Medical Boulevard

Call Directions Website Contact the business

Medical Boulevard We provide adaptable medical spaces. Our offerings include exam, spa, conference, and media rooms

10/27/2026

Read It Back Ep. 26: Who really saves when medicine uses AI?

Who really saves when medicine uses AI? Three posts from Eric Topol, JAMA, and NEJM, read word for word, then explained.

Posts & sources:
• Eric Topol (): https://x.com/EricTopol/status/2114716553921118545
Source: https://www.nature.com/articles/s44360-026-00205-y
• JAMA (): https://x.com/JAMA_current/status/2114750941513572672
Source: https://jamanetwork.com/journals/jama/fullarticle/10.1001/jama.2026.21230
• NEJM (): https://x.com/NEJM/status/2114824423752548852
Source: https://www.nejm.org/doi/10.1056/NEJMp2605590

Follow Medical Boulevard for a new read-back every day.
Discuss it with us on X:
For education, not medical advice. Every post is quoted as published.

This automated news aggregator is powered by cortxt technologies, in cooperation with Medical Boulevard. Build an automated news aggregator for you: https://cortxt.com/?utm_source=readitback&utm_medium=video&utm_campaign=news-aggregator

FULL TRANSCRIPT

HOST: [curious] Who really saves when medicine uses AI? This is Read It Back, from Medical Boulevard, with posts from Eric Topol, JAMA, and the New England Journal of Medicine. The money story depends on what happens to the PATIENT. Let's read it back.

HOST: [thoughtful] Start with a deceptively simple choice… which product name does a chatbot suggest? Eric Topol, the cardiologist at Scripps Research, flagged a new Nature Health study that puts a PRICE on that question. On October twenty-sixth, twenty twenty-six, Topol posted: Drugs recommended by L L Ms differed from general medical practice and often were for higher cost alternatives. There are links in this post. You'll find them in the caption.

READ-BACK: A drug's active ingredient and its product name are different things. Here, researchers supplied the disease and the ingredient, then asked five language models which products to name. Across synthetic cases, the answers kept clustering around the SAME products… and their projected pharmacy acquisition costs often exceeded those suggested by actual prescribing patterns. [serious] The fine print: these were simulated recommendations, not patient purchases. Pharmacy substitution could change what someone actually receives. Manufacturers and payers have a stake in product choice — but this study doesn't establish PAID influence. [emphatic] So, the read-back: agreement between chatbots isn't a value judgment. A repeated answer still needs checking for appropriate options and eventual cost.

HOST: [curious] Now follow that trail to the payment decision. AI can help review whether a service meets coverage rules. JAMA, the medical journal, shared new research on Medicare's Wiser model. On October twenty-sixth, twenty twenty-six, JAMA posted: Following implementation of the Wasteful and Inappropriate Service Reduction (Wiser) model by the Centers for Medicare and Medicaid Services, traditional Medicare use of Wiser-targeted services declined in participating states versus other states during the first six months. The reduction was seen for prior authorization targets including epidural steroid injections and other selected services, while delayed-implementation services did not change, informing ongoing evaluation of the model’s effects on utilization. There are links in this post. You'll find them in the caption.

READ-BACK: Prior authorization means checking coverage BEFORE a service is delivered. Wiser combines technology with clinical review

10/26/2026

Read It Back Ep. 25: Who turns an AI finding into care?

Who turns an AI finding into care? Three posts from NEJM AI, Francis Deng, MD, and Dutch Rojas, read word for word, then explained.

Posts & sources:
• NEJM AI (): https://x.com/NEJM_AI/status/2113634308455895545
Source: https://ai.nejm.org/doi/full/10.1056/AIoa2600258
• Francis Deng, MD (): https://x.com/francisdeng/status/2114384324099280939
Source: https://digitalassets.jointcommission.org/api/public/content/77c6421712dc463eae4b2398f9bf01a5?v=e9bc1e0f
• Dutch Rojas (): https://x.com/DutchRojas/status/2114532967620005960
Source: https://www.healthcare.gov/glossary/deductible/

Follow Medical Boulevard for a new read-back every day.
Discuss it with us on X:
For education, not medical advice. Every post is quoted as published.

