*DataPoint* Accelerating AI with Python-native Ray and the Importance of Open Source in AI

All right, hopefully you can hear me. Hopefully this is going to work. Figured I would share this. A lot of the folks I follow on LinkedIn, they, um they do some cool stuff. They go to some cool places like Dubai, Monaco, et cetera. Shout out to you, Lauren, for the inspiration. And I figured I would stream today because I haven't streamed in a while this month. My bad. Totally on me. A lot of family stuff going on. I'll explain

another time in the future. But said family stuff brings me to this wonderful place in basically rural Pennsylvania, somewhere between Pittsburgh and Philly. And I am literally at an airbnb by a stream. I don't know how well you can see that, but there's a little fire pit there. Rock garden. I don't know what that is. But I figured I would stream by a stream because it's kind of a dad joke, I guess. And I got to clean this camera on this laptop. It's a bit of a dad

joke. And tomorrow is Father's Day, so I figured I would just share that. A couple of things. One, I want to thank everyone for the positive feedback I got on the latest edition of the Frank Diggs Data newsletter. And if you're not already subscribed on LinkedIn, please do so. Thank you for that.

Ray is a distributed computing framework for basically AI and ML workloads, more so than, I think, data engineering workloads. And it's one of those things where it says it's going to replace Spark. I think it will. Will it do that in six months? Six quarters? I think that's the only thing that is the big question. Obviously, it's a new technology, and I think that the fact that it's based on python is a definite plus. The fact that it can really kind of free people from

Pi Spark. I know it says sounds silly, right? Like free people from Pi spark. Like it's this horrible thing, but Pi Spark is not proper python. Yeah, I know there's been a lot of work done bridging those two worlds, but I also think that the problem space that yeah, I can't be here. Okay, cool. The problem space that Spark was meant to solve goes back about ten, maybe 15 years at this point. Same thing that happened with Hadoop. The problem Hadoop was meant to solve dates back to

the early 2000s. So these technologies will either adapt let me rephrase that the technology products themselves. These are open source projects. Open source projects tend to do best when they solve one particular problem in a particular way. I know I'm going to get a lot of hate mail on that, but ultimately what I'm saying is that let's focus on the problem we're trying to solve, not necessarily the tools that we're

using to solve them. And I think Ray is definitely positioned to be an excellent future facing thing. Hey, my LinkedIn comment. Hey, Boris. How's it going? Let's see. Let me get the chat overlay going. There we go. Usually I don't stream from the

MacBook. I'm on the MacBook now. It's a MacBook M two, not an ultra. I got this last fall to replace my personal laptop because I am falling out of love with Windows because of Windows Eleven, but that is something I've ranted about on multiple live streams, so I'm going to drop that there. So as for me, I'm enjoying my stream side streaming. I should make that a thing, but ultimately I think that check out the article, let me know what you think. What do you think is going to be? I know

there's another technology. What prompted me to write it was actually a question that came up at work the other day, and I had seen something come across either my Twitter feed or my LinkedIn feed about another framework that is claiming to be like the next spark, right? The next big data processing framework. I still think Ray has more on it, not the least of which is the Python native features.

And if you look at the work that's being done to increase speed in Python, whether that is some of the latest optimizations, or Mojo, which is basically a kind of like TypeScript what TypeScript did for JavaScript, mojo is doing that. So you're going to get a lot more performance enhancements depending on who you talk to. The performance enhancements are ten to twelve X, or in some extreme cases,

thousands of times faster. I think those are extreme cases, but ten times optimization, ten times speed improvement is going to be a big deal, particularly when you're dealing with neural network learning and all the computational power that goes there. If you marry that to a big data distributed processing framework, well, then think about what that open up. I just noticed my hair is a little wild. It's time for me to get a

haircut. So speaking of AI and some AI goodness, I want folks to know about this upcoming AI webinar. It's a joint webinar with Red Hat and intel where we're going to talk about AI application benchmarking. I know one of the speakers, she's awesome. I don't know the other speaker, but it's going to be a joint presentation between Intel, Red Hat, and we're going to talk about how OpenVINO and Red Hat OpenShift. Data science can be used together to optimize your data

workflows. All right, I think I have another comment. There we go. Oh, that was me. I love web based software because it's so easy. Browser based streaming software or any software. The updates are all automatic, so sometimes the user interface, they've changed that. So just want folks to be aware of that upcoming webinar. There's another webinar. I think it's invite only that I'll be doing next week. It's for IBM. Watson X IBM partners only, I think. And IBM employees.

I'll be speaking on that, on the power of open source artificial intelligence and how the partnership between IBM and Red Hat can really accelerate your data endeavors. There's another webinar going on that I do have, the QR code, but it's not pre loaded into the system. You have to forgive me. I've been traveling with the kids business. Travel alone is a challenge, but traveling with the family now that they're out of school adds an extra level of

fun. However, I do have the privilege of having my extra production assistant here. Let's go and meet her. Hey, you want to say hello? Are you camera shy? You camera shy? Come on. There you go. She's a Weimer on her hi. I know, she's very excitable. Yes, the link to the meeting webinar. Let me post that again. Thank you, Thomas. Hope things are well. There we go. You can scan that QR code. Let me hide that. Scan that QR code.

