Quality control for AI in Drug Discovery. Independent audits + signed Validation Passports. Molecules, Sequences, Structures. Validation one can Trust.

Delaware, USA
Bits in Bio is coming to Miami. First mixer: Wednesday, Nov 18, 6:30 to 9 PM at La Tropical in Wynwood. Free, for everyone in South Florida working where software, AI, and biology meet. Hosted by our founder with Bits in Bio. luma.com/uk92lg9g
17
A Validation Passport is bound to 3 things: 1. The exact evidence evaluated 2. The engine release 3. The date Change any one of them and it's a different record.
1
13
Calling everyone in the Bio community. For anyone located in FL or in South Florida during Nov.18.2026 You should definitely attend 💯
I am hosting Bits in Bio Presents, A Mixer in Miami. Join me! Bits in Bio is a global community of more than 13,000 people working where software, AI, and biology meet. On Wednesday, November 18, we're bringing it to Miami for the first time, 6:30 to 9:00 PM at Cerveceria La Tropical in Wynwood. No panels. No pitches. No slides. Just everyone in South Florida building at the intersection of computation and biology, in one room. Researchers, clinicians, founders, engineers, students, investors, and anyone trying to break in. UM, FIU, the Miller School, Nova Southeastern, and everyone in between. Why La Tropical? It started in Havana in 1888, became Cuba's most beloved brewery, and was reborn here in Wynwood. As the son of Cuban immigrants starting something new in this city, that one's personal. Coming alone? Good. Most people will be. Introducing people is the job that night, so come find me. Huge thanks to @AaronBlotnick and @bitsinbio for making this happen. luma.com/uk92lg9g via @LumaHQ
1
1
17
Interesting article for those wanting a good read on the current landscape 🧠
1
11
Some evaluations shouldn't be run. LP-GATE checks first whether a submission can be evaluated at all, and turns it away with a reason if it can't. A confident answer to the wrong question is worse than no answer.
1
1
23
114 AI Drug Discovery partnerships were signed in 2025, up from 84 the year before Most of the headline value is milestones, Not the Cash. Pharma pays a little up front for access and a lot later for proof. What happens between those two payments is the part nobody can measure?
1
1
1
30
Source: DealForma, AI/ML drug discovery 2025 review.
9
TRAIRC TECHNOLOGIES retweeted
Most of the number in a model-license deck cannot be re-run by the person writing the check. I wrote down why, and what a record would have to be.
Article

A leaderboard is not a record

Someone is about to pay for a computational platform. The deck has a chart. The chart says the model works. Ask what the number was measured on, against which molecules, under a split that was locked

1
1
6
456
Four state, not a score. In Tolerance Marginal Out of Tolerance Could not be measured The fourth carries the most weight. A system that never says "couldn't check this" isn't checking anything
3
38
A Validation Passport doesn't tell you a model is good It states what the evidence behind a number can and cannot support, says out loud where nothing could be evaluated, and binds all of it to the exact data and code that produced it. Anyone can verify it LP-GATE
1
2
35
Manufacturing has spec limits. A part is in tolerance Marginal Out of tolerance Couldn't be measured LP-GATE reports those same four states for a computational claim. The fourth one carries the most weight. A system that never says "couldn't measure this" is decoration.
1
1
42
"We checked for leakage" Is not a finding Which checks? At what thresholds? Against what corpus? Declared before or after you saw the numbers??? LP-GATE answers all four...in writing and signed ✅
1
46
A model is only as trustworthy as the experiment that measured it. Most of this field measures models carefully and measures the measurement... Not at all 💀 LP-GATE is the second measurement ✅
1
1
30
LP-GATE reports four things: Passed Warn Failed Not Evaluated The fourth is the one nobody else reports. If a question wasn't examined, we say so. Otherwise silence reads as a Pass, and silence is where most of the optimism in this field lives
1
1
27
Every Industry that ships something consequential has a step whose job is to catch problems before they cost money. Pharma manufacturing Semiconductors Aerospace AI Drug Discovery has a benchmark score and a deadline. LP-GATE is the missing step 🧬
1
3
36
Most Validation tools return a number. A number is something you take on Faith. LP-GATE returns a signed record anyone can verify without an account and without trusting us. That difference is the entire product🧬
1
25
LP-GATE doesn't score your molecules. It checks whether the evidence behind your claim holds: - Leakage - Split integrity - Controls - Calibration Then it says what Passed, what Failed, and what was never evaluated at all. Audit✅ Not appraisal🚫
1
2
77
Every result has two numbers The one that was reported The one that survives an independent check In most of this field, only the first has ever been computed
1
22
A benchmark split is a hypothesis, that these two sets are Independent. Nobody tests the hypothesis 🚫 It gets assumed and then reported as a fact 💀 That isn't a science problem. It's a missing control
1
24