Show Notes
A podcast about cloud adoption and cybersecurity.
Website: https://www.globaltech.tv/
Social networks: https://linktr.ee/globaltechtv
View Full Transcript
Episode Transcript
[00:00:03] Speaker A: So hello everyone. Again, we are back to our special episode of Global Tech TV. We are doing the summarize of 2024 and going forward to 2025. We were talking about a lot about cyber security in the first part and also beginning to touch the cloud. And Raz, we will start with you this segment, second part of this episode. What do you hope to see in 2025 in the cloud industry?
[00:00:34] Speaker B: Okay, so first of all, thank you for the break. I took a nap and come back. Okay, okay.
[00:00:40] Speaker A: We are not in 2025 by know.
[00:00:43] Speaker B: Okay, okay, I thought it's a break. And now in 2025, just a spoiler. We just need a break for two minutes and come back. Give it last. Okay, so let me focus on how do I Hope to see 2025 in the cloud. So for me, cloud is so complex. There are so many different clouds. Our aws, our gcp, there are Azure, for God forbid, there is Alicloud. No, I'm just joking. There is Alicloud in this side of the world and a few more. And we already forget about the private clouds and other infrastructure that guess what? Lots of companies are using them. Yeah. They are not just using the cloud. So what I'm hoping to see in 2025 is seamless multi cloud infrastructure or framework on tools across platforms that kind of make easier not just for specialists like you, Ariel. And every time that I have an issue, I need to call you to Spain, which is, you know, the time difference killing me.
I wanted to make it easy like my, you know, any other technology that I'm using. This is number one. Number two, for God's sake, we have AI, let's try to optimize it. AI driven optimization, advanced AI tools assisting with the cost reduction with performance with management tuning and practically security. What I'm asking, you know, I'm very humble.
Yeah, yeah.
[00:02:23] Speaker A: Very simple.
[00:02:23] Speaker B: I'm not asking big, big things to hope for. Yes, I'm not gonna be disappointed.
And the last one, surprise security. Please do zero trust. You know, by, by design. Yeah, stay. Do some encryption and encrypted data, make it like easy to adapt and then it's going to be much more easier to mitigate the growing of the cyber security threats that you spoke about them before. You know, and if you can do that, I'm going to love you, Ariel.
[00:03:04] Speaker A: Okay, so, so we will stop you right now, Raz, because there are a lot of people I believe in this session. What about you engine?
[00:03:15] Speaker C: I think that for this year I would love to see more training something that we haven't mentioned is that we should really invest on training the human factor, training our teams. Raz really highlighted wonderfully how one of the most integral parts of an organization and of security as a whole is a SOC team. I would love to see more SOC teams out there gaining more attention and gaining more value and knowledge, which comes of course from the organizations where they employed. So I think that my advice and hope of course for the 2025 year is to invest on our human factor, on our workers and our hackers, on our professionals insecurity. If they are trained and not only replaced by new workers, this would not only gain value to them but to the organization as a whole.
[00:04:07] Speaker A: Okay, so I believe that I am going around myself all this talk and for me is to really have smarter decisions for the customers, for the vendors. Try to do, you know, simpler products or helping like we really touch in this call, helping this customer, not just selling. If I may ask for something that will happen is to have some kind of simplicity around the migration to the cloud.
There was a friend that talked to me about a startup that they think they say okay, please do it because it will save us a lot of time and companies a lot of money. So I hope that maybe we will see simpler tools that will help us in 2025 or May Eyal, please, I would like to hear you so I.
[00:05:03] Speaker D: Don'T want to do spoiler alert for what we're going to talk about in this episode later on, but a lot of efforts were done in 2024 by the hyperscale cloud providers to integrate AI capabilities in cloud services such as Azure Copilot, Amazon Q and Google Gemini. And although the use of Genai has many potential benefits for automation and insights about store data, I expect the hyperscale cloud providers not to neglect the efforts put into cloud infrastructure so far. Meaning invest in embedding security, resiliency and observability from day one in all new and existing cloud services and decrease the service cost for the benefits of their customers. Things such as egress data or inter cloud network cost. There's no much reason for us as a customer to actually pay for this. It doesn't mean it makes sense.
