Serge Reda is co-founder of AlphaRithm AI and chair of the technology committee at Fordham's Real Estate Institute, where he also teaches. Trained as an architect, he spent 22 years in development, acquisitions and asset management, ending at a publicly traded REIT where he used Python, which he had learned as a hobby, to manage a large construction cost data set. He knows the gap between real estate and technology from both sides, including how few jobs exist for people who bridge it.

Serge Reda, Co-Founder, AlphaRithm AI

Watch the episode ↗

In this conversation

How a construction invoice data set and a Python hobby turned Serge from a developer into a data scientist.

Why an early adopter's advantage becomes everyone else's cost of doing business, from the internet to AI to quantum computing.

Why Serge says commercial real estate is not a large language model industry, and chatbots are the lowest hanging fruit.

How investors' need for a simple story produced a proptech market of single-purpose tools that do not talk to each other.

The role most firms lack: someone with real estate and technology skills who pulls siloed data into one place.

The commercial real estate industry is not really a large language model industry. It's a much more scientific and analytically focused industry.

Serge Reda

Chapters

From architecture to data science

Twenty-two years in development, then a data set too big for spreadsheets.

Early adopters pay first

How today's edge becomes tomorrow's cost of doing business.

Not a language model industry

Why real estate needs analytical models more than chatbots.

By the numbers

Months to reach 100 million monthly users

ChatGPT2

TikTok9

Instagram30

Serge says use of basic large language model tools has exploded, yet the analytical tools commercial real estate needs still have to be built.

Source: Reuters, citing a UBS study, "ChatGPT sets record for fastest-growing user base" (February 1, 2023) ↗

Why proptech became point tools

Capital rewards a simple story, and real estate is not simple.

The missing word is data

Silos, workarounds in Python, and the job that does not exist yet.

One big data set

What pooled building data could do, by analogy with ChatGPT.

The competitive advantage that's enjoyed by the early adopters just ends up becoming a cost of doing business later on by all the ones that did not adopt it.

Serge Reda

Takeaways

Do not wait for software vendors to integrate. Serge's workaround was to pull raw data out of each system and join it himself in Python, which gave him a view of construction risk that siloed reports could not.

Most users touch only a fraction of any SaaS product. The value sits in combining the data those products already collect across leasing, asset management, accounting and construction.

Firms need people with both real estate knowledge and technology skills on staff, not only as consultants. Few such roles exist today, which is why Serge is pushing Fordham students to build those skills.

This construction software isn't going to talk to that payments software. They're never going to make an integration.

Serge Reda