30.
✨ New Tool which converts any scan video to a scrollable viewer format
☯️Introducing a free offline web app - Raddoc's scroller
Ever wanted to share an interesting CT or MRI case with a colleague, but found it difficult? Sending a video is clumsy because they can't control the scrolling, and sending a folder with hundreds of images is a hassle.
I've created a simple web app for this!
Introducing Raddoc's scroller, a tool that turns your screen-recorded/ video recorded scan videos into a smooth, interactive viewer, just like the ones on Radiopaedia!
☯️How it Works (It's Super Easy!):
Open the above web app "RADDOC' scroller.html" in any browser. You have to switch on to desk top mode then if you are using it in your smartphone.
Record a Video: On your computer or phone, simply screen-record/video record yourself scrolling through the scan from start to finish. Save it as an MP4, MOV, etc.
Upload to RADDOC's scroller: Open the above web app in any browser and upload your video file.
In seconds, the app extracts the frames. You can now use your mouse wheel, arrow keys, or touchscreen to scroll back and forth through the image slices smoothly.
📤 The Best Part: Share It With Anyone!
Once you're happy with your scrollable scan, click the "Export as HTML" button.
This creates a single, self-contained HTML file that you can send to anyone. They just need to open the file in any web browser (like Chrome or anyother browser) to view and interact with your scan. No special software needed!
☯️Perfect for:
- Getting a quick second opinion from a colleague.
- Presenting cases in online discussions.
- Give it a try and let me know what you think!
☯️Subscribe our channels to get similar interesting updates
https://whatsapp.com/channel/0029Vb2S2bW0G0Xq94mR721T
raddoc96" rel="nofollow">https://youtube.com/@raddoc96
/channel/raddocs
/channel/radiology_chatgpt
/channel/chatgpt_prom
### Analysis of "The Decreasing Value of Chain of Thought in Prompting"
---
### Question 1: What is the central argument of this report regarding the 'Chain of Thought' (CoT) prompting technique?
Simple Answer:
The main point of this report is that the popular technique of telling an AI to "think step-by-step" — known as Chain-of-Thought (CoT) — is not the magic fix for getting better answers that many people believe it is. Its usefulness is decreasing, especially with newer and more advanced AI models. Whether you should use it depends entirely on which AI you are using and what you are asking it to do. It is not a one-size-fits-all solution.
---
### Question 2: How does the effectiveness of CoT prompting change depending on the type of AI model used?
Simple Answer:
The authors tested two different categories of AI models and found a clear difference in how they responded to the "think step-by-step" instruction.
* For "Non-Reasoning" Models (like standard GPT-4o, Sonnet 3.5): These are the general-purpose AI models that most people use today. For these, telling them to think step-by-step can still be useful. It generally boosts their *average* performance, helping them get more complex questions right. However, the report found a strange side effect: it can also cause them to make mistakes on easy questions they would have otherwise answered correctly.
* For "Reasoning" Models (like the experimental o3-mini, o4-mini): These are newer, more powerful models specifically designed to be good at complex reasoning from the start. For these models, telling them to "think step-by-step" provides almost no benefit. They are already thinking in a structured way internally. Forcing them to write it all out just wastes time and money without making the answers any more accurate.
In short, the trick is becoming less useful as AI models get smarter and more capable on their own.
---
### Question 3: What is the key trade-off the authors discovered when using CoT prompting?
Simple Answer:
The key trade-off is a classic case of "higher average vs. perfect consistency."
Imagine a student taking a math test.
* Without showing their work (Direct Answer): The student might quickly answer all the easy questions correctly but get stuck on the hard ones. They get a decent score but miss the high-value problems.
* By showing their work (CoT / Step-by-Step): The student now carefully works through every problem. This helps them solve the *hard* problems they would have missed. However, in the process, they might make a simple calculation error on an *easy* problem, which they would have gotten right before.
This is exactly what the report found with some AI models. Using "step-by-step" prompting:
* Benefit: The AI's *average score* across all questions went up, because it could solve more difficult problems.
* Cost/Downside: The AI's *perfect consistency* went down. It started making occasional errors on questions it would have aced every single time when asked for a direct answer. It became less reliable on the "easy stuff."
So, you have to decide what's more important for your task: a better average performance, or guaranteed correctness on the simpler parts?