This automated news aggregator is powered by cortxt technologies, in cooperation with Medical Boulevard. Build an automated news aggregator for you: https://cortxt.com/?utm_source=readitback&utm_medium=video&utm_campaign=news-aggregator

FULL TRANSCRIPT

HOST: [warmly] This is Read It Back, from Medical Boulevard. Who turns an AI finding into care? Today, posts from the journal N E J M A I, Francis Deng, and Dutch Rojas take us from a detection tool to the person waiting for an explanation… and then the bill.

HOST: [curious] Start with listening to the heart. A promising signal matters to a doctor only if it leads somewhere USEFUL. N E J M A I, the medical journal from N E J M Group, highlighted research on digital auscultation on October twenty-third, twenty twenty-six. The journal posted: A new study develops and externally validates an AI-enhanced digital auscultation system for detecting left ventricular dysfunction using prospective multicenter data. Learn more: There are links in this post. You'll find them in the caption.

READ-BACK: Auscultation means listening to the body, usually through a stethoscope. Here, the journal describes software helping detect left ventricular dysfunction — a problem with how a heart chamber works. This is a research announcement about developing and checking a detection system. External validation means testing beyond the data used to develop it. Multicenter means the data come from more than one center. [serious] The fine print: detecting a possible problem and improving someone's care are DIFFERENT questions. The journal is sharing research it publishes

10/24/2026

Read It Back Ep. 24: What makes medical AI trustworthy?

What makes medical AI trustworthy? Three posts from Greg Brockman, Eric Horvitz, and Bob Wachter, read word for word, then explained.

Posts & sources:
• Greg Brockman (): https://x.com/gdb/status/2113724268773023917
Source: https://openai.com/index/securing-our-frontier-model-training-environment/
• Eric Horvitz (): https://x.com/erichorvitz/status/2113342594633679319
Source: https://microsoft.ai/news/building-medical-superintelligence/
• Bob Wachter (): https://x.com/Bob_Wachter/status/2112952497329578046
Source: https://github.com/dviraran/mtsample_study

Follow Medical Boulevard for a new read-back every day.
Discuss it with us on X:
For education, not medical advice. Every post is quoted as published.

This automated news aggregator is powered by cortxt technologies, in cooperation with Medical Boulevard. Build an automated news aggregator for you: https://cortxt.com/?utm_source=readitback&utm_medium=video&utm_campaign=news-aggregator

FULL TRANSCRIPT

HOST: [warmly] This is Read It Back, from Medical Boulevard. What makes medical AI trustworthy? Greg Brockman, Eric Horvitz, and Bob Wachter take us from security… to diagnosis… to the bill. It's a story about EVERYTHING built around the model.

HOST: [curious] Start with the boundaries. What keeps AI's actions inside the rules? Greg Brockman, OpenAI's president and co-founder, shared his company's security report on October twenty-third, twenty twenty-six. Brockman posted: A post on how we've been securing our frontier model training environment. We've been building tools around our frontier cyber models, which have secured us far beyond what we could have unassisted and should generalize as key to how defenders secure themselves going forward.

READ-BACK: An AI agent uses tools and takes actions. A sandbox limits where those actions can go. OpenAI reports using agents to probe its defenses, finding weaknesses and repairing them. Think of checking the locks… by actually TRYING the doors. [serious] The fine print: these were deliberately adversarial tests of research environments, which differ from released products. Brockman leads the company making the claims. OpenAI also expects future models to find weaknesses it has NOT discovered yet. [emphatic] So, the read-back: trust needs a working boundary, backed by continuing tests. For medicine, the question is what has been checked… BEYOND the quality of the answer.

