What do you think?
Rate this book
ChatGPT, MD: How AI-Empowered Patients & Doctors Can Take Back Control of American Medicine
Doctors and patients have lost control of American medicine. Generative AI can put the power back in their hands, save millions of lives, and restore the doctor-patient relationship.Welcome to ChatGPT, MD. In this unique collaboration, renowned healthcare leader Dr. Robert Pearl teams up with the artificial intelligence system ChatGPT to examine the transformational power of generative AI in medicine. Together, the co-authors present a vision for the future where doctors, patients, and AI join forces to reclaim control of American healthcare from prevailing corporate interests. With engaging narratives and real-world examples, ChatGPT, MD reveals how advanced AI technologies can reduce medical errors, enhance diagnostic precision, ease the growing burden on healthcare professionals, and—most important—democratize medical expertise, arming patients with the kinds of tools and knowledge once reserved only for doctors. While the future looks incredibly promising, the journey will be fraught with challenges. In ChatGPT, MD, Dr. Pearl tackles tough questions concerning technological bias, patient privacy, and the threat of job displacement in an AI-driven healthcare system. An indispensable guide for healthcare professionals, patients, and anyone who’s disenchanted with the current healthcare system and invested in its future.
297 pages, Kindle Edition
Published April 9, 2024
87 people are currently reading
666 people want to read
About the author
Robert Pearl
3 books19 followersRobert Pearl, MD is a healthcare leader, author, educator, columnist and podcaster. For 18 years, he served as CEO of The Permanente Medical Group (Kaiser Permanente). He is also former president of The Mid-Atlantic Permanente Medical Group. In these roles he led 10,000 physicians, 38,000 staff and was responsible for the nationally recognized medical care of 5 million Kaiser Permanente members on the west and east coasts.
He is a clinical professor of plastic surgery at Stanford University School of Medicine and on the faculty at the Stanford Graduate School of Business, where he teaches courses on healthcare strategy, technology, and leadership. Pearl is board certified in plastic and reconstructive surgery, receiving his medical degree from Yale, followed by a residency in plastic and reconstructive surgery at Stanford University.
He's the author of three books: "Mistreated: Why We Think We're Getting Good Healthcare—And Why We're Usually Wrong," a Washington Post bestseller (2017); "Uncaring: How the Culture of Medicine Kills Doctors & Patients," a Kirkus star recipient (2021); and his newest book "ChatGPT, MD: How AI-Empowered Patients & Doctors Can Take Back Control of American Medicine" (April 2024). All profits from sales of his books go to Doctors Without Borders.
He is a LinkedIn "Top Voice" in healthcare and host of the popular podcasts Fixing Healthcare and Medicine: The Truth. He publishes two monthly healthcare newsletters reaching 50,000+ combined subscribers. A frequent keynote speaker, Pearl has presented at The World Healthcare Congress, the Commonwealth Club, TEDx, HLTH, NCQA Quality Talks, the National Primary Care Transformation Summit, American Society of Plastic Surgeons, and international conferences in Brazil, Australia, India, and beyond.
Dr. Pearl's insights on generative AI in healthcare have been featured in Associated Press, USA Today, MSN, FOX Business, Forbes, Fast Company, WIRED, Global News, Modern Healthcare, Medscape, Medpage Today, AI in Healthcare, Doximity, Becker's Hospital Review, the Advisory Board, the Journal of AHIMA, and more.
He is a clinical professor of plastic surgery at Stanford University School of Medicine and on the faculty at the Stanford Graduate School of Business, where he teaches courses on healthcare strategy, technology, and leadership. Pearl is board certified in plastic and reconstructive surgery, receiving his medical degree from Yale, followed by a residency in plastic and reconstructive surgery at Stanford University.
He's the author of three books: "Mistreated: Why We Think We're Getting Good Healthcare—And Why We're Usually Wrong," a Washington Post bestseller (2017); "Uncaring: How the Culture of Medicine Kills Doctors & Patients," a Kirkus star recipient (2021); and his newest book "ChatGPT, MD: How AI-Empowered Patients & Doctors Can Take Back Control of American Medicine" (April 2024). All profits from sales of his books go to Doctors Without Borders.