---
### Question 4: What is the impact of using CoT prompting on the cost and speed of getting an answer from an AI?
Simple Answer:
The impact is significant and negative. Using the "think step-by-step" method makes the process much slower and more expensive.
* Time: The report's charts (Figure S1) show that getting an answer took anywhere from 35% to 600% longer when using the step-by-step prompt compared to a direct request. A task that takes 2-3 seconds could suddenly take 15-20 seconds.
29.
☯️Create Stunning AI Presentations for FREE! 🚀
Hey everyone! I'm thrilled to announce my new web app: RADDOC's AI Presentation Generator! ☯
Tired of spending hours on slides?
Let AI do the heavy lifting! This app helps you go from idea to impressive HTML presentations in minutes.
🔗 Try it NOW:
https://aistudio.google.com/app/apps/drive/1h59exwjjYu8-LcTj0-uMROe5Ltk8wNq1?showPreview=true
👉 IMPORTANT FIRST STEP: You'll need to sign in to Google AI Studio at the link above before you can use the app. It's quick and easy!
What can it do for you? So much! 👇
📄 Multiple Input Options: Upload your text files, PDFs, DOCX, HTML, JSON, or simply paste your content. You can even just give it a topic!
🧠 Smart AI Models: Choose from powerful Gemini models like gemini-2.5-flash for speed or gemini-2.5-pro (when available and selected by me for specific tasks) for depth.
🌐 Google Search Integration: If you provide a topic, the AI can use Google Search to gather fresh information for your slides (and it will list the sources!).
🖼️ Image Magic:
Provide direct URLs for your images.
Generate images with AI (powered by Imagen 3.0!) right within the app based on your descriptions.
Quickly search Google Images for inspiration.
🎨 Themes & Customization: Select from various Reveal.js themes to match your style.
✨ Iterative Refinement: Get an initial draft, then enhance it! Add images, provide further text instructions (like "make it more formal" or "add a slide about X"), and even use the full history of your changes for better AI context.
🎤 Voice-to-Text: Dictate your content for topic queries, pasted text, and image descriptions using your microphone!
🎲 Experimental Interactive Slides: Feeling adventurous? Let the AI design a unique, engaging interactive slide for your presentation!
📝 Plain Text for PowerPoint: Export the slide content as plain text, perfect for manually creating or importing into PowerPoint or other editors.
👁️ Live Preview & Download: See your presentation come alive in the preview and download the final HTML file to present anywhere.
🛠️ AI Error Helper: If something goes wrong, the app can even try to explain the error in simple terms using AI!
How it Works (Quick Guide):
1.Go to the link (and sign in to Google AI Studio if you haven't!).
2.Setup: Provide your content (files, text, topic), choose slide preferences.
3.Generate: Get your initial HTML presentation.
4.Refine: Add images (URLs or AI-generated), give more instructions, add interactive elements.
5.Download: Grab your awesome HTML presentation and share it with the world!
This app is built to be user-friendly and powerful. I'm really excited for you to try it out and see what you can create!
Let me know what you think! Your feedback is invaluable. 🙏
RADDOC ☯
🔴For other free and interesting non academic use cases of AI
👇
/channel/chatgpt_prom
Fellow Radiologists, ☢️ Take a look at this new development in AI: LLaVA-Rad.
It's a smaller, open-access AI model specifically trained for generating CXR report findings.
Accuracy Focus: The research shows it's getting really good at generating factually correct findings sections for Chest X-rays. ✅
Smarter Evaluation: They also introduced a new way to evaluate these AI reports (using GPT-4) that seems to align much better with how we radiologists actually judge report quality, compared to older metrics. 📊 Finally!
Accessible Tech: Being lightweight and open-source means it's not just for big tech companies – it opens doors for more practical integration down the line.
Definitely an interesting piece of research pointing towards more usable AI assistance in the reading room.
http://youtube.com/post/UgkxhUFMiJv7tgIluwfbC6qY6lNV7_dE_XaS?si=xbLn_Rikg0rrFA8b
Читать полностью…
🔴For other free and interesting non academic use cases of AI
👇
/channel/chatgpt_prom
It is a free course on AI AGENTS.