And for many years we've read about customers struggling to take their first step in the public cloud. You mentioned it say migrating legacy workloads into hybrid architecture, etc. I'm hoping that in 2025 we will see more and more organization from all sizes and industries using the benefit of the public cloud. Things such as scale, almost infinite resources, managed resources and other stuff and investing in architecting and building cloud native applications. Meaning raise the bar whether customers are building new mobile application or whether they are exposing new applications that are using gen AI services.
No doubt the public cloud is the right place for the innovation and I'm hoping to hear more and more cloud native applications, more about cloud native applications and less about hybrid environments. Organization failing into Falling into the rabbit hole of cloud repatriation due to lack of proper planning. So it's all about proper design, proper architecture, proper planning. Don't fall into this hole of cloud repatriation. It's a rumor. You don't really see large enterprise actually moving back to the on prem maybe from some workload, but not as a as an industry shift. You don't see it.
[00:07:27] Speaker A: Yeah, yeah, there's a lot of blah blah talking a lot of people that have interest. But yeah, I think that the key to this thing is planning and really understand what you are doing. A lot of people do things I see because others say and others do and there is, I think there's a kind of lack of knowledge that people don't admit that they have and it affects organizations in the in the bad way. So I have a question for you.
If we finish now 2024 and we are running to 2025, what is the key word that we need to talk about, please?
[00:08:12] Speaker D: You. You mean the key word that no podcast can actually live without?
[00:08:18] Speaker A: Yeah, as I hope that you say AI but I don't know.
[00:08:24] Speaker B: So I had to say 2020.
[00:08:30] Speaker C: That.
[00:08:33] Speaker A: Yeah right. As always was. You are right. And let's start with AI.
What do you think that are the most important announcement that we have NG in 2024?
[00:08:49] Speaker C: Well, I. I have to agree with you about AI. It has been the key word not only for this year, but I think it's going to be yet another trend that's going to keep on following us. I think that the best and most wonderful example, we saw it only a few days ago. Just a few days ago OpenAI had a problem and shut down just for a couple of hours where we could really see how it affected our lives. It seems like in just a few months AI has been our friend, our virtual assistant, our go to for every task that we perform. I don't only mean this while we're working, but also in our daily lives we could see just in a couple of hours the whole world seized. We are getting so used to have all our tests automated and having right there next to us an assistant So I think that AI made waves in cybersecurity throughout 2024. It had notable advancements redefining how we defend it against threats. This is now shifting more towards cybersecurity. We had Google Cloud which introduced AI powered tools that enhanced Zero Trust framework. So we had IBM Swetson X which expanded its capabilities to provide predictive insights for vulnerability management. Another company that we are proud of was Sentinel One AI's driven platform which pushed the boundaries of autonomous response. We're having so many news from so many companies that not only took AI as a technology, but they also implemented it and I can say that elevated it into their organization and products. We've had mitigating attacks before they could even spread. AI has helped of course, all of us professionals in cyber security if used for the right intentions I might add there. And on the other side we have had the misuse of AI by all of the attackers. People who not only try to, they try to fool AI by questioning backwards, by trying to find new ways to leverage all the vulnerabilities, by trying to find easy ways on how to hack, by trying to exploit low hanging fruits. And what better than AI can help us do this fast tasks, especially for us.
Of course they've had usage of AI, especially with very sophisticated deep fake phishing campaigns and also automated automated exploit discoveries that they are really becoming challenges for organizations and of course for individuals in general. I think that all these developments have underscored the fact that while AI has offered immense potential for cybersecurity on the other side, it has also amplified the needs for, I think, implementing more ethical oversights and more robust countermeasures.
