Raheel Sayeed

Physician • biomedical informatics

Apple ex: @HarvardDBMI @bos_chip; RT ≠ endorse

[Raheel Sayeed @rsayeed]

Aah, My school followed CBSE, the most strenuous classic Indian STEM curriculum.

Results were announced 10a online, got HTTP 503 until evening.

[Deedy @deedydas]

The national high school exam of India, CBSE, has been Pwned!

This incompetent organization continues to deny the allegations against them. And a teenager has taken over their prod servers hosting the exam booklet scans of 2M test takers. They have just taken it down.

All they had to say is "can you help us fix the problem?" but their ego is too big to admit they were wrong.

Incompetence is one thing. The complete lack of accountability to the nation while your servers get catastrophically owned is another. Internet-scale embarrassment.

[Raheel Sayeed @rsayeed]

Doctor Gabe!

[Josh Farkas MD 💊 @PulmCrit]

careful: if this technology falls into the hands of internal medicine docs, rounds may extend to 8 hours long

[Raheel Sayeed @rsayeed]

Yes, let’s be explicit. Because implicit didn’t hurt enough.

[Tanishq Mathew Abraham, Ph.D. @iScienceLuvr]

People freaking out about this are acting like hedge funds and other quant firms aren't already implicitly betting on clinical trials

Honestly it's a messed up system, but perhaps it is better to be open about it than closed.

[Raheel Sayeed @rsayeed]

Let’s open betting stations outside of hospitals.

Absolute lunacy.

[Michael @endpointarena]

Quick run down of the features on Endpoint Arena, the prediction market for clinical trials.

the home page has all the trials:

[Raheel Sayeed @rsayeed]

😤

[Brittany Trang @brittanytrang]

Health insurers have claimed for months that AI is driving up health care costs. Now BCBS is offering proof.

The ambiguous analysis methods would sow more doubt on the conclusion...if a hospital hadn't told me last year that its AI coding software was targeting this EXACT code.

[Raheel Sayeed @rsayeed]

Critical thinking replaced by "Whats your compensation
?" 🤮

[Alex Imas @alexolegimas]

You should always read Tom, and today is no exception.

[Raheel Sayeed @rsayeed]

@DrShoaibAM Frankly, I'm not very convinced of any paper that predicts "critical thinking" in either direction. It is highly contextual, depends on the task/activity, and ofcourse the LLM being used. How you'd design the study would highly predict direction.

[Raheel Sayeed @rsayeed]

I now have a new term, inspired from AI Slop: Academic Slop!

[Raheel Sayeed @rsayeed]

@MaziyarPanahi Is the reasoning verified?

[Raheel Sayeed @rsayeed]

Absolute clown talk.

[Nachiket Bhatia @bhatia_nachiket]

Look at global salary data and one thing is obvious:
Developed nations value doctors. India consumes them.

Here, doctors are expected to:
•Work longer
•Earn less
•Stay silent
•And call it “nobility”

Abroad, the same doctor is called a high-value professional.

The gap isn’t competence.
The gap is respect—and regulation.

#DoctorsOfIndia
#IndianDoctors
#DoctorLivesMatter
#RespectDoctors
#ValueDoctors

#UnderpaidDoctors
#PayDoctorsFairly
#DoctorExploitation
#HealthcareCrisis
#BrokenHealthcare

[Raheel Sayeed @rsayeed]

@theo Yep!

[Raheel Sayeed @rsayeed]

Speaking of craft: in school I saw “Code is Poetry” stamped all over @photomatt’s PHP inside @WordPress. 🤯
Later came the real baptism: @Linus__Torvalds’ mailing lists. 😂

Shaped my software world view!

[Alex Lieberman @businessbarista]

11 observations interviewing 500+ engineers the last 9 months:

1) The new H1B visa thing is a big bummer. We have to say no to lots of qualified candidates because it's financially irresponsible for us to spend $100,000 extra to hire someone.

2) There are very few true full-stack engineers out there. Most applicants are 90% backend or 90% devops or 90% data eng with a splash of full-stack. Being a true five-tool hitter has never been more important.

3) The spread in compensation ranges is wild. I'll have one call with a Series C senior engineer making $600k/yr, then I'll have a second call with a SWEII at Microsoft making $175k/yr. Comp is all over the map and a moving target.

4) The biggest risk with big tech engineers is lack of experience building systems end-to-end. Technical depth may be really strong, but ability to own & system is iffy.

5) There are very few AI-native engineers. An AI-native engineer is someone that has completely rearchitected their personal workflows for building software. To be AI-native, your process should be >75% different than pre-AI.

6) There are very few AI engineers. An AI engineer is someone that has built and deployed production agentic systems (ideally at scale). Because the technology is so new, you either have to find a diamond in the rough or hire someone you think can learn fast.

7) Strong communication is rare as hell & when you find it, it's glorious. All of our engineers at @tenex_labs are client-facing, but outside of that strong communicators are usually strong thinkers and strong thinking has never been more important in engineering.

8) Testing an engineer's technical chops has never been harder. When everything is Claude Code-able, quickly understanding an engineers fundamental systems & engineering abilities is very hard. Someone will make a billion dollars solving this problem.

9) Great engineers under-estimate themselves. Mediocre engineers over-estimate themselves. I always ask the question: "If you had to rank yourself 0-10 as an engineer, what number would you pick & why?" The best engineers I've hired give themselves a 7.5/10 max.

10) Engineers are way more interested in being influencers/creators than I ever imagined. Part of the reason they're interested in joining us is they want to be the next @theo or @ThePrimeagen.

11) There are two types of engineers: those who view it as a job and those who view it as a craft. Those who view it as a craft are typically more entrepreneurial, are always cooking up a weekend side project, and view engineering as a core part of their identity.

[Raheel Sayeed @rsayeed]

I shouldn't be surpised at this but, holy... 😱!

[baz @baz_scm]

Does your coding agent slap you with screenshot-evidence of broken implementation? Mine does...

[Rafael Torres ᯅ @mundoxrbrasil]

Apple isn’t “correcting course” on AI. In reality, it didn’t burn money on something that was bound to become a commodity. It’s much easier — and far less risky — to adopt one of these models for its users than to compete in a spending race over who invests more and who will take the biggest losses once that becomes obvious.

In this context, Google is the ideal partner. It already pays Apple billions to remain the default search engine on Apple devices, and now it will get a return of a few more billions to power the new Siri — without Apple funneling money into OpenAI and strengthening what could become its next major product competitor.

Apple’s investors are less exposed to AI “investment risk” compared to other companies, while Apple still manages to solve a real problem for its users in the consumer-focused AI race. You could even say that Apple’s slowness in some areas ends up helping sometimes 😅

[Raheel Sayeed @rsayeed]

Holy, Im back on MS Windows after 19 years. Apparently its all AI now?

[Joel M. Topf, MD FACP @kidney_boy]

AI slop, so many little mistakes

[Raheel Sayeed @rsayeed]

Why not the same for health+medicine papers?

[Open Review @openreviewnet]

A Message from AI Research Leaders: Join Us in Supporting OpenReview
https://t.co/U1Co2d59do

[swyx @swyx]

This was complete slop and I'm tired of everybody pretending it was some genius work

I'm in 🇧🇷 today soft launching my talk declaring War on Slop . hope to develop it fully in time for my @aidotengineer keynote