👇
https://huggingface.us17.list-manage.com/subscribe?u=7f57e683fa28b51bfc493d048&id=9ed45a3ef6
AGI/ASI: The article emphasizes that the arrival of AGI, capable of human-level performance across broad tasks, could be a turning point. The potential for ASI, surpassing human intelligence, further intensifies the concern. The timeline for AGI is uncertain but speculated to be within the next few years, with the leap to ASI potentially following rapidly. It's suggested that not just large language models (LLMs), but other architectures may be needed for AGI. At this stage, the economic advantages of employing AGI/ASI for interpretation could lead to significant displacement of human radiologists across all subspecialties.
4. What is the author's perspective on the timeline for these AI advancements, and how does he address the counterarguments regarding the limitations of current AI technology?
Answer:
The author acknowledges that predictions regarding AI timelines are inherently uncertain. He cites Geoffrey Hinton's initial prediction of AI outperforming radiologists within 5-10 years (made in 2016), later revised to be overly conservative. Other figures, like Sam Altman and Dario Amodei, suggest AGI could emerge as soon as 2027. The author also references Yann LeCun's argument that LLMs alone won't suffice for AGI and that more complex architectures are needed, highlighting that the path to AGI/ASI may involve multiple approaches.
Regarding counterarguments about AI's current limitations, the author recognizes that narrow AI tools are not yet capable of replacing radiologists entirely. However, he emphasizes that these tools should be viewed within the context of a broader trajectory towards AGI/ASI. He argues that the development of agentic AI represents a significant step forward and that the potential for rapid advancement should not be underestimated.
5. What are the broader implications of the potential widespread adoption of AGI/ASI in radiology, beyond the immediate impact on the workforce, as suggested by the article?
Answer:
The article primarily focuses on the workforce implications, but it hints at broader systemic changes:
Healthcare Economics: The widespread adoption of AGI/ASI could significantly alter the economics of healthcare, potentially reducing costs associated with radiology services. This could have implications for reimbursement models, hospital budgets, and the overall financial structure of the healthcare system.
Access to Care: AI-driven interpretation could potentially improve access to radiology services, particularly in underserved areas or during off-hours. However, it also raises questions about the equitable distribution of these benefits and the potential for disparities. The article touches on the concept of "smart triage" and protocol optimization software, which could be implemented using AI to manage imaging volumes.
The Nature of Medical Expertise: The article implicitly raises questions about the future role of human expertise in medicine. If AI can perform diagnostic tasks at a level comparable to or exceeding that of humans, what will be the unique value proposition of human radiologists? This may involve a shift towards more complex, interventional procedures, or a greater emphasis on the human aspects of patient care. The author suggests that the transition to AI-driven interpretation will be gradual, with human radiologists initially serving as a "secondary check" before AI takes on more autonomous roles. The scope of human-driven interpretation will steadily shrink.
27.Introducing 🆓 Radiology PDF Tutor:
Reading and understanding a full radiology PDF can feel overwhelming, but Radiology PDF Tutor makes it simple, fast, and engaging.
How It Works:
1️⃣ Sign in with your Google account here
👇
https://aistudio.google.com/
2⃣ Then access this exact website
👇
https://aistudio.google.com/app/u/0/prompts/1pLNK44j3WWSr3dZznAeY_GCtbi5aIDt1
3⃣ Upload Your PDF. Select 'Upload File' to upload a radiology article or book chapter.
4⃣ Learn Step-by-Step. After uploading, click the run icon and wait a few seconds. The tutor will start teaching you one page or section at a time.
5⃣ Control the Pace. Once you finish a section, let the tutor know, and it will move to the next one.
6⃣ Ask Questions Anytime. Have doubts? Pause and ask for clarification before continuing.
Why Use Radiology PDF Tutor?
✔️ Simplifies Complex PDFs – Breaks content into digestible sections for easier understanding.
✔️ Interactive Learning – Learn at your own pace and get answers to your questions.
✔️ Fast & Convenient – Makes learning radiology engaging and productive.
🔗 Here’s the link for Radiology PDF Tutor
https://aistudio.google.com/app/u/0/prompts/1pLNK44j3WWSr3dZznAeY_GCtbi5aIDt1
Tip:
Save this link as a shortcut or bookmark to access it conveniently whenever you need it.
Avoid making a copy—it stores unnecessary data which can clutter the model and cause confusion.
Using the original link ensures that a fresh, new session opens every time, making it easier and hassle-free.
