How to Actually Make Money in Stocks

01 /The Concept

Seasoned equity investors rarely bet everything on a single stock, even when they're highly confident in a company. They typically buy into several companies expressing the same "investment thesis" – because it's impossible to know in advance which one will break out. The portfolio weights aren't equal, though: they reflect relative financial performance and get periodically rebalanced as conditions change. This approach is called index investing.

The best-known index is the S&P 500 – 500 of the largest publicly traded companies, weighted by market cap, maintained by S&P (Standard & Poors). Most other indices are benchmarked against it, because it remains one of the most reliable long-term performers in public markets.

But the holy grail for many investors is constructing a custom index that beats the S&P 500.

Generated Assets gives investors the sandbox for exactly that kind of experimentation. The platform makes it fast and easy to translate any investment thesis into an index – a specific list of stocks to buy – without doing all the research manually.

All it takes is describing the thesis to the platform's AI in plain language.

The homepage showcases examples: "B2B companies that generate less than 10% of revenue from consulting," "companies with business models AI hasn't disrupted yet," "companies minimally exposed to new import tariffs," "companies with more than 50 million social media followers," "subscription businesses with the lowest churn rates."

In response, the AI identifies matching companies and assembles them into an index. The list can be adjusted manually, and backtested performance against the S&P 500 is shown immediately.

If the results look promising, investors can start building a position in that index or publish it publicly. Public indices are ranked by current performance. Some of the top-rated at the time of this review:

"Companies with very large user bases but currently very low revenue per user." The thesis: sooner or later these companies will crack an effective monetization model, enabling them to rapidly extract value from their existing audiences – which should drive the stock price up.

"Companies developing technology, infrastructure, or services for missions to Mars." The thesis is clearly rooted in conviction that Elon Musk will eventually follow through on his colonization ambitions – a bet on the emerging Mars economy before it exists.

Generated Assets isn't a standalone startup – it's a product launched by investment brokerage Public.com. It was first spotted on Product Hunt a few days ago.

02 /Why It Matters

There are essentially two formalized approaches to equity investing – beyond pure intuition.

Technical analysis looks at stock price charts and patterns. The underlying assumption is that market participants behave somewhat predictably based on external signals: news flow, price movements, what other investors are doing. It works, sometimes, but tends toward the read-the-tea-leaves end of the spectrum.

Fundamental analysis focuses on the internal numbers – the financial results that public companies are required to disclose in quarterly and annual reports. The assumption is that financial performance will eventually be reflected in the stock price. The problem is that financial performance is backward-looking, and investing is fundamentally about the future.

Both approaches remain popular primarily because the underlying data is publicly available. But for more nuanced and forward-looking decisions, investors need to factor in a much wider range of inputs – and the teams at large investment firms spend enormous resources doing exactly that.

AI can now find and cross-reference many of those signals far faster than any human team – which will inevitably reshape the analytical methods available to investors and democratize access to more sophisticated strategies. That could meaningfully disrupt a market currently dominated by large brokerages and funds running proprietary analysis with large teams.

Similar AI-powered analytical platforms are beginning to emerge in other business domains.

Openmart ([related review](/review/jeto-uzhe-dengi-no-mozhno-zarabotat-eshhjo-bolshe)), a Y Combinator graduate, raised $2.75 million last summer for an AI that builds a database of small local businesses that could become buyers or distributors for large national brands. The platform lets clients find companies on almost any criteria – pubs with jukeboxes, restaurants using specific ingredients – tasks that would take a human researcher days to complete.

Daash ([related review](/review/hochesh-znat-skolko-i-chego-prodajut-tvoi-konkurenty)) raised $5.5 million in January for a platform that lets beauty product manufacturers track competitor sales. It cross-references consumer surveys with indirect company data to estimate not just total competitor volume but per-product and per-channel breakdowns.

Verata ([related review](/review/chtoby-pobedit-konkurentov-nuzhno-znat-chto-u-nih-proishodit)), another recent Y Combinator graduate, built a platform helping private equity firms identify strong executives for their portfolio companies. Its AI can estimate how specific candidates have influenced revenue or fundraising at previous employers – far more informative than a résumé full of self-praise.

Incidentally, that executive signal could be a powerful feature for investment AI too: if it turns out that a company just recruited someone with a strong track record in a comparable role at a comparable business, that's worth factoring into a growth forecast.

03 /Opportunities

AI is now capable of doing more than improving existing analytical methods – it can fundamentally change the structure of how analysis, decision-making, and business processes work.

A recent example: Auctor ([related review](/review/hochesh-perestat-terjat-na-jetom-dengi)), currently in Y Combinator, claims to have invented a new model for custom software development and implementation that differs from both Agile and Waterfall. Simplified: the platform creates a digital twin of the client, which can be queried for feedback without chasing down human stakeholders. That twin also monitors the project as it progresses, flagging deviations from requirements in real time.

The broad directional opportunity here is building AI platforms that can transform the underlying processes of information discovery, analysis, and business operations.

Investing is one of those domains. When a well-known financier was asked how he made such effective decisions, his answer was: "Because I study connections, not isolated facts." AI can analyze those connections far more broadly and granularly than any human analyst – without requiring enormous investments of time from the people making the final call.

And to be clear – this is not the same as asking ChatGPT "which stocks should I buy?" In that form, a general-purpose AI just summarizes whatever opinions it finds online. Commissioning real research requires careful prompt engineering and feeding in large amounts of additional data – and without that structure, you end up back at conventional technical or fundamental analysis.

Specialized platforms should do this better and more reliably. So the choice is: use the next generation of investment platforms when they arrive, or go build one yourself so others can use it.