Video: Generative AI Applications Summit: First Look

Hello and thank you for joining us. I am Steven Ramirez.

I’m the CEO of Beyond the Arc and I also have the pleasure this year of being the co-chair of the Generative AI Application Summit.

And with me is my co-host and co-partner in crime, Martin Sial.

Martin, if you would also be so kind as to just do a little brief introduction.

Yeah.

Thank you so much for setting this up, Stephen.

I’m Martin Sel, the CEO and founder of generative AI net.

And up until last year I was also the generative AI lead for EMEA at Infosys Consulting.

And now I am doing, I’m in the business building and get utilizing the generative AI wave and all of these applications.

Yeah.

To build value, to create value for people.

Excellent.

And I wanna thank people who are joining us.

We are going out live over several platforms today.

We are going to be broadcasting on LinkedIn, on Twitter and on Facebook.

And I know from our last experience on LinkedIn, if you are on LinkedIn and you are not hearing audio, then you may need to refresh the page to make sure that you can hear the wonderful things that we are actually saying.

We’re also going to have the comments that are gonna be live, Elise, I believe we are.

And so please go ahead and use that.

We’d love to get your participation as we go along.

And so feel free to make comments, ask questions.

We’ll try to, I think we’ll be able to pull some of those questions in as we go.

So let me give you just a little bit of a preview of the about 20 minutes or so that we’re gonna spend with you this fine morning.

So first of all, we wanna just talk a little bit about generative AI and really the process and the upside of building applications in generative AI.

So that is really gonna be where this conversation is going to start.

This is an incredibly fast-moving field with lots of experience now kind of coming together with companies and organizations all across the spectrum.

So we’re just gonna talk a little bit about that.

And then that’s a good lead-in to talk about the event that we are planning.

We’re going to be, as I mentioned, we’re a co-host for generative AI application Summit.

And so we’ll talk a little bit about the keynote speakers and something very exciting that we’re doing, and this is why I think you should stick around for this conversation.

Martin and I have put together what we’re calling the generative AI applications Accelerator workshop.

And this workshop is really hands-on.

And so in this session that we’ll be hosting in Phoenix as a part of the event and internal rules, maybe there’s a strategy how they want to answer certain emails.

And so their clients are asking lots of questions basically against these regulations.

And then internal employees have to craft the email.

Yeah.

They have to read through all the documentation and then answer it, understand what actually is the question, cross-read it and then send it out.

So this can take up to half a day.

Yeah.

To answer if it’s a complex ask.

And we have built a chatbot internally utilizing a RAG system, a retrieval-augmented generation system where all of the databases, all of these knowledge corpus is plugged in and through a semantic search algorithm, the respective relevant pieces of the various databases are then used in order to craft an answer by sending the context to a language model.

And in this specific part where you have sort of like the chunks and the knowledge pieces of the databases that are relevant for that specific ask, when you put this together, this is called then prompt engineering.

Here’s where you can really gain a lot of great.

Yeah.

Where you can really make great products by good prompt engineering.

And so like this, we have built them the chatbot that then answers the ask in literally under two minutes.

And the representative sort of like cross-reads it, maybe has asked another question towards that response.

There’s another iteration then but can then literally copy and paste it and send it to their client.

And we have saved up to 80% per iteration, 80% of time and effort.

So there are huge wins.

And this is what I think mostly drives the professional.

I think it’s super exciting.

And just to amplify I think a couple of the points that you’ve made.

I mean, so first of all, I have not really gone down the path with Copilot very much, but I’ll tell you, I was very interested in some reading I did over the weekend around using Microsoft Copilot with my tool of choice, PowerPoint, and being able to use a system that ingests documents, comes up with insights and, like everybody’s dream, makes you almost the near-perfect, not near-perfect but not bad, first draft PowerPoint in a matter of minutes.

I mean, I think that just the day-to-day application of generative AI, like if you’re not using it yet kind of on your desktop, I think that kind of workflow will just blow your mind, right?

Like anybody who’s, and the ability to go from, as you said, from Excel or Word or PowerPoint, right?

I mean, that’s really getting us towards, yeah, we’re gonna save this.

