Transcription complète de On Call Sysadmin Podcast #14: Secret AI, Floppies, and 256 Core processors

fair. 05:55.514 --> 05:58.394 You don't get fired by kicking the can down the road. 05:58.394 --> 06:00.634 You get fired by a project going south. 06:00.634 --> 06:01.034 So. 06:02.074 --> 06:04.554 Well, until it breaks and then who's to blame there? 06:04.794 --> 06:05.514 That's true. 06:05.674 --> 06:05.954 Yeah. 06:06.394 --> 06:06.954 100. 06:07.114 --> 06:07.354 Yeah. 06:07.954 --> 06:09.034 That's a really good point. 06:09.354 --> 06:20.314 Well, they're saying that for the new system, the air traffic control system, that a little over 1/3 of the systems were understandable just because they didn't have parts or funds to replace any of my gosh. 06:20.314 --> 06:21.154 I mean, where do you even go? 06:21.154 --> 06:22.074 Do you go to eBay? 06:23.144 --> 06:24.744 To find old floppy drives? 06:25.384 --> 06:25.984 You have to, I guess. 06:26.024 --> 06:27.064 You have to. 06:27.064 --> 06:27.384 Yeah. 06:27.384 --> 06:32.984 I mean, I guess that's the importance of donating your old e-waste. 06:33.784 --> 06:34.344 True. 06:34.464 --> 06:34.624 Yes. 06:35.704 --> 06:36.184 Or you could. 06:36.184 --> 06:37.784 your local air traffic control. 06:37.984 --> 06:38.144 Yeah. 06:38.424 --> 06:42.904 Or you could sit on it and try to sell it, you know, years down the road and make a fortune. 06:42.984 --> 06:43.864 That's true. 06:44.024 --> 06:44.424 Yep. 06:45.224 --> 06:45.384 Yeah. 06:45.384 --> 06:50.744 There's sometimes there's a bathtub in the bathtub curve and the price of 06:51.914 --> 06:55.274 server infrastructure components really expensive to begin with. 06:55.434 --> 07:04.154 They go down to near 0 after a short period and then maybe decades later, your investment will pay off. 07:05.994 --> 07:07.194 The new Beanie Babies. 07:07.434 --> 07:08.074 That's right. 07:11.234 --> 07:14.194 Maybe I need to start hoarding zip disks now for when that comes up. 07:16.634 --> 07:18.794 Is that what they're upgrading to from floppy to zip? 07:20.784 --> 07:28.504 I'm sure there's some critical infrastructure out there right now running on zip disks that in 10 years they're going to say, oh crap, we need to replace this. 07:32.504 --> 07:43.944 Well, here's another one that I saw this week was workers are increasingly looking for ways to hide their AI usage in the workplace. 07:44.584 --> 07:47.264 And not good at all. 07:47.264 --> 07:48.264 And this was 07:50.144 --> 07:55.744 a study that was done looking at Google search terms in the last year. 07:55.744 --> 08:14.624 So from September 2025 to September 2026, and there were hundreds over 160,000 times where somebody was searched for something similar to unblock AI or free AI, no sign up. 08:15.714 --> 08:18.434 Or how to delete my chat history. 08:18.674 --> 08:30.434 No, So employees are using AI whether or not there are controls in place that say that employees are allowed to use it. 08:30.674 --> 08:37.394 And that's something that I think should worry, I think any organization, especially for free ones. 08:37.394 --> 08:38.834 Free AI, no signup. 08:38.834 --> 08:44.434 Man, can you imagine what those companies are doing with the data that you, the spreadsheets, the 08:45.504 --> 08:50.224 The medical charts that are getting dropped into those systems. 08:50.784 --> 08:52.304 Yeah, that's not good. 08:54.504 --> 09:02.624 You definitely want to obviously restrict AI to some capacity, but like you said, people are going to use it regardless. 09:02.624 --> 09:10.864 It's better that you allow them to use it based on your parameters as opposed to them going to these free things and just throwing your data everywhere. 09:11.264 --> 09:11.464 Yeah. 09:12.544 --> 09:13.824 I mean, it's so I mean, if. 09:14.344 --> 09:15.984 Your company hasn't thought about it so far. 09:15.984 --> 09:17.224 Now's the time to do it. 09:17.384 --> 09:24.824 They were saying in the study period, searches for things like this is up almost 40% in one year. 09:24.824 --> 09:35.224 So AI tools are here and employees are using it, whether or not you have official tools in place for you to do it. 09:35.224 --> 09:41.304 So I mean, this just goes to show that, hey, concerns around governance, security, 09:42.394 --> 09:49.194 data for things that you put into an AI chatbot are as important as ever. 09:49.914 --> 09:55.914 And so it's time to get a handle on shadow AI usage in the workplace. 09:56.714 --> 09:56.994 Yeah. 09:57.514 --> 10:11.114 And I would bet even if like you don't have restrictions in place, but you at least like openly allow AI in some capacity, I bet that would cut down on some of this like free usage. 