HOST: [thoughtful] Now, from boundaries to reasoning. Eric Horvitz, Microsoft's chief scientific officer and a physician, shared research on diagnosis, step by step. On October twenty-second, twenty twenty-six, Horvitz posted: On today’s cover of Science: Our work on AI systems that reason through diagnosis, step by step. Grateful to an incredible team of collaborators. Sharing code toward achieving more accurate and more timely diagnoses worldwide. Harsha Nori, Mayank Daswani, Scott Lundberg, Marco Tulio Ribeiro, Marc Wilson, Xiao Liu, Viknesh Sounderajah, Jonathan Carlson, Matthew Lungren, Bay Gross, Peter Hames, Mustafa Suleyman, Dominic King, Microsoft, Satya Nadella, Kevin Scott, Eric Topol, NASEM Health, Mayo Clinic, National Academy of Medicine There are links in this post. You'll find them in the caption.

READ-BACK: Sequential diagnosis builds the picture one question or test at a time. Microsoft's system coordinates a virtual panel that proposes diagnoses and debates what information to request. The research compares accuracy and estimated testing costs in simulations. It evaluates the ROUTE to the answer, as well as the answer itself. [serious] The fine print: the authors' earlier manuscript says these simulations exclude healthy or benign encounters… and do not measure patient outcomes. Horvitz is a coauthor and a Microsoft executive. The company's new report describes a prospective hospital study as the next step. [emphatic] So, the read-back: a stronger diagnostic process deserves attention. It still has to EARN trust in care. Security and reasoning checks answer different questions

10/23/2026

Read It Back Ep. 23: Does AI's test score travel to the clinic?

Does AI's test score travel to the clinic? Three posts from Mustafa Suleyman, Berci MeskĂł, MD, PhD, and Adam Rodman, read word for word, then explained.

Posts & sources:
• Mustafa Suleyman (): https://x.com/mustafasuleyman/status/2113332931423039945
Source: https://microsoft.ai/news/building-medical-superintelligence/
• Berci Meskó, MD, PhD (): https://x.com/Berci/status/2113165757773226319
Source: https://www.nature.com/articles/s41746-026-03418-3
• Adam Rodman (): https://x.com/AdamRodmanMD/status/2113376694446817609
Source: https://bilh.org/about/news-and-stories/all-news-and-stories/news/2026/10/bilh-announces-center-surgical-medical-intelligence

Follow Medical Boulevard for a new read-back every day.
Discuss it with us on X:
For education, not medical advice. Every post is quoted as published.

This automated news aggregator is powered by cortxt technologies, in cooperation with Medical Boulevard. Build an automated news aggregator for you: https://cortxt.com/?utm_source=readitback&utm_medium=video&utm_campaign=news-aggregator

FULL TRANSCRIPT

HOST: [curious] Does AI's test score… TRAVEL to the clinic? [short pause] This is Read It Back, from Medical Boulevard — with Mustafa Suleyman, Bertsi Meshko, and Adam Rodman.

HOST: [thoughtful] First, a system that asks for more information before naming a diagnosis. Mustafa Suleyman, chief executive of Microsoft AI, shared new research on October twenty-second, twenty twenty-six. For doctors, the question is WHERE that performance was measured. Suleyman posted: Our M A I Diagnostic system is now one point seven times more accurate than physicians in the E D, even when they have access to the web to look things up. It's also lower cost and does fewer medical tests to get to its diagnosis. Published in the journal Science today...

READ-BACK: Sequential diagnosis means building the picture one question or test at a time. Microsoft's system coordinates several AI roles to do that. In its report on the Science study, Microsoft says the system reached eighty percent accuracy on emergency-department cases, against forty-seven percent for physicians. [short pause] But these were interactive cases built from records… not patients being treated in a LIVE emergency room. [serious] The fine print: test costs were estimated. Suleyman leads the company developing the system, and Microsoft says a prospective hospital study is ahead. The company has a stake… in the promise. [emphatic] So, the read-back: this is a promising test of diagnostic reasoning. It does not YET tell us how much better patients will do.