He is a LinkedIn "Top Voice" in healthcare and host of the popular podcasts Fixing Healthcare and Medicine: The Truth. He publishes two monthly healthcare newsletters reaching 50,000+ combined subscribers. A frequent keynote speaker, Pearl has presented at The World Healthcare Congress, the Commonwealth Club, TEDx, HLTH, NCQA Quality Talks, the National Primary Care Transformation Summit, American Society of Plastic Surgeons, and international conferences in Brazil, Australia, India, and beyond.
Dr. Pearl's insights on generative AI in healthcare have been featured in Associated Press, USA Today, MSN, FOX Business, Forbes, Fast Company, WIRED, Global News, Modern Healthcare, Medscape, Medpage Today, AI in Healthcare, Doximity, Becker's Hospital Review, the Advisory Board, the Journal of AHIMA, and more.
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 - 10 of 10 reviews
March 23, 2025
Book Review: ChatGPT M.D. by Robert Pearl, M.D.
Robert Pearl’s ChatGPT M.D., a 218-page exploration of artificial intelligence’s role in the future of medicine, promises a bold vision but ultimately stumbles under its own weight. What could have been a tight, impactful 100-page book is bogged down by relentless repetition. Pearl spends much of the text hyping ChatGPT’s potential—its ability to revolutionize diagnostics, streamline workflows, and enhance patient care—only to undercut his own enthusiasm with a contradictory refrain: AI won’t replace the “human touch” in medicine. This claim feels hollow to anyone who’s endured the modern primary care experience, where the “cattle call” atmosphere—rushed visits with PAs or nurses rather than M.D.s—already lacks the personal connection Pearl romanticizes. I’m paying for a doctor, not a conveyor belt, yet that’s rarely what I get.
The book chugs along with this push-pull dynamic—AI is amazing, but don’t worry, it’s not that amazing—until it takes a bizarre detour in the final pages of Chapter 12. Here, Pearl pivots to a head-scratching critique of Tim Cook and Apple, lambasting the Apple Watch and Vision Pro for not being medical-grade devices. He argues their consumer focus sidesteps liability as medical providers, a tangent that feels both petty and disconnected from the ChatGPT narrative. From there, the book plummets to a lackluster close, filled with redundancy after redundancy (pun intended).
At its best, ChatGPT M.D. offers intriguing glimpses into AI’s medical promise. At its worst, it’s a bloated, contradictory slog that loses focus when it matters most. Trim the fat, skip the Apple rant, and you might have something worth recommending. As it stands, it’s a missed opportunity.
Robert Pearl’s ChatGPT M.D., a 218-page exploration of artificial intelligence’s role in the future of medicine, promises a bold vision but ultimately stumbles under its own weight. What could have been a tight, impactful 100-page book is bogged down by relentless repetition. Pearl spends much of the text hyping ChatGPT’s potential—its ability to revolutionize diagnostics, streamline workflows, and enhance patient care—only to undercut his own enthusiasm with a contradictory refrain: AI won’t replace the “human touch” in medicine. This claim feels hollow to anyone who’s endured the modern primary care experience, where the “cattle call” atmosphere—rushed visits with PAs or nurses rather than M.D.s—already lacks the personal connection Pearl romanticizes. I’m paying for a doctor, not a conveyor belt, yet that’s rarely what I get.
The book chugs along with this push-pull dynamic—AI is amazing, but don’t worry, it’s not that amazing—until it takes a bizarre detour in the final pages of Chapter 12. Here, Pearl pivots to a head-scratching critique of Tim Cook and Apple, lambasting the Apple Watch and Vision Pro for not being medical-grade devices. He argues their consumer focus sidesteps liability as medical providers, a tangent that feels both petty and disconnected from the ChatGPT narrative. From there, the book plummets to a lackluster close, filled with redundancy after redundancy (pun intended).
At its best, ChatGPT M.D. offers intriguing glimpses into AI’s medical promise. At its worst, it’s a bloated, contradictory slog that loses focus when it matters most. Trim the fat, skip the Apple rant, and you might have something worth recommending. As it stands, it’s a missed opportunity.
November 17, 2025
FREE AI templates inspired by book at end of review.
This is a good book that offers an inspiring vision for change within the U.S. medical system. It's filled with excellent ideas and practical insights on leadership. The author provides a genuinely hopeful perspective. My only minor critique is the consistent use of the abbreviation "GenAI." When discussing the future, it leaves me wondering: Is the author referring to the current state of generative AI (like ChatGPT) or a more advanced, potentially general intelligence AI? A clarification would have been helpful for future concepts.
