Syed

years building US healthcare software. I help physician founders and small clinics take AI built prototypes to production, HIPAA ready. building @WealEMR

[Syed @trilopsai]

Intake → Schedule → Chart → Rx → Bill → Telehealth.

One platform. Built for HRT clinics.

Plus AI Encounter: it drafts your notes while you focus on the patient.

Check out our flagship product at https://t.co/K40Et2kA3J

[WealMD @WealEMR]

WealMD now supports scheduling long running tasks, billing and AI Assistance.

https://t.co/BFUzwBemIU

[Syed @trilopsai]

@KrittanawongMD thanks, thats a useful distinction. diagnostic certainty is a real win for patients, it just deserves its own label when people quote the headline

[Syed @trilopsai]

@MattZirwas true, the hawthorne effect would ruin it lol. EHR audit logs could give u that baseline quietly tho, the data is already sitting there

[Syed @trilopsai]

@w_milczynska love this breakdown. only 3 scribes is a surprise given how crowded that market feels lol. would be cool to see how many of the 91 have a clinician on the founding team

[Syed @trilopsai]

@rorynotsorry the overhead people underestimate: licensing in every state u serve, malpractice, lab partnerships, and a HIPAA stack w BAAs for every vendor. the software is the cheap part now

[Syed @trilopsai]

@devyflow happy to connect. we work on the production side, getting AI built health apps ready for real clinics. access control, PHI flows, BAAs. whats ur model focused on?

[Syed @trilopsai]

@LiamEtherson huge unlock. next question clinic IT asks is the rest of the chain, where the transcript lands after and who touches it. zero retention on the speech side doesnt cover the EHR side

[Syed @trilopsai]

@pridgenmd To hear a professional approve this is mind-blowing, saw a post by an electrical engineer dissing a one shot app by Fable 5.2 for a complex circuit board saying it had so many it had so many technical issues

I still believe models are going to get nitty gritties right sooner

[Max Marchione @maxmarchione]

BREAKING: A new weight-loss drug that hits both the GLP-1 and glucagon receptors passed its heart safety test in a global Phase 3 trial.

In June, it showed up to 16.6% weight loss (vs. 3.2% on placebo) in its main obesity trial.

[Syed @trilopsai]

@dusutimothy @blockfuselabs love the problem. small tip from the health IT side, "100% private" gets questioned fast by hospital IT lol. saying exactly what u store, where, and who can see it builds more trust

[Syed @trilopsai]

@statnews a tool to check the chatbot is the right move. curious if CrossCover shows its own sources or just a confidence score, clinicians wont trust a second black box either

[Syed @trilopsai]

@morgancheatham @Roche @Path_AI @R1RCM @ouraring @CounselHealth the caution critique always missed the upside. clinicians know which tiny edge case ends up in a lawsuit, and thats exactly what gets a product through procurement

[Syed @trilopsai]

@DrIanWeissman tragic. consumer chatbots should at minimum stop and send people back to their prescriber whenever a med change comes up. thats a guardrail, not a hard engineering problem

[Syed @trilopsai]

before u paste a patient case into a chatbot, one question: does ur org have a BAA w that tool

no BAA means ur handing PHI to a vendor that never agreed to protect it. strip names, dates and MRNs, or use the tool ur hospital actually approved

[Syed @trilopsai]

@TheCancerSutra the make it into our healthcare system part is the bottleneck nobody funds. a model that works still needs an EHR hookup, a payer code and someone to maintain it for 10 yrs

[Syed @trilopsai]

AUROC 0.92 sounds amazing until u find out it fires 40 alerts a shift lol

when someone pitches u a clinical AI model ask for 3 things: how often it alerts, how often its right when it does (PPV), and what the team is supposed to do next

[Syed @trilopsai]

@QuanquanGu agree on the urgency. the part AI can compress fastest might be the boring middle, trial recruitment and data cleaning. a lot of that decade is waiting on paperwork, not science

[Syed @trilopsai]

@martinvars catching all 30 is impressive. how many false flags came w it? and who owns the follow up once the patient leaves the ED, thats usually where these programs break

[Syed @trilopsai]

@jefftangx the 92% yoy is the part people skip. oura can afford to buy care delivery now. open question is whether a ring company can run a medical group w all the licensing and compliance that comes w it

[Syed @trilopsai]

AI scribe tip nobody says out loud: the note it writes becomes ur note the second u sign it

If it hears "denies chest pain" as "chest pain" thats on ur signature, not the vendor. read the assessment and med list every time, thats where the expensive mistakes hide