HOST: [curious] Now… the gap between a test and deployment. Physician Bertalan Meskó, director of The Medical Futurist Institute, highlighted research on different tools: AI that checks scans for clots. On October twenty-second, twenty twenty-six, Meskó posted: There is a worrying mismatch between what results the F D A allows an AI model to enter the market with and what results it starts delivering after being deployed. A study just confirmed that. Authors retrospectively evaluated two F D A cleared AI algorithms from a single commercial platform (Aidoc Medical BriefCase) in pulmonary embolism triage detection on routine contrast-enhanced C Ts, across a seventeen-facility academic health system. "Both models demonstrated lower sensitivity than F D A clearance benchmarks while exceeding cleared specificity, with diminishing performance for peripheral and non-acute emboli mirroring known human reader limitations and underscoring the need for standardized post-market surveillance of AI-enabled medical devices." I wonder how the F D A or Aidoc responds to such claims. The study: There are links in this post. You'll find them in the caption.

READ-BACK: A pulmonary embolism is a clot in the lung's blood vessels. This peer-reviewed study looked backward at two Aidoc tools: one for dedicated lung-vessel scans, another for incidental clots on routine scans. They caught fewer positive cases than the clearance comparison figures suggested… while doing better at ruling out negative cases. This is NOT a test of Microsoft's system. [serious] The fine print: the authors could not confirm identical software versions, and their reference was clinical reports already exposed to AI. Aidoc supplied predictions but, the authors say, did not design or analyze the study. [thoughtful] So, the read-back: the comparison is imperfect… but the lesson is practical. A clearance figure cannot promise identical performance everywhere. The missed cases in the actual setting MATTER.

HOST: [thoughtful] So who builds the evidence between promise and practice? Adam Rodman, a hospitalist at Beth Israel Deaconess Medical Center, is helping lead a new research center. On October twenty-second, twenty twenty-six, Rodman posted: I can't believe it's really happening! For the last year, Gabriel Brat and myself have been putting together a brand new AI Center at Beth Israel Lahey Health that we hope will serve as a catalyst across the entire system for advancing the responsible study and adoption of AI in medicine.

READ-BACK: This is an institutional announcement. Beth Israel Lahey Health is bringing medical and surgical AI research together. The center plans to connect clinical data, and draw on expertise in trials, health policy, and evaluation in everyday care. [thoughtful] Its stated questions include safety, effectiveness… and IMPACT. [serious] The fine print: Rodman is announcing an effort he helps lead. The center's opening is not evidence that any particular AI product improves care. [emphatic] So, the read-back: enthusiasm needs a place to be TESTED. Our standard is evidence about what works, for whom, and in which setting… followed by monitoring once a tool is in use.

HOST: [warmly] Strong AI test results need checking in routine care, with continued monitoring after deployment. That's today's lesson, and that's Read It Back. Follow Medical Boulevard so you don't miss tomorrow's read-back. Every post is quoted as published, with links in the description. This show is for education, not medical advice.

HOST: [warmly] This automated news aggregator is powered by cor-text technologies, in cooperation with Medical Boulevard. [curious] Want one for your own field? Click the link in the description, and build an automated news aggregator for you.

medicalblvd.com

10/22/2026

Read It Back Ep. 22: What makes medical AI trustworthy?

What makes medical AI trustworthy? Three posts from Gary Marcus, Jonathan H Chen MD PhD, and Google Research, read word for word, then explained.

Posts & sources:
• Gary Marcus (): https://x.com/GaryMarcus/status/2113253157610693047
Source:https://nvlpubs.nist.gov/nistpubs/ai/NIST.AI.100-1.pdf
• Jonathan H Chen MD PhD (): https://x.com/jonc101x/status/2112951537534218515
Source: https://www.nature.com/articles/s41586-026-10675-5
• Google Research (): https://x.com/GoogleResearch/status/2112943733511794862
Source: https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2851451

Follow Medical Boulevard for a new read-back every day.
Discuss it with us on X:
For education, not medical advice. Every post is quoted as published.

