What if Natural Language interface was just part of every applications features. Crazy thought. Can agent power AI outperform the latest models? Today we introduce the concept.
You can help it. It's going to be a loud show. I am on-site. I dropped off my car and I'm recording at the dealership. So he knows what sounds well here. But we are going to do a followup story to AI, ambient clinical listening from yesterday. And we're going to do a couple other AI stories that I followed my name's bill Russell.
I'm the former CIO for 16 hospital system creator, which we felt. So the channels in the bench dedicated to transforming healthcare, one connection. At a time. And it's like our show sponsors for investigating, developing the next generation of health leaders. Moldable service now. Enterprise health parlance. Certified health and handout.
Check them out this week off. Dot com slash today. ──── All right. Hey, don't forget. You're on this podcast with a friend or colleague. You. It's a foundation for daily or weekly discussions on the topics that are relevant to you in the industry. One. More people to be mentoring and that's what we're driving for.
They can subscribe wherever you listen to podcasts. ─── All right. Let's get to the articles. ── First one is a follow-up problem yesterday. And it's really interesting to me. Because I said that I believed that the cost of ambient scribes, ambient listening, what's going to continue to come down. And then this story came across my transom. And the base come your launches free AI powered, ambient scribe available to 350,000 plus providers across 300. 50 hospitals. So today. We are excited to announce availability free. You get the whole idea. AI powered scribe. Capable of generating soap notes and automating documentation integrated. Into the EMR. This school is billboard processed food. At about the patient's strong line patient keeper in New York, engage. And. ── DLS RCM and RPM. And available with integrations into epic theater, Meditech more. With this release over 250,000 fighters, 350 locations, 40 major hospitals. We describe as a finalist in the BA AI technology talent. Which evaluated over 350. Plus scribing products. It has been shown to save three hours a day and back from the patient time.
It's interesting. We went through the pricing model yesterday. In the show. I was saying how I thought there was going to be pressure. To drive it down. I know a lot of organizations. ──── Focus in on whatever their current technology is, but I don't think the cost of. Switching. I'm swapping out and technology's going to be back right here. And I think more and more, we're going to see this stuff integrated into
Microsoft Mechanics
primary tools. ── And. And potentially you could see something that's just baked into epic.
I know that sounds crazy. Given their partnership. With various companies. But I think most of our applications, this will just become a feature of the application. Some short. Natural language front end. Four. Putting information into the. ──────━
Into the application, whatever that application has to be, it could be the EMR can be any other clinical application that happens. ─────────
All right. Here's another one I want to cover is open AI and Microsoft reportedly planning. A hundred billion dollar data center. ─ And so the question I was asking you is, what does this tell us?
What are we. What does Microsoft believe in order to make this kind of investment? Literally, they believe it. ─────
Hi. It's the future. Of computing. As the. Almost anyone at this point. But I think the other thing that's tells us. Is that they believe in a. ── They are going to be. ─ A foundational element of. AI. ─ Computed moving forward. If you look at these like Ambien square, I. Pick out a bunch of old work. Off of open AI, so they could be the natural language front end for all sorts. Technology partners that are out there. They could also have Visibility into what they're going to be able to do with. You could be 5, 6, 7. You name it? And potentially simplify this whole environment. To the point where you have. One model to rule them all. If you go. ──── Now. I find that interesting.
I did watch a video and I'm going to cover this video tomorrow. And today's show. ── And it talks about. AI agents. ─ And the ability for AI agents to produce better results. In fact, the ability for AI agents. To produce better results with chatbot. 3.5. ── And the various agents can be adversarial or they can be cooperative. ─ Meaning. We do this in our brain all the time, let's do this. You write a paper, write an article, you write something. ─ And then we will go back and proofread. When we proofread it. We proofread it as operative, but also somewhat as an adversarial. We're looking for the mistakes.
We're looking for punctuation mistakes. We're looking for. Grammatical mistakes. We're looking for. Mistakes in the articles so that we can correct it. I
think of Chad GPP four as a single threaded AI engine, you ask it a question. It responds with something. ───
Then in an ancient world, you would have these other agents that interrogate the information. Is that quote actually in the transcript that we gave. ── Is the a E. Is the spelling. Correct. It's the punctuation? Correct? Is this a good blog article or is it a good article for a magazine? Is this a good article for a newspaper? Is this a. Research paper. You know what. You could have adversaries who were asking questions like, Hey. ─ Is this, should this be, could it be. ─ Is this correct?
And back. That kind of stuff. ─ And there's. This whole body of work. It's goes on around this. And new technologies coming up around this. That take. Just basic stuff out of chat TPP. JPP. 3.5. And put agents around it. And just with those agents. Yet above tattooed BP four. Quality. ─── So think about that because the cost of 3.5 is a lot less than four. ─ At four will be less than five.
So you have this escalating Fox and the escalated compute power. That's required. But if we implement a. ─ Architecture. That allows for agents to. Maybe move through the information multiple times, which is by the way, how we've worked for. Ever. You could produce those ─── results. If not. Better results than the current models of the. Coming down. So that's just something to keep an eye on.
I'll do a show on that tomorrow. But this whole AI space is interesting to me actually. I see people doing talks. Some channels. And discussions going on. And I wonder how far people have. Gone into this. ── How much research are you doing? ── From where I sit at this point, this is a. ─ Transformative technology. ── And it's moving faster than any other technology. ────━───
In my brief history on this planet. You could say that. Cell phones really changed things and say the internet changed things. We can say the PC changed things. They changed things over time. This is changing industries very rapidly, and I believe we're going to see the same thing in healthcare.
I think we're going to see it higher quality. Results from using these generative models using agents. And so that's what I've been covering tomorrow. Interesting time to be around. Hopefully you didn't pick up too much background noise. During this. ─ During this. Recording. And hopefully tomorrow I will be back in front of my for my desk. To a three or four minutes. All right.
That's all for today. Don't forget. Share this podcast with a friend or colleague. You should, as a foundation for mentoring. We want to thank our channel sponsors who are investing in our mission and develop the next generation of health leaders. Notable service now, enterprise health. Parlance 📍 certified health. And to help check them out at this week. Slash today.
Thanks for listening. That's all for now. ──────────────────────────────────────────────