With Radiology PDF Tutor, you can finally make sense of radiology PDFs, one page at a time, while enjoying an interactive and engaging learning experience. Try it now!
For more similar interesting Radiology tools
👇
*️⃣/channel/radiology_chatgpt
*️⃣https://whatsapp.com/channel/0029Vb2S2bW0G0Xq94mR721T
*️⃣/channel/raddocs
26.Radiology report creating..
To merge your positive findings into the normal report template you already have, follow these steps:
1. Join the Discord server using this link: https://discord.gg/2ksrXjkwnx.
2. Navigate to the bot9-report-template-with-findings-merger channel.
3. Send only the positive findings in a single message in that channel.
4. After receiving the AI's response, send the normal template text as the next message.
5. The AI will merge the positive findings within the normal report template.
Before that, if you want to write all the positive findings of the scan using speech-to-text quickly, refer to this message:
/channel/radiology_chatgpt/261
If the final report produced by the AI cannot be pasted into MS Word due to font size or symbol issues:
Paste the final report into https://markdowntohtml.com/
Delete the existing text in the "Enter markdown" box there, then paste the final report generated by the AI within that box.
You will receive a properly formatted radiology report below that box then.
🔴For other free and interesting non academic use cases of AI
👇
/channel/chatgpt_prom
25. Advanced Radiology Report Creating AI
https://miniapps.ai/raddoc
☯A Free, accurate speech-to-text conversion tool for creating radiology reports.
👇
http://youtube.com/post/Ugkxtk-h_tq49V_dEZoLLVySWwoB-LO48LHs?si=cXQTgpy3mDuwJbV5
* Cost: AI companies often charge based on how much text the AI has to process (measured in "tokens"). When an AI "thinks step-by-step," it generates a lot more text (all the reasoning steps) before giving the final answer. This increase in text directly translates to a higher cost for the user.
Therefore, especially when the accuracy benefit is small or non-existent (as with newer models), you are paying a high price in time and money for very little gain.
---
### Question 5: What is the authors' main purpose for writing this report, and who is their intended audience?
Simple Answer:
The authors' main purpose is to provide practical, data-driven advice to people who actually use AI in their daily work. They want to move beyond hype and give clear guidance on what works and what doesn't.
The report explicitly states its audience is "business, education, and policy leaders." Essentially, this is for anyone who isn't an AI researcher but needs to make informed decisions about how to use AI effectively and efficiently. They are trying to correct a common misunderstanding and help people save time and money by not using an outdated or inappropriate technique.
---
### Question 6: Based on the report's findings, what is the practical advice for someone using AI today? When should they use CoT, and when should they avoid it?
Simple Answer:
Here is the practical advice from the report, boiled down into simple rules:
#### When you SHOULD consider using "Think Step-by-Step":
* You are using an older or less advanced AI model that isn't great at reasoning on its own.
* The task is very complex, and a slightly better *average* performance is more important than getting every single simple part perfect.
* You are not concerned about speed or cost.
#### When you SHOULD AVOID using "Think Step-by-Step":
* You are using a state-of-the-art AI model (like the latest from Google, OpenAI, etc.), as these models likely already do this by default. Telling them to do it is redundant.
* Perfect consistency is crucial. If you need the AI to be 100% reliable on tasks it can easily do, CoT might introduce random errors.
* Speed and cost are important. If you need a fast, cheap answer, CoT is one of the worst things you can do.
One final, very important piece of advice from the report: Be careful about forcing the AI to *only* give you the final answer (e.g., "Just provide the letter of the correct answer and nothing else"). The report found that many modern AIs think step-by-step by default, even when not asked. Forcing a direct answer can prevent this helpful internal process and actually make the AI's answer *worse*. The best approach is often to just ask the question naturally and let the AI respond as it sees fit.
🤖✨ Discovering How AI “Thinks” Behind the Scenes! ✨🤖
Hey everyone! 🎉 Have you ever wondered how those smart AI assistants (like ChatGPT) come up with their answers? A recent study digs deep into just that—and the results are super cool, even if you’re not a techie! Here’s the scoop in plain language:
🌟 What’s the Big Idea?
Researchers wanted to see two things separately:
1️⃣ Knowledge – What facts or “stuff” the AI already knows.
2️⃣ Reasoning – How the AI uses those facts to connect the dots and solve problems.