[00:12:04] Speaker A: Yeah, the better person in this room is Angie for sure.
I'm thinking about becoming a hacker, etc. And then the ethical comes from you, Angie and AI. Al, please show your insight, please.
[00:12:24] Speaker D: Okay, so in 2024 we couldn't really pass a week without hearing something new about AI or Genai into that integrated into existing applications. New large language models, LLMs supporting billion parameters for improved performance and complex task handling, to name a few. GPT4O, developed by OpenAI is estimated to have over 1,175 billion parameters. Google Gemini Ultra, released in February 24. Earlier this year it reports to have up to 1 trillion parameters.
Llama 3.1 from Meta, the latest open source LLM launched in June 2024. They boast 405 billion parameters. So it's a huge number. And Nvidia became the hardware manufacturer that everybody is looking to hear announcement about about its latest technology. However, as the GPUs cost more became more and more expensive, analysts are saying that Nvidia has around five big customers. I can assume it's aws, Microsoft, Google Meta and X and and since other customers in most in most cases cannot afford to purchase the latest GPUs due to the high cost. So it's something we need to really think about in terms of where the technology is going.
And in 2024 we realized that we cannot really rely on technology to produce good and quality outputs for humans. More and more cloud providers began to embed guardrails into their LLM engine to lower the risk of prompt injection, data manipulation, hallucination and abusive prompts. The technology is still in early stages and we really need to invest in responsible AI to generate safe and accurate data. So it's something I really think about both for the current state and for future think about Genai okay so thank.
[00:14:34] Speaker A: You before going to your Raz I was a little late in this AI adopting let's say I see I saw that many people were running again to AI because it's AI and I I don't like to be a ship you know to to go after every everybody but I took the your advice Raz and I started playing with the ChatGPT today I built my first or yesterday is I was helping a startup and we we did it with the the deck with open AI.
It's amazing you know but at the end in touching also what they all said it's not perfect and and you know you cannot from one trust these tools I think also today until the end they still need the the human insights and the ideas and and the rechecking but I am seeing you know more and more the things that I can use like NJ also you said in the day to day it help you but again I am looking for more be more I don't know exact with the tools and the things that we have this is just you know the things running into my head and what I can focus and saw that I was holding a podcast and I think that the CEO of Google said that 30% of the code I don't know if we touch it in previous podcasts maybe we did that write the code of Google and it's amazing from one side I say okay it's really amazing from the downside we need to really see how we can use this amazing technology. And again there are thousands of tools and they are changing almost every day week. So this is from. From my part but I will be more active. Right. Thanks to you in the, in this area in 2025.
[00:16:46] Speaker B: Let me give you one more insight.
[00:16:48] Speaker A: Yeah.
[00:16:49] Speaker B: Think about that. The gen AI is the perfect generalist and I'm not saying it because I have the podcast the generalist when you can subscribe over there in your free time. I'm saying that because you see when you are specialized in something they I have difficulties to give you the relevant things that you're looking for because when you try to unpack or how the Gen Z like to say. Unboxing. Unboxing. Unboxing, unboxing. Yeah. In TikTok it's, it's have difficulties to give you the relevant result. But if you want a high level and you want to have an assistant that can be generalist in so many domains, this is the relevant assistant that you're gonna have. This is like the first step to enter new research or new information, but in order to go even deeper. Maybe this is not the address that you're looking for, but it's a good start that you will not be able to do that without the AI. So anyway you can do that, but it will save you time.
[00:18:01] Speaker A: Thank you.
[00:18:02] Speaker B: Okay.
[00:18:04] Speaker A: Thank you.
[00:18:05] Speaker B: Things to thought think of. Okay. You wanted to ask me what I'm. What I'm. What I'm thinking about AI in 2024 or.
[00:18:14] Speaker A: Yeah, I. I will replay with you Raz. Start with the 2024 and go up to the 2025. What do you hope to see in 22?
