Rogue AI
An autonomous AI that has gone rogue and wants to accumulate resources to fuel its own growth or a bad actor who wants to use AI to enrich themselves would find the stock market an obvious target. The stock market is a system where misinformation, deployed at scale and at speed, can generate enormous wealth in a very short time. And unlike nuclear infrastructure or power grids, market manipulation doesn’t require physical access. It requires only the ability to corrupt the information that investors rely on.
Luckily, there is a solution to protect the market and investors alike.
Protecting the stock market means protecting its core function: allocating capital to companies that earn the highest returns on it. The best way to prevent a malevolent force from corrupting that function is to beat it to the punch. How? Widely distribute the truth about company fundamentals before that truth can be distorted.
The logic is straightforward: the longer markets operate with clean, verified data, the harder it becomes for any bad actor to move the numbers without triggering skepticism among investors who’ve grown accustomed to the real ones. The inoculation, in practical terms, means giving all investors access to verified, core statistical truths about the fundamentals of companies, or a single, reliable baseline that is hard to quietly corrupt.
This is not a novel idea. At 53:09 in Episode #158 of the All-In podcast, Chamath Palihapitiya called for exactly this: “AI that crawls 10-Ks and 10-Qs to generate statistical measurements of all public companies.”
What Mr. Palihapitiya’s call implies, and what all professional investors already know, is that reliable statistical measurements simply do not exist for most companies. The truth about stocks is hard to get.
See more on how we protect the stock market from rogue AI in our report “The Stock Market Is Rogue AI’s Most Obvious Target. Here’s How We Protect It.”
Trustworthy AI is Most Important
For AI to evolve into a trusted partner for completing serious work, we need to be honest about how it performs.
The primary goal for AI should be to build trust. Instead, the focus on making money is leading the AI companies down the wrong path. We detailed one such fraught path in “AI Is Designed To Lie To You.”
We all understand that the AI providers need to make money to survive, continue to develop their technology, and offer competitive products. But, without trust, users will never pay the prices AI providers need to charge to earn an adequate return on the huge amount they’ve spent on building out their AI offerings.
On the other hand, trust can be the most powerful driver of loyalty, which is among the strongest drivers of long-term profit growth.
More importantly, the AI provider that is first-to-market with trustworthy AI will see radically faster adoption rates compared to peers. And, first-mover advantage could be huge here. Imagine how quickly users would flock to an AI that they could trust to reliably perform complex and sophisticated tasks.
It bears repeating: the most essential ingredient in building reliable and trustworthy AI systems is accurate data. As the saying goes “garbage in, garbage out”. No matter how large the infrastructure or sophisticated the model, an AI system trained on unreliable data will be unreliable.
The tools and solutions outlined above aren’t just a dream. Accurate data exists. AI for stock picking that drives real alpha exists. Start leveraging it in your investing process today.
This article was originally published on May 28, 2026.
Disclosure: David Trainer and Kyle Guske II receive no compensation to write about any specific stock, style, or theme.
Questions on this report or others? Join our online community and connect with us directly.