Instead of only checking whether the AI’s final answer was right or wrong, they looked at every step of the AI’s “thinking process.” Think of it as watching a student write out all their steps on a math or medical exam, rather than just seeing the final grade.
📊 Why Does That Matter?
More Transparency:
We often see only the AI’s final answer. By breaking down its process, the study shows exactly where the AI is strong (knowing facts) and where it might stumble (logical reasoning).
Better in Medicine vs. Math:
In medicine, having correct facts is crucial—one tiny mistake can lead to the wrong diagnosis. The study found that teaching AI more medical facts (through a process called “fine-tuning”) made it much better at medical questions.
In math, logic and step-by-step problem solving matter more. Here, a different training method (called “reinforcement learning”) helped the AI sharpen its reasoning skills.
Customized AI Training:
Since different fields (medicine, math, law, finance, etc.) have different needs, this research shows we can train AIs in a smarter way: focus on facts where it matters most, or focus on reasoning where that’s the priority.
🔍 Key Findings—Simplified!
When the AI was only taught medical facts, it got better at medical quizzes but its “how did you get that answer?” steps became a bit messy.
When the AI was trained to think more carefully through each step, it made fewer mistakes in its reasoning—even though it didn’t learn new facts as thoroughly.
For medical tasks, knowing the right facts was actually more important than fancy reasoning tricks. (Because medicine relies on accurate information!)
For math challenges, reasoning clearly was the key to success.
💡 What’s So Awesome About This?
Safer, More Trustworthy AI: In fields like medicine, we need to trust that the AI isn’t just guessing. This study’s approach shows exactly where the AI might be confident in a fact or where it might be on shaky logical ground.
Tailored AI Helpers: Imagine AI doctors that are really solid on medical facts, or AI tutors that excel at guiding you through each step of a tricky math problem. This research is a big step toward that future.
Understanding AI “Thought”: By breaking down how AI thinks, we can spot and fix mistakes earlier, making AI more reliable for everyone.
🎈 In a Nutshell:
This research peels back the curtain on AI decision-making, separating the “what it knows” from “how it reasons.” Whether it’s helping doctors or solving math puzzles, we can now train AIs to be smarter and safer in exactly the ways we need them to be. How cool is that? 🤩
Feel free to share and spread the word—AI is only going to get more amazing! 🚀✨
https://x.com/raddoc96
➡️ Follow me on twitter to get updates and nice AI use cases in medicine which will help in your work/study.
O3 in radiology (effect of reasoning being applied at image recognition level rather than text level alone)
https://x.com/ASK_MSK/status/1912594507398566027?t=evOxWsrTXiBqoOzG1mo49Q&s=19
https://x.com/raddoc96/status/1909675223798198631?t=WeTobCqVsZCghm8u5_iykQ&s=19
Читать полностью…
https://telegra.ph/Meet-the-Future-of-AI-Gemini-25-Pro-03-25
Читать полностью…
https://telegra.ph/AIs-New-Thinking-Trick-How-Exploring-Many-Solutions-Makes-AI-Smarter-03-18
Читать полностью…
28. New tool from RADDOC☯️
Struggling to Take Lecture Notes in class? Here’s a Free AI Tool to Help!
How It Works:
1️⃣ Record Your Radiology Lecture – Just record the audio of your lecture on your phone or any recording device.
2⃣Sign in with your Google account here
👇
https://aistudio.google.com/
3⃣Then access this exact website
👇
https://aistudio.google.com/app/u/0/prompts/1swyXQs1dp-CR0uD-3Owbzegp08nbkmoT?pli=1
4️⃣ Upload the Audio File there after clicking the plus icon and then clicking "upload file".Wait for sometimes to get it uploaded completely.Then , click send.
5️⃣ Get a Structured Transcript – The tool will instantly convert your audio into a well-structured and organized transcript of the lecture.
6️⃣Copy the generated text there, by clicking "Copy text" , then paste it into Google Docs.From there, you can export it as a professional-looking PDF containing the entire lecture content.
Why Use This Tool?
✔️ Saves Time – No need to manually jot down every detail.
✔️ Accurate Transcripts – Captures the entire lecture in an organized format even if the audio quality is poor.
✔️ Easy Sharing – Export as a PDF to study or share with your peers.
Tips:
* Save this exact link as a shortcut or bookmark to access it conveniently whenever you need it.