I'll post it in the comments. I think I already did mention this one in a previous LinkedIn post. Restream has really got to up their game because I don't think you can type. Oh, wait, it looks like you can. No, LinkedIn comments are read only. Still. Oh, well, what a problem to have, though. I mean, I remember when doing a live stream required, like, specialized hardware. Even before that, it was even more

specialized hardware. When I was doing live streams for the K Street office, we had to get a Tricaster, which is a $4,000 machine. And that was the cheap one, right? That was the mini one. But now I've done this on my phone through a browser, which is just absolutely phenomenal. So for me to complain and I can't send a text message to LinkedIn from here is kind of funny, right? It's about the psychology of human expectations. But seriously, Restream is an

awesome platform. If they added that, it would be the perfect platform. So with that, I'm going to thank you, Thomas. Thank you, Boris, for commenting. I really appreciate it. I'll leave that QR code up there for another minute, but let me know what you think. Do you think I should put more content up on my LinkedIn newsletter? I do have a sizable subscription there. I am going to work on building my own email list because I just think that's just a good thing to do.

Anyone has advice on what platform to use, please let me know in the comments below. Definitely looking for some advice on that. And there'll be a new episode of Data Driven coming out. Now that my oldest son is out of school for the year, I'm going to make it a summer job for him to learn how to edit podcasts. Figured it'd be a good skill for him to have, so hopefully I can get the backlog. We have a lot of great

shows. Seriously. I know you're thinking I would say that, but Andy and I shout out to you, andy brother from another mother. We've been just impressed with the quality of guests that we've spoken to. We've had some really smart people. Some one famous person, actually. One famous on the Internet person, anyway, who's not in the database. Actually. She's an entrepreneur. Awesome. Lauren Tickner. She's cool. I've admired her work for a while. She's very kind of no nonsense, kind of a

cool straight shooter. Very funny. Very smart, too, in terms of how she's built her brand and her business. I like this street. I mean, it's not the Burj Khalifa that I'm looking at, but we all do what we can with what we have. Who knows, maybe your future livestream will be over there. But in all seriousness, check out this webinar. I think this is a very exciting time to be in artificial intelligence. It's definitely an inflection point.

I think Chat GPT has really put gasoline on something that was already pretty well lit. And it's funny because 18 months ago I was saying that there was going to be an AI in the LinkedIn profile where I talk about how only quantum computing can save us from AI winter. Obviously, I was proven wrong. There's been a lot of innovation. I did not expect Chat GPT to come in 2022.

That was a type of technology and thing that I would have anticipated towards the middle of the decade. I knew it was coming, I just didn't think it would be coming this fast. So just goes to show you, even if you're an expert in the field, even if you're in this, this has surprised a lot of people, and obviously there's a lot of drama around Chat GPT. Some people kind of dismiss it as, well, the latest fad, or it doesn't do anything more than your phone does when you do the predictive

text. And while that's technically true, right, that is the equivalent of comparing a paper airplane with, I don't know, an Airbus A 380, or a Boeing Triple Seven, or an F 35 fighter. Whatever represents the state of the art for you in aviation. It's technically true because they are based on the same type of principles, right? Air, lift, thrust, drag, all

that sort of thing, gravity. But they are very different in terms of their implementation and skill and the amount of just technology in it, right. The amount of maturity. That's probably the word I was looking for. Fortunately, I need more coffee, basically. Unfortunately, this airbnb actually has a nice curry machine. I know there are a lot of curry haters out there, but you can't beat push button coffee. Do you sacrifice quality? Yeah, I guess you do.

But let's be real, if I make it from a drip cup, it's not going to be impeccable quality either. I'll save the coffee shops and the baristas to make the really good coffee, but for me, just the average kind of day, I'm cool with getting the curry cup.

So with that, I will definitely encourage you. Check out that webinar. Great speakers, great technology. And this is really the time for AI to shine. And if you look at all the ethical concerns in artificial intelligence in all of this stuff happening, open source is now more important than ever in terms of transparency, in terms of sharing innovation, sharing risks, right? Understanding these risks to society, that open source mentality, right?

That sounds bad, but that open source mindset, that's a better word is needed now more than ever, because this technology is going to affect each and every one of us, right? Obviously, for those in the industry, it's going to affect us first. But if you look at what Google is doing, I think when Wendy's replacing the drive through, when you make the order, it's going to do that through Google. This is going to impact everybody, right? From kind of entry level jobs to even middle

layer, white collar jobs. This is going to have a profound economic impact. Now, will it be barred? Will it be chat? GBT? Will it be OpenAI? Who knows? But the fact is, it's here. It's now. A lot of governments are trying to regulate it. Good luck. The cat's out of the bag. It's here. And the best way to approach it is through a community minded approach that's similar to open

source software. That mindset of let's be open, let's be transparent about what we're building, how we're building it, and make sure that we run the technology and the technology doesn't run us. So with that, I will end this live stream by a stream. And this kind of stream of consciousness thing was good enough. I might actually make this a podcast episode as well.

So if you are listening to the audio of that, just imagine a peaceful, tranquil stream in the middle of the country with a beautiful fire pit and a quaint little lodge type thing going on. That's where I am. I should have painted that picture early, but maybe I'll Bailey. Do that for me. So with that, I'm going to end the stream. And if you're wondering who the heck is Bailey, bailey is the artificial intelligence that we

have helped run the show. Go, that does a lot of the intros and the outros for data driven and impact quantum. Speaking of impact quantum, season three is in the works, I promise. So with that, I'll end the stream and you have a great day. And I probably won't stream on Father's Day, so if you celebrate Father's Day, have an awesome one and enjoy the rest of your day. You do have.