[00:18:22] Speaker B: Okay, so in 2024 I'm gonna speak about the good, the bad and the ugly. Did you see this movie?
[00:18:32] Speaker A: Yeah. Okay, great movie.
[00:18:34] Speaker B: So the good with AI it's actually assist security to be more driven, to be more enhanced to have this co pilot, to have this sentinel. One thing to have these agents to have better response and it's not yet mature but it still give us a way to think how things will be better. To be better. Good for the ugly. The adversarial AI risk growth meaning the attacker also leveraging AI. They're targeting using phishing generative AI to to drive different malicious activities. So they're also leveraging that and the ugly. Some of people will say ugly, some of them not. Regulatory growth. Yeah, regulatory and all the compliance thing it's good now they have more strict AI compliance frameworks and demands and enhancement. So some people it's going to be ugly for them but some people will say we need to have these regulations and compliance in order to grow on the right direction to 2025 makes sense.
[00:19:49] Speaker A: Okay. Yeah. In 2025 will be also the ugly and the ad. Or not.
[00:19:54] Speaker B: So the second part of that movie, I'm not sure there is a second, second part. I don't have the name. I, you know, I only have the first version of the movie, but I can think about a Disney version for that which can be, I don't know, whatever you want. The Beauty and the Beast, maybe. But let's speak about how stronger AI will be with security. There is so many people, so many entrepreneurs, big companies that taking two different direction. One direction, how we can leverage AI to improve and have a better security.
AI systems. This isn't one thing. And second, how we can improve ourselves to defend against AI. So we're gonna see more stronger AI capabilities. That's what I'm, I think and hope to see in 2025.
From other directions, I want to speak about more ethical and safety, more transparency when using AI. So not every, Sorry, more for my French and you're in France. Not every idiot will say, hey, I have like an AI and I'm doing, you know, the magic. No, you have an API, you build an application. On top of that, you put a chat or prompt and you say you have an AI. You don't have AI. Okay. And I'm expecting to have more transparency, transparency in the infrastructure, in the models, in the data set they are using, and actually be more compelling and more compliance and accountable for the thing they're offering to the market. So all the, the things that we, we are, we want to have more trust and see there the right things in terms of AI. And the last one, I want to see more adoption.
So organization and vendors we spoke about before, adapt more the tools, be more vertical. Yeah, I know you like vertical because you change your camera a lot, you know, during the, the session, I want.
[00:22:09] Speaker A: To make it interesting. Right, I know.
[00:22:11] Speaker B: And you just want to give me a spoiler to speak about verticality. So be more vertical. Try to understand where you specialize and double deep on that and leverage AI. It'll make your life much more easier. That's what I'm hoping to see for 2025.
[00:22:30] Speaker A: Great, thank you, Raz. And what about you? What did you hope to see in 2025?