In one part of book, he says how narrow AI applications are tested and approved like medical devices, but this can't apply to how broad generative AI is. OpenEvidence already has solution to that, pass medical licensing test by 100%. Passed USMLE, next needs to pass internal med board.
Templates:
He said a simple list of ten most likely iatrogenic mistakes would have made a big difference in pediatric cancer case in book.
10 iatrogenic List (decaiatrongenic)
10 most likely iatrogenic errors for such a case:
[List ten most likely iatrogenic errors to be made in this kind of case. ]
Sources:
What medical sources, outside of the case information, did AI/LLM use to fill out this template?
What medical evidence, outside of case information, did AI/LLM use?
What medical journal articles, outside of case information, did AI/LLM use?
What consensus statements did AI/LLM use?
What websites did AI/LLM use?
Aİ can give erroneous information, use discretion.
[Use APA citation format to identify sources.]
Turnabout is fair play so here’s 10 AI mistakes ( mechagenic ) for a case:
10 most likely AI mechagenic errors for such a case ( decamechagenic ):
[List ten most likely AI errors to be made in this kind of case. ]
Sources:
What medical sources, outside of the case information, did AI/LLM use to fill out this template?
What medical evidence, outside of case information, did AI/LLM use?
What medical journal articles, outside of case information, did AI/LLM use?
What consensus statements did AI/LLM use?
What websites did AI/LLM use?
Aİ can give erroneous information, use discretion.
[Use APA citation format to identify sources.]
Dr. Pearl also mentioned medication crosschecking, so here’s a template for that:
Medication Crosscheck
Dosage check:
[Review case for drug dosing errors. ]
Case drug contraindications:
[Check case for drug contraindications. ]
Drug interactions:
[Check case for drug interaction problems. ]
What are the normal expected blood work lab changes from each drug?
Sources:
What medical sources, outside of the case information, did AI/LLM use to fill out this template?
What medical evidence, outside of case information, did AI/LLM use?
What medical journal articles, outside of case information, did AI/LLM use?
What consensus statements did AI/LLM use?
What websites did AI/LLM use?
Aİ can give erroneous information, use discretion.
[Use APA citation format to identify sources.]
Dr. Pearl also focused on medical education for using AI, here's an AI Anki template for students or anyone (change numbers to change total cards generated):
Your task is to generate exactly 14 lines of text for Anki import based on a lecture. Each line represents a single flashcard in the format `Front of Card;Back of Card;Tags`. All 14 cards should be generated within the single section below.
CRITICAL RULES:
1. The output MUST be exactly 14 lines.
2. Each line MUST be in the format: `Front;Back;Tags`
3. The Front and Back content MUST NOT contain any semicolons (`;`).
4. Do NOT use quotes (`"`) around any of the fields.
5. There should be no empty lines between cards.
[Generate 14 Anki flashcards from the lecture, one per line, following the format: Front;Back;Tags]
Generate exactly 14 Anki flashcards based on the lecture, with each card on a new line.
**CRITICAL FORMATTING RULES:**
1. **Exactly 14 Lines:** The final output must contain exactly 14 lines, no more, no less.
2. **No Empty Lines:** There must be no blank lines between the cards.
3. **Format:** Each line MUST follow the strict format: `Front of Card;Back of Card;Tags`
4. **No Semicolons in Content:** The 'Front of Card' and 'Back of Card' content MUST NOT contain any semicolons (`;`). Use commas or other punctuation instead.
5. **No Quotes:** Do NOT use quotes (`"`) around any of the fields.
**CARD CONTENT:**
* **Card 1 (Lecture Title):** The front of the card should be the title of the lecture. The back of the card should be the main topic or subject of the lecture.
* **Cards 2-14 (Study Questions):** Generate 13 distinct study questions based on the lecture content. The front of each card must be a question, and the back must be the corresponding concise answer. The questions should cover key concepts, definitions, and important takeaways from the lecture.
