Today in health, it we're going to take a look at Google DeepMind efforts to suppress chatty BT. We're going to talk about how many AI platforms. Do you need at your health system? And we'll also discuss a little bit about vendor lock-in, which I'm seeing more and more, and I'm a little concerned about my name is bill Russell. I'm a former CIO for a 16 hospital system and creator of this week health instead of channels, dedicated to keeping health it staff current.

And engaged. We want to thank our show sponsors for helping to support our mission, to develop the next generation of health. Leaders short test and artist site. Check them out at this week. health.com/today. Having a child with cancer is one of the most painful and difficult situations a family can face this year. We are partnering with Alex's lemonade stand. We are up over $40,000 on our $50,000.

Goal. For 2023, and we are looking forward to a strong second half. And blowing away that 50,000 number. If you want to be a part of it, just go ahead and hit our website at the top. You can see a banner that's going to have Alex's lemonade. Stand on it. Go ahead and click there. And you can give. At that point.

We want to thank you for your generosity. And we look forward to. Seeing what we can do to fight childhood cancer. All right. , so this article that I've covered and then jump off into a couple other little details. I want to talk about. And it is from venture beat. More details emerge about Google DeepMind efforts to surpass chat GPT.

All right back in April, Google brain and deep mind joined forces as Google deep mind. With plans to take on the competitive threat posed by open AI. And it's game-changing chat GPT. Now, according to wired DeepMind CEO. Dermis D D miss D. Miss Hassabis says the company is working on a new system, Gemini, which was teased at Google IO in may.

That will tap techniques that helped AlphaGo defeat the go champion in 2016. And we talked about that awhile back and it's, it's really interesting. It's a amazingly complex game and AI was able to for the first time, Sort of invent a move that it really surprised the go champion and was able to defeat the go champion.

, with a move that hadn't really been seen before. And that was one of the things people were looking at to say, , is, are we starting to emerge? With artificial general intelligence, where computers start to think. Now clearly. , we don't think that that's what happened, but that's a sort of a, a little glimpse into what they're trying to develop here and the techniques they're trying to use.

According to the report has the best. Said the Jim Nye system will combine large language model technology. With reinforcement learning techniques used in AlphaGo with the goal of giving it new planning and problem solving capabilities. All right. So thinking about that, it's an AI model. Dedicated to planning and problem solving capabilities could be interesting. As you're looking at things you could potentially look at your entire health system, the workflow you could look at.

, yeah, workflow being one of those, which would be really fascinating. You could look at business model what's working. What's not working. , you could potentially do really complex models. , around changes and forecasting. Anyway. It goes on another recent report from the information cited, an anonymous source, who said that Google's researchers have been using YouTube, which Google owns to develop Gemini, which AI practitioners say could be an edge for Google deep mind sits. It can get more complete access to the video data than rivals that scraped the videos.

Also interesting. They will get access to the entire, this week health collection. Cause we put everything out on YouTube so that you guys can access it easily. So think about all those interviews, all those interviews with CEOs. , practitioners, nurses, doctors, you name it. We'll all be out there for it to train the model.

At Google's IO, , the CEO noted in a blog post that Gemini. The company's next generation foundation model is still in training. Gemini was created from the ground up. To be multimodal, highly efficient at tool and API integrations and built to enable future innovations. Like memory and planning. He wrote.

While still early, we're already seeing impressive multimodal capabilities not seen in prior models. Once fine tuned and rigorously tested for safety. Gemini will be available at various sizes and capabilities. The reports about Gemini come as Hassabis himself has himself has expressed concern about the long-term existential threats of AI. The wired report has speced says that no one really knows for sure that AI will become a major danger, but if it progresses, if progress continues at this pace,

There isn't much time to develop safeguards. , aspects also recently was one of the most prominent signatories of the statement of AI risk,

which warned against the risk of extinction. From advanced AI. If it's development is not properly managed. So. It's interesting. Somebody said to me, you know, you notice all the people who are signing this thing are people who don't have models and the people with models didn't sign it, meaning they want them to slow down so that they can catch up. That was sort of the inference that was made. I'm not sure I disagree with that, but I also believe that there should be safeguards around AI as we move forward.

, you know, I asked the question earlier on how many AI models will you have in your health system. And the answer is as many as it takes. And we may have special function, AI models. It's just like you have special function vehicles. You have trucks. You have, , ambulances, you have cars, you have, , potentially bikes and mopeds and stuff to get around campus.

, you will have different functioning AI models to do different tasks and you will leverage those models. To their, , to their training and to their highest and best use within your organization. I believe that's what we're going to see. We're going to see multiple models applied at different health systems.

To the area of their expertise, right? You're not going to see one model to rule them all. You're going to see multiple models that are efficient at what they do. So I think that is the case. I do want to warn against a vendor. Lock-in. I'm seeing more and more, and especially in this area of cameras in the hospital rooms.

I see people heading off in this direction and putting hospitals in camera. Eh, camera's in hospital rooms. And the number one question I would ask if I were putting cameras in hospital rooms. Period would be. Can I use the video stream for any application? If the answer to that is no find a different camera.

Even if it's tied to an existing vendor solution. Oh, I'm going with X, Y, Z solution for nurse sitting. But the solution is this camera and this camera will only feed that solution. Get a different solution. You don't want to be locked into this. You don't want to put thousands and thousands of cameras out there and have to rip them out. We've done this so many times in healthcare. It is absolutely crazy.

And I'm doing this under the idea of AI, because you still have to keep an eye on. As I get involved with these AI models. You know, where's my data going? How's it being used? , can I rescind my data? If I want to rescind my data? You, you have to keep your hands on the steering wheel, much like the example I gave about cameras.

You have to keep your hands on the steering wheel. I want cameras in those rooms that are essentially multipurpose. I can feed one system. I can feed the next system. When the application changes in three years, I can. Then take that feed and move it over to that system. , I would not treat the cameras.

As if they are throwaway devices. We can put in high, high, high resolution cameras. That are going to last us for 5, 6, 7, 8. 10 years into those. Rooms, but if you do it with vendor, lock-in in place, you're gonna end up with. Six different cameras because each one of the systems you're trying to use.

Will not allow you to use the feed from the camera for that system. For the love of Pete, please get ahead of this. , same thing with your AI models. Do not get locked in. So anyway, that's all for today. Do you know someone that would benefit from our channel, please forward them a note, let them know you're listening.

They can listen to wherever they download their podcasts. We are just about everywhere. You can think of. We want to thank our Chell sponsors once again, for being a part of our mission to develop the next generation of health leaders, short tests and 📍 artists, I check them out at this week. health.com/today.

Thanks for listening. That's all for now.