Dear friends, colleagues and mentors,
Imagine an investment becoming a living, breathing, thinking organism. Its information refreshes as new evidence arrives. Its memory preserves what was known and decided. Its analysis responds to changing conditions, with every number traceable to its supporting record.
You walk into an investment committee with leases, operating reports, site observations and market updates already connected. You can see which assumptions the investment rests on, how they have held up, and which missing fact would change your decision.
You still decide the investment view and the action. But you arrive with fresher evidence, more of it examined, and more of your assumptions tested.
That is what I mean by a super investor: an experienced professional whose ability to observe, question and remember extends beyond what one person - or one team - can reliably hold in mind.
Machine intelligence gives us new ways to build that capacity. And this market gives us a reason to begin.
Why now
On September 16, the Federal Reserve raised its policy rate range to 3.75-4.00%. By September 29, the 10-year Treasury yield stood at 5.26%. Investors cannot plan around an assumed return to cheap financing. Federal Reserve statement, interest-rate data
For a building, the pressure becomes concrete at refinancing.
Take a hypothetical $10m interest-only loan made in 2021 at 3.5%, supported by $750k of annual net operating income (NOI). Assume it matures in 2026 and is refinanced at 7% on the same balance.
Metric | Original loan in 2021 | Refinance in 2026: flat NOI | Refinance in 2026: NOI tracks inflation at 4.13% CAGR |
|---|---|---|---|
Loan balance | $10m | $10m | $10m |
Fed policy target range, for context | 0–0.25% | 3.75–4.00% | 3.75–4.00% |
Assumed loan interest rate | 3.5% | 7.0% | 7.0% |
Annual NOI | $750k | $750k | $918k |
Annual interest payment | $350k | $700k | $700k |
Coverage (NOI / interest) | 2.14x | 1.07x | 1.31x |
Below assumed 1.25x minimum coverage? | No | Yes | No |
Loan rates and the 1.25x minimum are illustrative. Fed ranges reflect September 2021 and September 2026 and provide context rather than a direct property-loan pricing benchmark. Both loans are interest-only; principal repayments would reduce coverage. Federal Reserve, 2021, 2026
The inflation case applies the August 2021–August 2026 CPI increase, equivalent to 4.13% compound annual growth. Figures are rounded; coverage uses unrounded amounts. BLS, 2021, 2026
The table shows how much income performance matters at refinancing: with flat NOI, coverage falls below the assumed minimum; with inflation-linked growth, it remains above it.
That income depends partly on the tenant’s own financial health. Higher borrowing costs can strain the tenant’s P&L, affecting expansion plans, demand for space and renewal negotiations. The owner’s financing and the tenant’s outlook have to be read together.
Will the tenant renew before the debt matures? At what concessions? What improvements would attract a replacement, and how long would that take?
The spreadsheet can price the scenarios. The difficult work is deciding which assumptions deserve confidence - and noticing when the evidence changes.
One connected thinking system
Investment work often moves through separate documents, models, conversations and reporting cycles. New information may arrive quickly while the analysis it should inform takes longer to catch up.
Machine intelligence can help connect fragmented information, test assumptions and keep analysis current. Where relevant data is sufficient, machine learning can offer another reference point on outcomes such as leasing duration or operating costs.
The value is in the connection. A fresh number should reach the analysis it affects without waiting to be copied through another report, model or email chain.
For the investor, three capabilities matter.
It connects evidence to consequences
A change in a tenant’s plans can affect several investment assumptions. A connected system helps investors identify those implications and decide what deserves closer attention.
The investor should be able to examine the evidence behind a conclusion, understand its assumptions and compare alternative interpretations.
It keeps the analysis current
Underwriting is a point in time; a building lives month to month.
As operating performance, leasing conditions and financing terms change, investors need to understand whether their original expectations still hold.
Fresh information is useful when it reaches a decision in time to matter. It should also arrive with a clear source and verification status.
Nothing should update silently. Investors need to see what changed and distinguish verified evidence from a proposed revision.
It preserves the reasoning
Teams accumulate experience, but the reasoning behind decisions fades. We remember the outcome more clearly than what we knew at the time.
Keeping the original expectation, the available evidence and each later revision helps a team distinguish a weak assumption from an unexpected event - and resist making every outcome look obvious in hindsight.
This is the capability I believe compounds most. Experience becomes more valuable when we can revisit it honestly and use it to improve the next decision.
What deserves care
A fluent answer can feel convincing before it has earned our confidence. Models can misread documents, omit qualifications and generate unsupported claims. Traceability lets us check an answer; it does not make it correct.
Real estate adds a second challenge: context is intensely local. Two buildings in the same national category can face different tenant demand, competing supply, power constraints and operating conditions. A broad pattern can miss the detail that decides the investment.
Start with a defined task in a defined market, with enough relevant observations and tested performance. Broader data supplies context; local evidence helps determine its relevance.
Ask machines to connect evidence, monitor changes, compare scenarios and raise questions worth investigating. Require explicit assumptions, controlled access to client information and a visible history of revisions.
People remain responsible for setting objectives, interpreting local conditions, negotiating and choosing which risks to accept. A tenant conversation or site visit may reveal something the record has missed.
The investor decides what to believe and what to do.
What this means for Reml
These principles shape Reml: fresh source information, traceable analysis, continuous monitoring and a visible history of changes.
We are building toward a connected thinking system where new evidence reaches the decisions it affects, and investors can follow the path from source to conclusion.
Leasing enquiries and other operational workflows are valuable uses of AI. Our focus is the professional investor’s judgment, where evidence must be inspectable and responsibility stays clear.
The super investor still walks the building, has the difficult conversation and owns the decision. Machine intelligence gives that investor more capacity to prepare, question and learn.
As you approach your next investment or asset review, which part of your process could a machine strengthen - and where does your own judgment need to become more deliberate?
I would like to hear your answer.
Yours Curiously,