Link - https://aistudio.google.com/app/u/0/prompts/1swyXQs1dp-CR0uD-3Owbzegp08nbkmoT?pli=1
* Avoid making a copy—it stores unnecessary data which can clutter the model and cause confusion.
* Using the original link ensures that a fresh, new session opens every time, making it easier and hassle-free.
Bonus🎁Another interesting tool:
Already created the lecture transcript PDF?
Now, there is another awesome thing you can do.
👇
Check out this post to learn how to make ai to teach you that lecture to you step by step in interactive manner.
👇
/channel/radiology_chatgpt/273
* Upload your lecture PDF there
in the link given in the above post.
* Use the tool to learn interactively at your pace.
* Ask questions and clarify doubts step-by-step while the AI teaches you the content.
With these AI tools, capturing, organizing, and learning from lecture is becoming easier. Try it today!
Subscribe our channels to get similar interesting updates
https://whatsapp.com/channel/0029Vb2S2bW0G0Xq94mR721T
raddoc96?si=xD3s2Tp7hLygeMdI" rel="nofollow">https://youtube.com/@raddoc96?si=xD3s2Tp7hLygeMdI
/channel/raddocs
/channel/radiology_chatgpt
/channel/chatgpt_prom
Here are five essential questions that capture the main points and core meaning of the provided text, specifically tailored for an audience of experienced radiologists. I've addressed the central themes, key supporting ideas, important facts, author's purpose, and significant implications in each question, followed by detailed answers:
1. What is the central argument presented regarding the future of radiology in the face of Artificial Intelligence (AI) advancements, particularly Artificial General Intelligence (AGI) and Artificial Superintelligence (ASI)?
Answer:
The central argument is that the radiology profession faces an existential challenge due to the rapid advancement of AI, specifically the potential emergence of AGI and ASI. Unlike the current "narrow AI" tools that augment radiologists' capabilities, AGI/ASI could potentially match or exceed human performance across the full spectrum of diagnostic tasks. This, coupled with a projected radiologist shortage and increasing imaging study volumes, creates a scenario where the economic and practical pressures may favor the widespread adoption of AI-driven interpretation, significantly reducing the need for human radiologists. This may not be a sudden replacement, but a staged transition from current AI to more autonomous reading, shrinking the scope of human-driven interpretation. The author posits that this transition is not driven by malicious intent from AI developers but rather as a pragmatic solution to systemic pressures within healthcare.
2. What are the key demographic and economic factors highlighted as contributing to the potential decline in the need for human radiologists, as outlined in the article?
Answer:
Several key factors are presented:
Demographic Cliff: A significant portion (53%) of practicing radiologists were over 55 in 2022, suggesting a large wave of retirements within the next decade. The influx of new radiologists (approximately 1,000-1,200 annually) is insufficient to offset this attrition, potentially leading to a 20%-25% workforce reduction by 2030. This will result in a higher workload per radiologist, with estimates suggesting an average of 35,000 cases per year, significantly higher than the current average.
Increasing Imaging Volume: The number of imaging studies in the U.S. is growing at a compound annual growth rate (CAGR) of approximately 4.2%, with projections exceeding 875 million studies annually by 2030. This growth is driven by factors like an aging population and increased utilization of imaging in emergency settings.
Economic Strain: While the demand for radiologists is increasing, downward reimbursement pressures limit the ability to increase salaries to attract and retain talent. This creates a financial strain on practices and healthcare systems, making cost-effective solutions like AI more appealing. The article highlights that by 2030, there could be fewer than 25,000 radiologists handling over 875 million studies, a stark contrast to the current scenario.
3. How does the article characterize the progression of AI in radiology, from "narrow AI" to the potential emergence of AGI/ASI, and what are the implications for the role of human radiologists at each stage?
Answer:
The article outlines a progression:
Narrow AI: Current FDA-cleared algorithms are described as "narrow AI," excelling at specific tasks like nodule detection or fracture identification. These tools are seen as augmenting radiologists' capabilities, increasing efficiency, but not replacing them entirely. Their accuracy often approaches, but doesn't surpass, that of humans.
Agentic AI: This is presented as a transitional phase, where AI can perform end-to-end interpretation for specific workflows (e.g., screening mammography). These systems require minimal human oversight and are expected to see increased adoption in the next 1-3 years. Agentic AI will begin to be used more in areas like overnight or emergency radiology, followed by subspecialty domains like complex MRI. This phase may be a stepping stone towards AGI/ASI.