‘Cause that’s getting us towards the AGI future that I know that your book touches on.

But before we go there, so we’re talking about building applications.

We’re talking about some of the promise and what you and I have actually both seen in our work, but now we get to go to this summit where we are convening dozens of companies and in total hundreds of participants to come and tell their AI stories, right?

So we’re going to, on our agenda, we have companies that are really sort of leading the charge and actually have applications that they’re starting to build.

I wanna talk a little bit about who’s gonna be on the agenda, but before I do, let’s talk about this accelerator that you and I are creating because I think that I haven’t seen anything like it at all the conferences that I go to.

Just give us the quick kind of 60-second intro into what we’re hoping to accomplish.

Yeah.

So drawing from our consulting experience, especially with these gen AI solutions, we have sort of like a two-step approach.

In the first step, we want to, together with the participants, identify which ideas within their specific organization can be shortlisted to examine further.

And the shortlisting goes really with a hands-on approach.

Participants come also with access to a chatbot language, where it can be ChatGPT, can be Claude 3 Opus or Gemini.

And they try out with certain prompting rules how to solve a specific problem.

So in that way, we shortlist, let’s say, the top three, top five or maybe it’s just one idea.

That’s step one.

Step two, we then map this idea or this shortlisted idea onto a sort of canvas gen AI accelerator canvas, where we have the data layer, we have the search layer, we have a generation layer, and then an application layer, which basically is on top of everything.

Yeah.

Application layer.

And participants then, with guiding questions by us, answer the questions and know, okay, this is the data that I have.

This could be auxiliary data.

Do I actually need a search functionality?

What is the generation layer?

What do I need to take into consideration?

Open source, closed source.

There’s a plethora of things to keep in mind.

And then actually, what application do I plug it into?

Another as a white label, or am I using maybe like an avatar, or is it a chatbot framework?

A classic one.

So we are elaborating sort of like on a broad scale on all of this.

And also we have a share out, all of us together.

Yeah.

Which will be, I think, super powerful.

Yeah.

I’m excited.

I think that this is really kind of a, if you are in an organization and you’re thinking about developing, well, first of all, you’re just thinking about how to use generative AI and you wanna take it beyond logging into a web interface and asking a question, right?

That you probably have mastered.

But if you really wanna build that into business applications that are doing, as Martin talked about, this augmented search capabilities using generative AI to then not just search but then to be able to use those documents for the new creation of content and insights and knowledge in very specific business use cases.

Well, that’s what we’re doing, right?

That’s really what this workshop is gonna be about.

And in addition, so the workshop I think is a key part of it, but Martin, I have to say it was hard work along the way, but we wound up with a pretty good agenda of companies that are gonna be presenting at the generative AI application summit.

Are there any highlights for you?

Like who are you looking forward to hearing from at the summit?

So yeah, there are several highlights.

I think just chronologically but also like from the interest perspective, I’m very much looking forward to Julian Ziman from Hugging Face, the chief evangelist from Hugging Face.

I already heard and read a lot of him and he is always very much on the bleeding edge of this tech, especially when it comes to how to use open source.

And if you are not aware of it, open source is very rapidly catching up with closed source performance.

Yeah.

And so I have also worked with plenty of clients that have chosen open source because they have said, Hey, I can have this.

What closed source is cheaper.

I’m in full control of the language model.

There are no guardrails by some other institute.

And also like two, three months later I’m at the same performance, so why would I even like go with close source?

That is a valid argument as well.

So I know Julian is very much deep in the trenches and he, like he, it’s definitely a great experience to listen to him and a learning jump right away.

Yeah.

I think, and if you’re not familiar with Hugging Face, I mean, it really has become kind of like the go-to hub for all open source gen AI.

It goes beyond Gen AI, but certainly from the gen AI perspective, like what they’ve been able to do, I think is absolutely amazing.

I think also too, I mean, I’m excited we have Hitesh Waha from Google.

So the counterpoint to the open source is the amazing work that I continue to hear, and we know that Google has just immense resources devoted and have for years around AI of all sorts.

So to have a kind of key representative from Google, I think is also gonna be fantastic.