This is a good book that offers an inspiring vision for change within the U.S. medical system. It's filled with excellent ideas and practical insights on leadership. The author provides a genuinely hopeful perspective. My only minor critique is the consistent use of the abbreviation "GenAI." When discussing the future, it leaves me wondering: Is the author referring to the current state of generative AI (like ChatGPT) or a more advanced, potentially general intelligence AI? A clarification would have been helpful for future concepts.
In one part of book, he says how narrow AI applications are tested and approved like medical devices, but this can't apply to how broad generative AI is. OpenEvidence already has solution to that, pass medical licensing test by 100%. Passed USMLE, next needs to pass internal med board.
Templates:
He said a simple list of ten most likely iatrogenic mistakes would have made a big difference in pediatric cancer case in book.
10 iatrogenic List (decaiatrongenic)
10 most likely iatrogenic errors for such a case:
[List ten most likely iatrogenic errors to be made in this kind of case. ]
Sources:
What medical sources, outside of the case information, did AI/LLM use to fill out this template?
What medical evidence, outside of case information, did AI/LLM use?
What medical journal articles, outside of case information, did AI/LLM use?
What consensus statements did AI/LLM use?
What websites did AI/LLM use?
Aİ can give erroneous information, use discretion.
[Use APA citation format to identify sources.]
Turnabout is fair play so here’s 10 AI mistakes ( mechagenic ) for a case:
10 most likely AI mechagenic errors for such a case ( decamechagenic ):
[List ten most likely AI errors to be made in this kind of case. ]
Sources:
What medical sources, outside of the case information, did AI/LLM use to fill out this template?
What medical evidence, outside of case information, did AI/LLM use?
What medical journal articles, outside of case information, did AI/LLM use?
What consensus statements did AI/LLM use?
What websites did AI/LLM use?
Aİ can give erroneous information, use discretion.
[Use APA citation format to identify sources.]
Dr. Pearl also mentioned medication crosschecking, so here’s a template for that:
Medication Crosscheck
Dosage check:
[Review case for drug dosing errors. ]
Case drug contraindications:
[Check case for drug contraindications. ]
Drug interactions:
[Check case for drug interaction problems. ]
What are the normal expected blood work lab changes from each drug?
Sources:
What medical sources, outside of the case information, did AI/LLM use to fill out this template?
What medical evidence, outside of case information, did AI/LLM use?
What medical journal articles, outside of case information, did AI/LLM use?
What consensus statements did AI/LLM use?
What websites did AI/LLM use?
Aİ can give erroneous information, use discretion.
[Use APA citation format to identify sources.]
Dr. Pearl also focused on medical education for using AI, here's an AI Anki template for students or anyone (change numbers to change total cards generated):
Your task is to generate exactly 14 lines of text for Anki import based on a lecture. Each line represents a single flashcard in the format `Front of Card;Back of Card;Tags`. All 14 cards should be generated within the single section below.
CRITICAL RULES:
1. The output MUST be exactly 14 lines.
2. Each line MUST be in the format: `Front;Back;Tags`
3. The Front and Back content MUST NOT contain any semicolons (`;`).
4. Do NOT use quotes (`"`) around any of the fields.
5. There should be no empty lines between cards.
[Generate 14 Anki flashcards from the lecture, one per line, following the format: Front;Back;Tags]
Generate exactly 14 Anki flashcards based on the lecture, with each card on a new line.
**CRITICAL FORMATTING RULES:**
1. **Exactly 14 Lines:** The final output must contain exactly 14 lines, no more, no less.
2. **No Empty Lines:** There must be no blank lines between the cards.
3. **Format:** Each line MUST follow the strict format: `Front of Card;Back of Card;Tags`
4. **No Semicolons in Content:** The 'Front of Card' and 'Back of Card' content MUST NOT contain any semicolons (`;`). Use commas or other punctuation instead.
5. **No Quotes:** Do NOT use quotes (`"`) around any of the fields.
**CARD CONTENT:**
* **Card 1 (Lecture Title):** The front of the card should be the title of the lecture. The back of the card should be the main topic or subject of the lecture.
* **Cards 2-14 (Study Questions):** Generate 13 distinct study questions based on the lecture content. The front of each card must be a question, and the back must be the corresponding concise answer. The questions should cover key concepts, definitions, and important takeaways from the lecture.