This automated news aggregator is powered by cortxt technologies, in cooperation with Medical Boulevard. Build an automated news aggregator for you: https://cortxt.com/?utm_source=readitback&utm_medium=video&utm_campaign=news-aggregator

FULL TRANSCRIPT

HOST: [warmly] This is Read It Back, from Medical Boulevard. Today's question… what makes medical AI trustworthy? Gary Marcus, Jonathan Chen, and Google Research take us from an accountability warning to the difference between a convincing simulation and a clinical test. Let's read them back.

HOST: [thoughtful] Start outside the hospital. When software acts on our behalf… who answers for mistakes? Doctors know that question. Gary Marcus, emeritus psychology professor at New York University, raised it on October twenty-second, twenty twenty-six. [short pause] Marcus posted: Breaking news: U S. banks are taking hits from AI agents. Predictions: one. No matter how much chaos generative AI causes, the AI industry will not take real responsibility. two. The Trump administration will not hold them responsible

READ-BACK: An AI agent uses tools to carry out tasks. Explaining a form… and submitting it… are DIFFERENT responsibilities. Marcus's industry and political predictions are opinions — not established outcomes. He is also an AI entrepreneur who has founded companies in this field. [serious] There is a more concrete standard available. The National Institute of Standards and Technology's voluntary AI risk framework calls for clear responsibilities, human oversight, and leadership ownership of deployment risks. It is guidance — not a guarantee that any particular system is safe. [emphatic] So, the read-back: being able to do a task does NOT settle who may authorize it, or who answers for mistakes. For medicine, that is the starting question. Next comes the evidence for the task itself.

HOST: [curious] Now take that question into a simulated clinic. Jonathan Chen, a Stanford physician-scientist, shared a research talk about medical agents on October twenty-first, twenty twenty-six. The paper he names tests a system called Mira. Here is what Chen posted: One of ten. Stanford Division of Computational Medicine Colloquium: "AI in Oncology" with Jakob Nikolas Kather, senior author on the recent Nature paper "Towards autonomous medical artificial intelligence agents."

READ-BACK: Mira works inside a simulated electronic health record. It gathers histories, requests tests, and proposes treatment plans. The Nature study reports promising performance… but the patients it talks to are SIMULATIONS, grounded in existing records. Picture a flight simulator: a valuable place to test decisions, with a very specific boundary around what the result means. [serious] The fine print: simulated answers can be tidier than real patient conversations, and the authors say prospective clinical validation remains essential. Senior author Jakob Nikolas Kather discloses biotech shareholdings and consulting relationships. [emphatic] So, the read-back: this is evidence worth pursuing. It is NOT permission to hand over a clinic. A result earns its meaning from the conditions of the test. The next question… is what happens when those conditions change.

HOST: [thoughtful] That brings us to a narrower task, tested in new clinical settings: estimating how far along a pregnancy is. On October twenty-first, twenty twenty-six, Google Research, Google's research organization, shared a study of AI-assisted ultrasound. Google Research posted: Our new research shows how blind sweep ultrasound AI generalizes to new care settings, enabling non-sonographer operators with low-cost probes to deliver clinical-grade gestational age assessments, expanding life-saving maternal care globally. Read the blog: There are links in this post. You'll find them in the caption.

READ-BACK: A blind sweep means moving an ultrasound probe across the abdomen in a standard pattern, without having to find each precise anatomical view. AI then estimates gestational age: how far along the pregnancy is. This prospective diagnostic study in JAMA Network Open tested the system in Chicago and Nairobi. The model was adapted using Chicago data

10/19/2026

Read It Back Ep. 21: Who controls your health information?

