Building The Search Engine for the Agentic Future

From transatlantic cables to telephones and spreadsheets, major technological changes over the centuries have sparked enormous disruptions to the financial industry.

Today we are on the brink of another breakthrough.

In this blog I will explain why recent developments in AI and onchain finance are set to change the way that you and I interact with markets forever—and how Faro will help make it happen.

Hello, Hyperliquid.

In case you've been living under a rock, you will have noticed that the long-held 'Mullet Thesis' (fintech front-end, DeFi back-end)  is actually coming to pass. (Shoutout to Ryan Sean Adams of Bankless for calling that).

Hyperliquid built perps in a way that actually rivaled centralized exchanges and plugged it into multiple consumer front ends via builder codes. Then we have other contenders like Lighter forging integrations with Robinhood. Coinbase incorporating Morpho on the backend. Paypal doing settlement on stablecoins. The list goes on.

In short, onchain finance is no longer just a niche side channel for crypto natives with a distaste for KYC. It's becoming the backend plumbing infrastructure enabling users to do more in a transparent and secure way. All through the familiar easy UX you'd get from Robinhood app or Revolut

We believe this thesis will continue to play out and become the default set up for everything from capital markets to intergovernmental trade 

But you probably know that already. The next bit is where things get interesting.

Superabundant intelligence

AI now looks to be the defining technological creation of our generation

Open source models are catching up with (and even outpacing) frontier labs, suggesting LLMs will soon be widely available to everyone. At the same time, the cost of compute is getting relentlessly cut each year by hardware and algorithm refinement.

Thus AI intelligence is not only getting cheaper, but also becoming more freely available—mirroring a familiar trajectory:

Early web browsers like Netscape and Opera were paid. Before they were killed by free and widely available alternatives. Today, browsers survive on what's wired around them like search defaults and ecosystems.

Back in the early days of computing, the OS was the product. But then Linux made it free creating abundance. Companies survived by moving into support and distribution (like RedHat) or building on top (AWS).

In both cases, as technology made the product more abundant, the scarcity and value migrated further up the stack.

In AI, this pattern is playing out even faster, with open source models quickly and cheaply harvesting outputs to learn models that the frontier labs spend hundreds of millions to train.

So what's left once this has all converged on abundant super intelligence?

Above the LLMs is what we're building at Faro: The application that orchestrates data, tools, and context to turn generalized intelligence from any AI model into a personalized financial terminal.

The ever-expanding data pool

A world where AI sits at the centre of human activity is a world producing exponentially more data. More agents, more interactions, higher frequency.

This shift will also change how we interact with financial markets, and in doing so, reshape them entirely.

Today, the default way for most people to touch financial markets is an index fund. The S&P 500 has returned roughly 9% annually for a century, and that track record has made it the one-size-fits-all answer to investment needs, subscribed to by everyone from pensioners to first-time investors, regardless of their actual goals.

In future, instead of buying an off-the-shelf product or outsourcing it all to a fund manager, you'll be able to disclose your situation privately to an AI–I have three kids, this much left on the mortgage, this much spare cash each month, this much tolerance for volatility–and have it build you a data-backed index fitted to your specific situation.

The result will be direct interaction with markets at whatever level of granularity you want, free of any third party's bias, incentives, or risk profile. 

You stop buying a standardized product and start manufacturing the outcomes you actually want.

The Search Engine for the Agentic Future

Navigating this ever-deepening pool of data requires what is effectively a search engine. 

This is the analogy I keep returning to when explaining Faro: the search engine for the agentic future.

We have this ever-expanding pool of data and what we're really building is a personalized mechanism that sits between one specific user, a manifestation of all the things they find valuable, the end data, and the expression environment, which is financial markets for us today.

The engine takes a question like "why is bitcoin down 10% today?" and maps it, together with what that specific user cares about, onto the ever-expanding pool of data we're collecting underneath, then it represents the answer back in the way that's maximally useful. Sort of like how YouTube serves you a subset of a massive content pool tuned precisely to your preferences, but beyond recommendation to encompass interpretation and synthesis of the data too.

Building this way means maximizing personalization: we construct a flexible harness with as many pivot points as possible, so that as a user interacts with Faro, the harness adapts and molds itself around that specific person.

This level of personalization is only made possible today by the recent development of onchain markets and LLMs.

We're excited to build on this and empower people to finally represent their own desires in how they interact with financial markets.

Follow @Faro and @Abishekfi for more updates as the build progresses.