Brandon Ballinger

Building @EmpiricalHlth - don't die of heart disease. Previously @AppCardiogram, @GetSift, @HealthCareGov rescue team, @BrexHQ, & @Google AI.

[Brandon Ballinger @bballinger]

@bhalligan @HubSpot @dharmesh Congratulations!

[Brandon Ballinger @bballinger]

@pdhsu Average LDL was 65 mg/dL in HORIZON, so my money's on participants just having too little cholesterol-associated risk to show an impact.

[Brandon Ballinger @bballinger]

@MaBMortensen @Drlipid @khurramn1 @DrMichaelShapir @rblument1 Really impressive!

[Brandon Ballinger @bballinger]

34% of U.S. women below age 50 have iron deficiency under new hematology guidelines, which raise the threshold of ferritin to 30, 50, or 100 ng/mL depending on circumstances:

[Brandon Ballinger @bballinger]

@MiguelriosEN I do wonder if AI orgs will have fundamentally different “physics” than human orgs.

For example, does Conway’s law still hold when communication bandwidth is 8 TBps?

[davidlee @davidlee]

1/ I'm hiring an investor at Samsung Next, based in the Bay Area. No investing experience preferred. You'd work for me.

[Brandon Ballinger @bballinger]

An LDL of 70 mg/dL would already put you around the 10th percentile.

Yet a new study shows cutting cholesterol even further, to <55 mg/dL, reduced cardiovascular events by a third.

IMO, surprising that reductions at the low end still seem to yield such large benefits. I'd expect things to saturate much earlier.

[Brandon Ballinger @bballinger]

@stephenrobles @EmpiricalHLTH is updated for Duo: https://t.co/I47yzmbP2m

[Sebastien Bubeck @SebastienBubeck]

3SAT in subexponential time (2^{n / log^eps(n)}) is maybe the biggest surprise I have seen during my 20 years following theoretical computer science (still remember the feeling 25 years ago seeing the diff a 3 versus a 2 can make for complexity 😅)

[Brandon Ballinger @bballinger]

@ferenstein @gjurvetson @swerdlin We trained a foundation model on blood tests and wearable data along the lines of what you propose. Could be adapted for this use case: https://t.co/RcdOrpmw5J

[Brandon Ballinger @bballinger]

@armenberjikly @EmpiricalHlth is building this.

[Empirical Health @EmpiricalHLTH]

iPhone Duo launches today. To test Empirical Health without hardware, we built a full 3D simulator with inverse camera projection, frosted-glass blur, hinge-anchored layouts, and path-traced lighting.

[Brandon Ballinger @bballinger]

@vintweeta @annewoj23 I wonder if this is true for all biomarkers? Their ideal use is building a daily dashboard of healthy behaviors, rather than putting the biomarkers themselves front-and-center.

[Brandon Ballinger @bballinger]

@martinastks Cool. The concept reminds me a bit of @Detour from @andrewmason, which I thought deserved more play.

[Brandon Ballinger @bballinger]

@akiffpremjee @EmpiricalHLTH There's care delivery -- we do prescriptions and doctor video visits in the advanced program, plus re-tests at 3 and 9 months so you can measure progress.

[Brandon Ballinger @bballinger]

@ChrisJBakke Long tradition of unix hackers using their initials.

[Brandon Ballinger @bballinger]

@nbt Good modern example of the insight that a CS PhD can get you. Lots of non-obvious big bets.

[Brandon Ballinger @bballinger]

@FrancoisChauba1 That's awful -- I'm so sorry for your loss.

[Brandon Ballinger @bballinger]

FDA is reviewing Grails @GalleriTest test Wednesday. Its accuracy is highest for cancers with no current screening guidelines: