Unveiling Watson X: IBM's Next-Generation AI Platform

0:00What can 160 years of experience teach you about the future? When it comes to protecting what matters, Pacific Life provides life insurance, retirement income, and employee benefits for people and businesses building a more confident tomorrow. Strategies rooted in strength and backed by experience. Ask a financial professional how Pacific Life can help you today. Pacific Life Insurance Company, Omaha, Nebraska, and in New York. Pacific Life and Annuity, Phoenix, Arizona. As more time passes since ChatGPT was launched, we're seeing more and more veterans in the tech and AI industry finally start to compete with different AI tools.

0:38And one of those veterans that has finally thrown its hat into the ring in a really big way is IBM, which recently has launched what they're calling Watson X Studio, which essentially is developed to help create and make generative AI easier for enterprises. So today on the podcast, we're going to dive into what exactly WatsonX is. But first, I wanted to talk a little bit about the history of IBM and Watson, because I actually think this is kind of interesting. Back in 2011, IBM unveiled another tool called Watson, which essentially was a question-answering computer system, like really similar to ChatGPT, but it wasn't trained on the traditional transformer model.

1:21It was a little bit more archaic in the design. In any case, it was developed by IBM's Deep QA project, a research team led by David Feroci. So Watson was named after IBM's founder and first CEO, which is Thomas J. Watson. So that's what this original version was, and I'm assuming that's what this new Watson X is. Also, I'm assuming this new Watson X, the name is just, you know, paying homage to this earlier version. But in any case, the reason this original Watson system was famous is because it competed on the TV show Jeopardy in 2011 and it won. You know, this does kind of look like a publicity stunt for IBM and Watson.

2:01But apparently there was another piece of software that IBM made a long time ago that won some chess tournament. And so now they were kind of looking for the new thing and this is what they did. So, yeah, back in 2011, it competed. it beat brad rutger and ken jennings winning the first place prize of a million dollars i'm sure the other two competitors were a little salty to say the least that you know they got that far and then someone put a computer system in there that beat them um and then interestingly enough after that right because it would appear as though this was kind of like a chat gpt version way back in 2011 this is one of the earlier versions but it wasn't really used for any commercial purposes until 2013 when IBM announced that they were going to use it to help with decisions in lung cancer treatment.

2:51So they, you know, they're using it for some interesting things. And then I'm not sure exactly what happened to it beyond that. It didn't seem to have, you know, made such a big splash that we now have Chai ChibiT. But in any case, that brings us to today when they have a new announcement with their new technology. So their new IBM X is IBM jumping back into this, you know, obviously they were pioneering this space and so they're going to be announcing essentially three different product sets in their generative ai toolkit for enterprises the first one is called watson.ai it's going to be the studio itself and also a foundational model library they also have watson.data which is a data store and then watson.governance a governance toolkit so we'll jump into what each of these do and why they are important so right off the bat i would say so IBM's Watson X is a platform where enterprises can train, tune, and then they can go and deploy their own AI models based off of their own corporate data.

3:51So it can train foundational models and machine learning models for cloud environments. And IBM curated and trained models are also going to be forming the backbone of the service. So there's also going to be like IBM has their own trained models in here. And this isn't the first company to announce this, there are other competitors. We're seeing other people like Microsoft Azure, Amazon's SageMaker Studio, and also Google's Vertex AI are all doing similar things to this. And so this is IBM getting into the ecosystem and really trying to deploy in a way that keeps enterprises on their software, on their AI studio from start to finish.

4:33So I'll go over the other things that they're doing, but it's just something that I found is interesting is essentially these foundational models and open source AI models are going to handle the gathering and sorting of training data, and then they're going to pass that data to the businesses that need it. So there's also a toolkit for ongoing AI governance, like we talked about, and essentially IBM wants to provide an end-to-end AI workflow that's going to let businesses go from not having any AI in their company to customizing and running all of it for them. And most importantly for IBM, keeping this all on their own platform so that you don't have to go to Microsoft, Amazon, or Google in order to do this.

5:16And then they can keep all of that cloud computing and all of that processing right on their own systems. They said, we built IBM Watson X for the needs of enterprises so that clients can be more than just users, they can become AI advanced, advantaged. So that was the chairman of IBM, one of their chief executive officer in their press release or whatever. But essentially, following this like end to end workflow model, their clients of IBM essentially are going to be able to build their own models from the ground up or adapt existing AI models that that IBM has on their platform. So the Watson X studio also includes tools for helping write code, similar to what we're seeing out of, you know, GitHub and Microsoft.

6:07So it has a whole bunch of these like custom built AI models for doing custom things. They have one for writing code, they have one called fm.nlp. That's essentially a collection of large language models tailored to specific industry domains, right? So like, they have, you know, really specific industry ones. And then they have one called FM dot geospattle, which essentially is just a model that uses NASA's satellite data to analyze weather, weather patterns and climate change. And so they have a couple, right, like these custom things, I'm assuming they have some sort of custom, exclusive deal with NASA on their satellite data that they use to process.

6:45So they're trying to find ways to use their own in-house custom data and create these little language models that would be useful to incorporate into different companies' own language models. So companies will be able to upload all of their own data, right, on the data store, and then they're going to be able to train their own models off of this, which I think is really powerful. And IBM is trying to make this something exclusive where you'd rather use them than others by coming up with their own AI models in here. And I think this is, it's something really interesting, because Google recently released a memo that said, we don't have a moat, and neither does open AI.

7:22And it's kind of like the fact, it's kind of like this interesting conundrum that everyone in AI is having right now, there's a whole bunch of these big tech players trying to develop these AI tools. But no, there's not really a moat on access to a lot of this data, like, you know, Reddit decided they're gonna start charging. So maybe that's like some sort of moat is the price. But in reality, people are going to get all of this data whether they pay for it or not and there isn't really a moat on the data it's publicly available and so it's going to be interesting to see what i believe is going to happen is all of these all of these companies are going to try to put moats by whatever data sets they have making them exclusive to their own products so if you want you know nasa's geo data you're going to have to go to ibm and use their platforms which makes sense it's pretty smart because like overall you want to use all of their products but it's gonna get ai is gonna get pretty commoditized where it's not going to be super different between different services and platforms for certain things and the only way that they're going to be able to draw on customers is with exclusive data sort of similar not far off from what you're seeing with disney or netflix right like netflix launched and it had all movies from everything on it and then as disney wanted to get into the streaming wars they cut off allowing all of their movies to be on netflix and it's exclusively on Disney.

8:39And, you know, we're seeing like Disney and Hulu and all these different platforms, try to make exclusive content for their platforms that you're going to go there, even though, I mean, by and large, it's all just movies, right? So in any case, this is what I'm going to see. This is what we're seeing with AI. This is how people are going to create moats in the future. Google, it's interesting because Google said we don't have a moat, but like in reality, if data was the moat, Google does have a lot of different data sets that other people don't. recently I was listening to a podcast and they were talking about the fact that Google's Google Bards capabilities now is able to read it has access to all the transcripts of all YouTube videos and so it's able to read the transcripts from YouTube videos and search those so you can say hey Google Bard pull up the YouTube transcripts from this channel and and tell me what XYZ person's opinion is on this thing.

9:34And it can go look at all their YouTube videos and tell you what their opinion is on something. Now, it's not incredibly accurate at the moment, but it will come up with responses. One bug I found when it was doing that is that it was mixing like two people in one conversation. It was mixing their opinions up because they're both in the same video. So not perfectly accurate, but you get the picture Google has access to that library, they could potentially in the future make it so you know, maybe they block everyone else from accessing the API to those transcripts. So it'll be interesting to see exactly what happens there.

10:07But in any case, back to Watson and what they're doing. A couple of their other major software products that they're including in this is Watson Code Assistant, AIO Ops, insights for greater visibility into IT operations, Watson Assistant and Watson Orchestrate for labor and customer service solutions, environmental intelligence solutions for EIS builders, which helps measure and respond to environmental risks. So at the recent 2023 conference, IBM announced the opening also of IBM Consulting Center for Excellence for Generative AI, man, that is a bit of a mouthful. I feel like they probably could have come up with a better shorter name for that.

10:46But in any case, essentially, what they're doing is they're bringing together about 1000 generative AI experts who are going to work in consulting for different businesses who want to help bring or essentially build and deploy WatsonX and AI into their company. So this makes a lot of sense, right? This new software solution is targeted for enterprises. It's going to be really powerful, but it's obviously got a lot of jargon, and it's going to need a lot of help from, I believe, these experts and these consultants to really help companies know and organizations know how they can incorporate AI into their company and how to do that in the best way, what best industry practices are.

11:26And so this is kind of another smart moat or competitive advantage, you could say, that IBM is offering by having, you know, a thousand, you know, really smart consultants that would cost a lot of money otherwise, slash they're hard to get a hold of because Microsoft, Google, and IBM are all snapping them up and paying really competitive salaries for them. This is going to be a big competitive advantage for people to want to get onto the IBM system because they have someone there that can hold their hand and help them work through the whole process that is an expert on this. So I think it's going to be left to where it's again, it's an arms race right now, it's gonna be really interesting to see who pulls ahead on this, whether that's IBM, Microsoft, Amazon, or Google, in this specific space, because we also have a couple other like newcomers like cohere and anthropic that are trying to get into this area as well getting people to build their own AI models on top of their platform.

12:18So this is going to be a space that's going to be heavily competed against. And we're going to follow it closely to see who the winner is over the next months and years.