June 22, 2024
Thank you to Dr Pearl for putting this book together which is no easy tasks. And here my opinions. On LinkedIn the author recommended that I read his book after one of my comments about the inability of AI to save healthcare. I read the book and it has the feel of a self-promotion book written by someone who wants to be a KOL on Gen AI. It's totally possible that Dr. Pearl wrote this purely selflessly and intends to inform the public - I'm only sharing my opinion.
The examples given in the book are about technology, spots, and car racing - so it's a heavily macho book. Remember, women are key decision makers in most households - addressing their perspective is, if nothing else, a good way to sell your book unless you're writing it for other reasons.
I'll start by saying that trying to predict the future has little value and the author states that legislature, payment systems, regulators, physicians, and patients all have to change for Gen AI to be an effective tool in healthcare. Maybe that could happen but it's never happened in the face of any previous tech.
He's talking about health 4.0 when we've regressed to previous models - so it's important to first understand how we got here before figuring out how we can get to somewhere else. As a previous CEO of a very wealthy HMO, Kaiser Permanente, I'm not sure how much he could really comprehend the day to day struggles of patients and physicians.
He reiterates the point many times that with more Gen AI we can free the doctor to have more time with the patient. What tech in the past ever allowed more time instead of just squeezing out more profits from the patient? And why would the doctor be of any value to spend more time with the patient if Gen AI is so capable, even being able to operate the robotic devices itself to perform the surgeries? A good editor and publisher can sometimes help tease out these plot holes.
Why would a patient take advice from ChatGPT if they aren't doing that already? How would their weight and BP be better if they are getting constant notifications when the same tactics right now are ineffective. Could it be that we have a food system that is designed to create disease to sell more food and to create illness which requires disease? Maybe not but it appears so.
Saying that ChatGPT will democratize health information is quite uninformed because health information is already democratized. We've been in the health information age since the 90s - 4 decades should be more than enough time for even the oldest authors to understand that the problem isn't knowledge - it's much more complex.
Complaining that hospitals are getting paid too little by Medicare and Medicaid can only come from the CEO of Kaiser Permanente. No physician, no patient, not even a politician is staying up late wondering how to get more profits to hospitals; we know they are already squeezing everyone dry, we're trying to figure out how to still keep them engaged without financializing more of medicine.
Comparing the Amazon system to healthcare is terrible because Amazon is terrible in every which way except for improving supply chains. The products we now use are far lower quality than ever before and the impact on the environment is likely not repairable in the next few decades. Nobody wants healthcare to be more like Amazon. The only part of Amazon we want is the great customer service.
Finally, the author mentions the ICU software that can predict which patient will crash. We don't need a software for that - nurses were able to do that better than doctors but now we are having fewer nurses, fewer doctors, and sicker patients just so the medical groups can profit more - especially groups like Kaiser Permanente which the author led as a CEO. Does anyone believe that such software would offer better patient outcomes or allow us to employ fewer capable nurses and doctors?
The examples given in the book are about technology, spots, and car racing - so it's a heavily macho book. Remember, women are key decision makers in most households - addressing their perspective is, if nothing else, a good way to sell your book unless you're writing it for other reasons.
I'll start by saying that trying to predict the future has little value and the author states that legislature, payment systems, regulators, physicians, and patients all have to change for Gen AI to be an effective tool in healthcare. Maybe that could happen but it's never happened in the face of any previous tech.
He's talking about health 4.0 when we've regressed to previous models - so it's important to first understand how we got here before figuring out how we can get to somewhere else. As a previous CEO of a very wealthy HMO, Kaiser Permanente, I'm not sure how much he could really comprehend the day to day struggles of patients and physicians.
He reiterates the point many times that with more Gen AI we can free the doctor to have more time with the patient. What tech in the past ever allowed more time instead of just squeezing out more profits from the patient? And why would the doctor be of any value to spend more time with the patient if Gen AI is so capable, even being able to operate the robotic devices itself to perform the surgeries? A good editor and publisher can sometimes help tease out these plot holes.
Why would a patient take advice from ChatGPT if they aren't doing that already? How would their weight and BP be better if they are getting constant notifications when the same tactics right now are ineffective. Could it be that we have a food system that is designed to create disease to sell more food and to create illness which requires disease? Maybe not but it appears so.
Saying that ChatGPT will democratize health information is quite uninformed because health information is already democratized. We've been in the health information age since the 90s - 4 decades should be more than enough time for even the oldest authors to understand that the problem isn't knowledge - it's much more complex.
