Five instruments that answer an expensive question while the money is still yours.
The price, before the price.
Formularies, reference prices, thresholds.
A band, not a number.
Because the honest answer is a range.
And the cliff, marked.
A launch price is far easier to lower than to raise.
illustrative
Fold what is worth folding.
One number, taken from the sequence alone.
No alignment. No database. No GPU.
Milliseconds, on a laptop.
Wrong toward folding, on purpose.
A wasted run is recoverable. A structure never attempted is not.
measured record, provisional until the scoring toolchain is ratified; the chain drawn is PDB 7SQG, chain A, the pilot target
It is already in the blood.
The cheapest test in the building, and the most often ordered.
Read whole, not value by value.
Twelve values, one pattern.
The panel worth a second look.
Flagged for review, never diagnosed.
illustrative
Know where it goes before it exists.
Transport, targeting, clearance.
Argued on paper, before anything is made.
Before a synthesis run is booked.
The earliest of the five. We say so.
And the least finished.
Your name on the result.
A real question, not an exercise with a known answer.
A plan frozen before anyone looks.
Pre-registered, and measured against ground truth.
The retraction published, when one is due.
With your name still on it.
A tool that saves sixty per cent of your compute and quietly breaks one prediction in four is not a result, it is an advertisement. Every live number here names its set and its date, and a retracted figure stays printed beside the one that replaced it.
Optifold’s accuracy figures carry the status PROVISIONAL_UNRATIFIED while its scoring toolchain is ratified.
“I aim to empower under-developed or underprivileged communities and their protein research.”
Oliver Choi, founder of ExAnte Labs
ExAnte Labs builds tools that answer an expensive question before the money is spent. Ex ante is the economist’s term for a judgment made before the fact, from forecast rather than outcome, and every instrument here is built on that side of the ledger: the price before the price, the fold before the GPU, the reading before the next draw, the particle before the synthesis.
The lab’s one rule is that a saving is never published without the rate at which its recommendation is wrong, and that a retracted figure stays printed beside the one that replaced it.
Oliver Choi
Founder. Author of record on the Optifold handbook.
The rest of the team is the internship cohort, which is how the lab is meant to grow. These roles are open:
Intern applications go to the same address. Send the question you would want to work on, not a CV.
The address above is a placeholder until the lab’s domain is live.