Difference between previous llms(gpt4o/claude 3.5 sonnet/meta llama) and recent thinking/reasoning llms(o1/o3)
Think of older LLMs (like early GPT models) as GPS navigation systems that could only predict the next turn. They were like saying "Based on this road, the next turn is probably right" without understanding the full journey.
The problem with RLHF (Reinforcement Learning from Human Feedback) was like trying to teach a driver using only a simple "good/bad" rating system. Imagine rating a driver only on whether they arrived at the destination, without considering their route choices, safety, or efficiency. This limited feedback system couldn't scale well for teaching more complex driving skills.
Now, let's understand O1/O3 models:
1. The Tree of Possibilities Analogy:
Imagine you're solving a maze, but instead of just going step by step, you:
- Can see multiple possible paths ahead
- Have a "gut feeling" about which paths are dead ends
- Can quickly backtrack when you realize a path isn't promising
- Develop an instinct for which turns usually lead to the exit
O1/O3 models are trained similarly - they don't just predict the next step, they develop an "instinct" for exploring multiple solution paths simultaneously and choosing the most promising ones.
2. The Master Chess Player Analogy:
- A novice chess player thinks about one move at a time
- A master chess player develops intuition about good moves by:
* Seeing multiple possible move sequences
* Having an instinct for which positions are advantageous
* Quickly discarding bad lines of play
* Efficiently focusing on the most promising strategies
O1/O3 models are like these master players - they've developed intuition through exploring countless solution paths during training.
3. The Restaurant Kitchen Analogy:
- Old LLMs were like a cook following a recipe step by step
- O1/O3 models are like experienced chefs who:
* Know multiple ways to make a dish
* Can adapt when ingredients are missing
* Have instincts about which techniques will work best
* Can efficiently switch between different cooking methods if one isn't working
The "parallel processing" mentioned (like O1-pro) is like having multiple expert chefs working independently on different aspects of a meal, each using their expertise to solve their part of the problem.
To sum up: O1/O3 models are revolutionary because they're not just learning to follow steps (like older models) or respond to simple feedback (like RLHF models). Instead, they're developing sophisticated instincts for problem-solving by exploring and evaluating many possible solution paths during their training. This makes them more flexible and efficient at finding solutions, similar to how human experts develop intuition in their fields.
Our WhatsApp channel
https://whatsapp.com/channel/0029Vb2S2bW0G0Xq94mR721T
📎My Twitter Account - https://x.com/raddoc96
Читать полностью…
Stanford launched a free Google Deep Research clone called STORM.
It uses GPT 4-o + Bing Search under the hood to generate long cited reports from many websites in ~3mins.
It's also completely open-source and free to use.
👇
https://storm.genie.stanford.edu/
I recommend you to use my original link as bookmark / homepage only as it will be convenient and easy to use everytime instead of making a copy... because, if u r copying that, all the voice messages will be stored each time unnecessarily and it will lead to confusion in the model.
But, if u save my link and use it each time, new chat will open freshly to use.
24.
🚨 Transform Radiology Dictation (SPEECH TO TEXT) to Text with FREE Google AI Studio! 🚨
📢 You can now use the Google AI Studio link I’ve shared below for high-quality Radiology Speech-to-Text Conversion.
https://aistudio.google.com/app/u/0/prompts/1HeFhBkFAWNa7r2vNkGVzBNuvTBlIOPOf?pli=1
🎙️ How It Works:
You can either record your radiology dictation (from 2 minutes up to 50 minutes) and upload the audio file or you can even directly record the speech in Google ai studio itself.
The tool will recognize radiology-specific content and ignore unrelated chit-chat or casual conversations in the recording.
This is NOT for live transcription or short sentences—it's designed for detailed, large-scale dictations.
🛠️ How to Access:
1. Sign in with your Google account with the above given Google ai studio link.
2. Upload your audio file/Click record audio option for live speech to text feature
3. Get a seamless, high-quality transcript of your radiology report.
Save this link and revolutionize how you create radiology reports! 📄✨
https://aistudio.google.com/app/u/0/prompts/1HeFhBkFAWNa7r2vNkGVzBNuvTBlIOPOf?pli=1
@RADIOLOGY_CHATGPT
@RADDOCS