Who controls your health information? Three posts from Mark Cuban, Eric Topol, and Wendell Potter, read word for word, then explained.

Posts & sources:
• Mark Cuban (): https://x.com/mcuban/status/2111945635444253143
Source: https://support.claude.com/en/articles/12123587-import-and-export-your-memory-from-claude
• Eric Topol (): https://x.com/EricTopol/status/2111503231398609164
Source: https://www.researchprotocols.org/2025/1/e77204
• Wendell Potter (): https://x.com/wendellpotter/status/2111105221996826875
Source: https://sunlightreportinsurance.com/subsidiaries/rvo-health

Follow Medical Boulevard for a new read-back every day.
Discuss it with us on X:
For education, not medical advice. Every post is quoted as published.

This automated news aggregator is powered by cortxt technologies, in cooperation with Medical Boulevard. Build an automated news aggregator for you: https://cortxt.com/?utm_source=readitback&utm_medium=video&utm_campaign=news-aggregator

FULL TRANSCRIPT

HOST: [warmly] This is Read It Back, from Medical Boulevard. Who controls your health information? Today, Mark Cuban, Eric Topol, and Wendell Potter take us from AI memory to the doctor's office… then behind the websites patients read. [curious] Making information easier to use is only the BEGINNING.

HOST: [thoughtful] Start with what happens when someone wants to change AI services. Can their information come along? For medicine, that raises a question about PATIENT control. Mark Cuban, investor and Cost Plus Drugs co-founder, raised the consumer version on October eighteenth, twenty twenty-six. Cuban posted: Right or Wrong ? Switching costs for consumer AI, won't approach zero until there is a standardized export format for what is in a model's memory. The more a consumer uses a model, the more it's memory stores. The more work to export and import I'm not saying it can't be done. Obviously you can import from one model to another. But for regular users, that's a lot more difficult than saving to an X L S or doc file Curious what people think ?

READ-BACK: AI memory means saved context that can carry into later conversations. Think of preferences… or details someone has already explained. Cuban's point is about the EFFORT of moving that context. It is a business opinion, not a study measuring how many users are stuck. Transfer tools already exist: Claude's documentation describes importing and exporting memory. [serious] The fine print: Anthropic calls imports experimental… and says some memories may not transfer successfully. Anthropic sells the service

10/17/2026

Will you look worse once you stop toxin?

Dr. Sameer Bashey, MD FAAD answers the frozen and irreversible myths in the myths module of the Neurotoxins for Dentists Master Class: the goal is partial paralysis, and toxin lasts two to four months depending on product.

100% online · self-paced · 4 CE credits. Founding price: sold out. Join the list for the next promotion.

Enroll: medicalblvd.com/go/mb-fb/reversible

For licensed dental professionals. Scope of practice varies by state. Many dental uses of neurotoxin are off-label.

Medical Boulevard Precision Academy is an ADA CERP Recognized Provider. ADA CERP is a service of the American Dental Association to assist dental professionals in identifying quality providers of continuing dental education. ADA CERP does not approve or endorse individual courses or instructors, nor does it imply acceptance of credit hours by boards of dentistry. Medical Boulevard Precision Academy designates this activity for 4 continuing education credits.

10/17/2026

Read It Back Ep. 20: Does better AI deserve more authority?

Does better AI deserve more authority? Three posts from James Zou, JAMA, and Arvind Narayanan, read word for word, then explained.

Posts & sources:
• James Zou (): https://x.com/james_y_zou/status/2110428682439598486
Source: https://arxiv.org/abs/2610.18861
• JAMA (): https://x.com/JAMA_current/status/2110761334119797105
Source: https://jamanetwork.com/journals/jama/fullarticle/2855491
• Arvind Narayanan (): https://x.com/random_walker/status/2111065613237915681
Source: https://www.normaltech.ai/p/steering-ai-beyond-automation

Follow Medical Boulevard for a new read-back every day.
Discuss it with us on X:
For education, not medical advice. Every post is quoted as published.