Complaining that hospitals are getting paid too little by Medicare and Medicaid can only come from the CEO of Kaiser Permanente. No physician, no patient, not even a politician is staying up late wondering how to get more profits to hospitals; we know they are already squeezing everyone dry, we're trying to figure out how to still keep them engaged without financializing more of medicine.
Comparing the Amazon system to healthcare is terrible because Amazon is terrible in every which way except for improving supply chains. The products we now use are far lower quality than ever before and the impact on the environment is likely not repairable in the next few decades. Nobody wants healthcare to be more like Amazon. The only part of Amazon we want is the great customer service.
Finally, the author mentions the ICU software that can predict which patient will crash. We don't need a software for that - nurses were able to do that better than doctors but now we are having fewer nurses, fewer doctors, and sicker patients just so the medical groups can profit more - especially groups like Kaiser Permanente which the author led as a CEO. Does anyone believe that such software would offer better patient outcomes or allow us to employ fewer capable nurses and doctors?
April 20, 2024
Each year, our medical conferences feature more lectures emphasizing the inevitable integration of AI into every area of our practice. It's an extremely exciting but also controversial and overwhelming subject matter and in CHATGPT, MD, Dr. Pearl presents the material in such an intelligible way.
The book features the most up to date uses of generative AI in science and medicine as well as insightful predictions of its potential to mend decades long faults in the current system and bring the doctor-patient relationship back to the forefront. Dr. Pearl emphasizes a collaborative approach to empower both parties, calming misgivings about an "AI takeover", and proving generative AI is the tech that might finally equalize the power imbalance between a corporate run healthcare system and it's doctors and patients.
Like many, I've lost loved ones to misdiagnoses and unnecessary medical errors. As a professional, it pains me to see the misinformation patients turn to because they can't receive prompt medical care. Reading CHATGPT, MD made me excited about the future where, when assisted by powerful technology, doctors and patients are finally armed to take back control of medicine.
The book features the most up to date uses of generative AI in science and medicine as well as insightful predictions of its potential to mend decades long faults in the current system and bring the doctor-patient relationship back to the forefront. Dr. Pearl emphasizes a collaborative approach to empower both parties, calming misgivings about an "AI takeover", and proving generative AI is the tech that might finally equalize the power imbalance between a corporate run healthcare system and it's doctors and patients.
Like many, I've lost loved ones to misdiagnoses and unnecessary medical errors. As a professional, it pains me to see the misinformation patients turn to because they can't receive prompt medical care. Reading CHATGPT, MD made me excited about the future where, when assisted by powerful technology, doctors and patients are finally armed to take back control of medicine.
September 24, 2025
An important topic, and some interesting food for thought. I listened to a podcast with Robert Pearl about this topic before reading the book. The podcast conveyed the ideas much more succinctly than the book, which repeats itself a bit. I've also found Robert's bias is toward optimism, and some of the claims presented with confidence in the book are contested. Overall a useful read to understand an optimistic case for the future.
October 3, 2025
Robert Pearl offers a timely and pragmatic look at how AI — especially tools like ChatGPT — is reshaping healthcare. He walks the fine line between optimism and caution, showing how doctors and patients can reclaim agency in a system increasingly driven by technology and profit. While some sections dive deep into policy and ethics (which may feel heavy), the book’s core message — that AI won’t replace people, but people who use AI will outpace those who don’t — hits home.
June 27, 2024
this book is actually so interesting and I liked hearing the ways we can harness AI for the benefit of patients. it’ll be interesting to see the way things actually play out over the course of a few years.
July 17, 2024
This has some interesting ideas. A lot of it was a discussion of issues with current healthcare system and how something like ChatGPT might help both the patient and the physician. The author used ChatGPT to assist in writing the book and claimed that made the writing process much faster.
September 16, 2025
I did not like this much at all. So repetitive and ultimately pointless. Pearl offers basically no insight into the use of AI in clinical medicine and just pontificates endlessly about vaguely related tangents.
April 18, 2024
Very interesting insights. If you have any interest in seeing a better healthcare model that is patient centric, relies on value based reimbursement over volume based reimbursement, and allows generative AI to enhance physician capabilities, read this book!
Displaying 1 - 10 of 10 reviews