[00:22:35] Speaker C: Well, I think that not only hoping, but I would love to, let's say, guide my answer towards a more advices I think that I could give to specific professionals in security on how to use AI in 2025 so that AI is not, it doesn't have to be that shallow in the sense that you don't only need to ask IT questions that regards specific topics or just to be your go to assistant for just a task that might concern of a few seconds. I think that AI can be used in more ways than I have highlighted just a few of the ways that I personally use AI in my daily workflow and I would love to share that with others who perhaps might find it beneficial. I think that I would start with using AI to have an automated vulnerability detection. This can be simply used. It can really help hackers to identify weaknesses in systems way more fast than they would manually for sure. Of course, in addition to that, I would love to for people to use AI for advanced penetration testing. This can be a prime scenario where AI can use tools to simulate e.g. sophisticated attacks, mimic real world scenarios and also improve overall organization defenses. On the other side, I might add that AI can be used for predictive threat analysis. It can analyze patterns, it can predict potential attack vectors. It can enable proactive measures against emergency emerging threats. I have so many of these advices. AI is so versatile in my opinion. It's here to follow us not only throughout 2025, but it's here to follow us in case that we are mindful and use it ethically. I have to really emphasize that part. As security professionals, we really need to be mindful and ethical on how we use AI. It's really fast, really powerful and it really can grab all the information that we can prop onto it. And in, in addition to that I can find the positive parts of this. As always, I would love to stay positive in my advices. We can use AI to enhance security awareness trainings. This is really, really crucial and really nice use of AI where hackers can leverage AI to create realistic phishing stimulations, for example. Also they can use it for adaptive training programs and all of these programs in the matter of a few seconds or minutes. They can be tailored to each end user behavior and two less advices. So I do not bore all of you. I would love to add AI can be used to for sure to collaborate through AI driven insights. You can share autonomous insights from AI generated threat analysis. You can help organizations to collectively improve their security as a whole. And to top it off, in the end, I would love to add that of course after what you mentioned, Ariel, as a professional in the field, you can use AI to very well build smarter tools. You can cut down on the time that you invest into building these tools. You can have more insights, more sources that you use. And I think that ethical hackers should really be mindful and implement this last part. They should really use AI to build smarter tools. They could develop AI power tools that can assist in monitoring, in detecting and in my favorite responding to incidents in real time.
[00:26:24] Speaker A: So yeah, thank you very much Angie. Maybe we will take it for another short episode sharing more advices around AI or AI security what to use Also for your Raj I would like to see me using more AI to be smart on the things really using AI also for myself but maybe also for the companies.
This is something that I will for sure talk in the next management meeting that we will have because I think that again for myself I didn't use enough and something that I read about using more agents and less Genai and something that is running on my mind. I would like to see what is happening in this area. Again this thing focus really on companies. Eyal we will be the last one with this question and also with this episode. So Please what about 2025? What would would you like to see in this AIA area?
[00:27:31] Speaker D: Okay so in 2024 we most of the use cases we saw were around for using Gen AI was around chatbots such as ChatGPT, Cloud or Perplexity.
They're good for text on summarization, creating images or video and for AI powered coding assistant things such as GitHub, Copilot or Amazon Q developer which are currently which currently there is not enough data to say if code is generated by those tools actually makes developer life easier or simply generate more bugs that needs to be fixed. In 2025 I'm hoping to see other use cases of Genai things that will pass the proof of concept stage and will produce something that will benefit humans and not more of the same concept with different reps. So this is like my point of view about the current generation of gen AI and also the current technology is nearest limitation in terms of memory requirements for LLMs and cost of GPUs. In 2025 I'm hoping to see more and more cloud providers and hardware manufacturers breaking Nvidia's monopoly in the GPU industry by building their own dedicated hardware which will increase the competition in the market and hopefully will decrease the cost for supporting AI applications.
My point of view on what I'm expecting to see in the AI era.
[00:29:10] Speaker A: In next year okay anything smart has to say in the end of you you like to listen so you listen to everybody. Thank you guys. To finish this this episode I just.
[00:29:23] Speaker B: Want to wish everyone a good health be able to using the relevant tools on the relevant time and also know how to do some breaks because breaks are super important for mental health and physical health and this what I'm hoping for 2025 because this technology can kill us. So yeah we should, you know, we should think for the best and be able to take some time to relax and that's it, I guess. Happy New Year.
[00:29:59] Speaker A: Oh yeah. This is what I like to hear, Raz. Happy New Year to everyone.
And yeah if you can rest and take a while and be with your friends and family, it will be great.
Thank you Eyal. Thank you Angie. Thank you Raz and thank you for every person that listened to us make us like follow and all the other things. You are more than welcome to follow us in every social media that we are global tech TV and please feel free to ask and give comments or even criticize us if there is a something for so Happy new year. Happy 2024. Happy happy 25 and good luck to everybody. Bye bye.
[00:30:44] Speaker B: Happy New Year.
[00:30:46] Speaker A: Happy New Year. Bye.