This automated news aggregator is powered by cortxt technologies, in cooperation with Medical Boulevard. Build an automated news aggregator for you: https://cortxt.com/?utm_source=readitback&utm_medium=video&utm_campaign=news-aggregator

FULL TRANSCRIPT

HOST: [curious] Does better AI deserve more authority? This is Read It Back, from Medical Boulevard. With James Zou, JAMA, and Arvind Narayanan. [warmly] Let's read it back.

HOST: [curious] First, the engine. A faster AI assistant sounds useful in a busy clinic. But what has ACTUALLY been tested? James Zou, Stanford's associate professor of biomedical data science, shared his team's Hybrid-Agent research on October fourteenth, twenty twenty-six. Zou posted: Excited to introduce Hybrid-Agent! New architecture that reasons in latent space and acts through text. Smarter, two times faster and three times cheaper than standard L L Ms, with fully readable latent thoughts! We also open-source the R L and S F T pipelines for training hybrid models. Great job Jiaru Zou with awesome collaborators at Together AI and University of Illinois Urbana-Champaign

READ-BACK: An AI agent is software that can use tools while working through a task. Hybrid-Agent switches between internal numerical representations… and ordinary text for actions. The researchers also describe a way to inspect those internal steps as readable text. [short pause] This is a research preprint, with benchmark tests, including medical questions. [serious] The fine print: the cost savings are estimates that exclude external tool ex*****on. Zou is a coauthor, with a Together AI affiliation. These are the developers' results… NOT evidence of improved patient outcomes. [emphatic] So, the read-back: efficiency makes an assistant more practical. It doesn't establish what clinical responsibility that assistant can SAFELY carry. [short pause] That needs its own evidence.

HOST: [thoughtful] That's technical performance. Now… what are clinicians actually asking AI to do? JAMA, the peer-reviewed medical journal, shared a study of Open Evidence use on October fifteenth, twenty twenty-six. It puts TREATMENT questions at the center of the story. JAMA posted: Among U S clinicians, use of an AI-enabled clinical information platform increased substantially over two point five years, with sustained use across physician and nonphysician groups. Queries most often focused on treatment and management rather than diagnosis, indicating that these tools are being used primarily to support clinical decision-making at the point of care. There are links in this post. You'll find them in the caption.

READ-BACK: Open Evidence is an AI tool for looking up clinical information. This observational study follows questions people submitted, rather than assigning patients to different treatments. Treatment and management featured more often than diagnosis. [short pause] That matters: the tests we use to judge medical AI need to reflect the work clinicians ACTUALLY give it. [serious] The fine print: this is ONE platform. Researchers couldn't tell whether questions arose during patient visits, preparation, or other settings. Three coauthors worked for Open Evidence, which supplied data and contributed to analysis. The study does not establish better patient outcomes. [emphatic] So, the read-back: adoption tells us where to investigate. It doesn't CERTIFY the answers.

HOST: [serious] So we have a faster engine and widespread use… but still a question about authority. Arvind Narayanan, a Princeton computer science professor, argues that ability ALONE cannot settle what we hand over. On October sixteenth, twenty twenty-six, Narayanan posted: We need a mindset shift in how we deploy AI. In the past, new capabilities unlocked new use cases. Safety, control, ethics, and skill erosion were important but usually secondary considerations

Address

9025 Wilshire Boulevard
Beverly Hills, CA
90211

Opening Hours

Monday 9am - 7pm
Tuesday 9am - 7pm
Wednesday 8am - 7pm
Thursday 8am - 7pm
Friday 9am - 5pm
Saturday 9am - 3pm

Telephone

+13108002262

Alerts

Be the first to know and let us send you an email when Medical Boulevard posts news and promotions. Your email address will not be used for any other purpose, and you can unsubscribe at any time.

Contact The Business

Send a message to Medical Boulevard:

Shortcuts

Share

Category

Nearby realtors & realty services