And maybe you could say a little bit about Matthew gi.

Give us a little preview of what Matthew’s like.

Yeah.

And also to Hitesh, as you mentioned just now, he’s not only working at Google, he’s also a gen AI ambassador.

So that’s really like his core work.

And as there is gonna be a Q&A session potentially afterwards.

Yeah.

Then I think I’m looking forward also to engage there on a conversation.

Yeah.

Matthew, I think on the second day, then let me quickly check.

Yeah, he’s, oh no, right.

He’s on the first day.

And he comes from the data engineering department and bridges basically the gap between data engineering and generative AI machine learning.

Also, yeah, from Ary, he together with Joe Rice, I was there on their LinkedIn Live roughly a month ago.

They’re having these Monday morning chats, quite well watched regularly.

No, but it’s also gonna be a super interesting conversation.

Yeah.

On what is happening also with data engineering consulting, PwC, so really kind of a nice mix.

And so we’re really excited about sort of seeing everybody and really having this opportunity to dig into that application development.

You know, I’ll just mention the website is on the screen.

We also have a discount.

We’re the chair, so we wanna see you there.

And so the discount code is scrolling across now.

Check it out.

I hope that you can join us in Phoenix.

We’re gonna just, as we sort of wrap up, Martin, before we kind of end today, I do wanna just ask you a little bit about the book.

So tell us the title and give us just a couple of key highlights.

Yeah.

Thank you for asking.

So I wrote Generative AI Navigate.

You don’t see it because of the blurring algorithm, but it’s Navigating the Course to the Artificial General Intelligence Future.

What I’m basically, I wrote it myself, published by Wiley.

And what this is mostly about is first I’m doing a very brief look on the history.

It’s not all of a sudden.

You can even see, we had the first chatbot actually in 65 by Joseph Weisbaum, 1965.

Wow.

It was called Eliza.

Super interesting.

And it was actually quite funny.

You were writing a sentence about, yeah, I like my family or whatever.

And then the chatbot picked up the word family and said, oh, tell me more about your family, you know, and basically just asked questions of keynote keywords.

So a bit more simple.

Yeah.

And then a little history lookout and look over the history.

And then what is currently happening in Generative AI, it’s not only language models, it’s also image generation.

It’s 3D object generation, super powerful Opus AI, by the way, is a very strong company where you type in certain text, chair with this, and it generates the 3D version of it.

Especially interesting for virtual reality, et cetera.

We have sound generation, music generation, avatar generation.

It’s really getting better and better video generations.

Like you have heard also about Sora, right?

Sora by OpenAI, which is really strong in generating videos.

Yeah.

And this is already 2K.

Yeah.

And not 4K yet, but 2K and can produce up to a minute.

Now think, if I think a bit further soon.

Well, Martin, before my current career path, I spent 10 years working in Hollywood in the movie business.

And believe me, if you think about just what it would take to create a minute of animation, and the fact that you can now do that with two sentences and it generates in minutes, right?

As opposed to teams of animators and render farms with like dozens.

And from where gen AI just sort of burst on the scene, maybe for most of us in the 18 months, I know that you are one of the OG, right?

You’ve been following this trend over the last eight or nine years from its very beginning, Martin.

And so it’s really exciting to now see this take shape.

I think we should probably leave it there.

I dunno if you have any final comment you wanna make before we wrap today.

I’m very much looking forward to the Generative AI Application Summit.

I think it’s gonna be a great energy.

We will also have a dinner together and the joint workshop that we’ll have together.

So I think we designed it in a way that there’s lots of cross, I’m not sure if this is the right way.

In German I would say it that way, cross-pollination of ideas and sort of like synergizing, have a synergetic field and environment.

I’m very much looking forward to that.

Highly excited.

Yeah, me too.

You guys have had the information.

I hope that you can join us.

And in any case, if you’re interested in generative AI, both for Martin and I, you’ll see on the LinkedIn side our profiles are there.

Reach out to us.

We’d love to connect.

We’re gonna have another sort of follow-up conversation in another week or two.

Hope you can join us then.

But for now, thank you guys so much for joining us.

Really appreciate you being part of this session and look forward to seeing you next time.