10:11.724 --> 10:13.564 Yeah, it would. 10:14.284 --> 10:17.244 So, I mean, really, here's where it's at, right? 10:17.324 --> 10:29.084 I mean, you've got to provide some sort of tool for employees to use, or at least at first gain an understanding of what tools are being used. 10:30.924 --> 10:32.044 Or what they want to use. 10:32.124 --> 10:33.004 Or what they want to use. 10:33.004 --> 10:33.404 Yeah. 10:33.524 --> 10:34.684 Or think they want to use. 10:35.004 --> 10:35.404 Right. 10:37.164 --> 10:39.444 and then understand the use cases that they have. 10:39.524 --> 10:44.004 And then at that point, you can determine how are you going to govern the data that goes into AI chat. 10:46.324 --> 11:05.604 Because the worst thing is for an employee to sign up for a free account for ChatGPT or Claude or Grok or whatever it happens to be, or deepseek, heaven forbid, you know, these unregulated Chinese providers. 11:06.084 --> 11:06.804 And 11:07.744 --> 11:22.944 any data, chat, spreadsheet, document, anything that you put into those, your acceptable use or your terms of service, tell them that they can use any data that you put in there for training of the chatbot. 11:23.754 --> 11:34.114 And that's the last thing you want is some proprietary company information being used as a reply to a future query into the chat bot. 11:35.154 -->

11:35.394 Yeah. 11:36.914 --> 11:48.914 I mean, especially when it's things like, financial data or medical information or, trade secrets, things like that. 11:50.514 --> 11:51.074 So 11:52.524 --> 12:04.684 and that goes to user training too, because even if you do have a paid version of whatever AI you're using, I mean, you still need to know not to tell it certain things or give it certain things. 12:05.804 --> 12:07.364 Yeah, I mean, that's a really good point. 12:07.364 --> 12:16.444 Things that can put your company at risk, like, you know, that we had talked about this a couple days ago is that, you know, we're starting to see more 12:16.984 --> 12:40.984 usage of it for things like contract review, because hey, if I can have Claude review it for free instead of going to my attorney and having them, charge me $15,000 to look at it and give me notes, I can spend a few tokens or not pay anything to have it do it or ask legal questions where your 12:43.504 --> 12:47.784 an e-mail or a phone call to your attorney would have a client privilege, attorney-client privilege. 12:48.344 --> 12:52.424 Anything you put into a chatbot like that is discoverable in it that's already been decided on. 12:52.424 --> 12:58.744 You know, there have been legal cases already where they say, hey, you do not have attorney-client privilege with ChatGPT. 12:59.624 --> 13:11.064 And so if you go in there and you admit to some wrongdoing, knowingly or not, that's something that if you get a lawsuit, they're going to discover that. 13:11.304 --> 13:12.104 And 13:12.664 --> 13:16.464 that's going to come back to bite you. 13:16.624 --> 13:32.064 So you've got to have that employee training and at least get the buy-in and understanding on why it's such a big deal that you really shouldn't go around the safeguards and just start putting company data anywhere and everywhere else that you can think of. 13:32.184 --> 13:37.264 And also there may be government regulation around it as well. 13:37.264 --> 13:37.664 So 13:38.664 --> 13:53.224 if you for HIPAA, I heard a story was talking with somebody about this was a doctor's office putting medical charts into a chat bot and having it do a summarization. 13:55.384 --> 13:58.184 I mean, that's absolutely not right. 13:58.184 --> 14:03.224 You're that's a that's a blatant HIPAA violation, at least in my opinion. 14:03.224 --> 14:03.344 And 14:05.104 --> 14:06.624 can get you in some serious trouble. 14:06.624 --> 14:13.824 And that can be, 15,000 or up to $75,000 fine per incident, per time that you do that. 14:13.824 --> 14:34.144 So there's probably going to be some regulatory rules that are going to come down around some of this stuff because we're playing catch up on its capability and where it's used and where the data goes and the controls around it. 14:34.144 --> 14:34.544 So 14:35.194 --> 14:37.834 the right thing to do is just to get ahead of it. 14:38.234 --> 14:50.234 And let's look at some enterprise plan that is designed to keep your company data proprietary and go from there. 14:51.274 --> 14:58.394 Yeah, and that's a whole another discussion there is around HIPAA, because yeah, it's a thing, but it's not really an enforced thing. 14:58.394 --> 14:59.754 And even if they were to add 15:00.904 --> 15:04.464 AI requirements in HIPAA, it's still not going to be enforced. 15:04.464 --> 15:07.944 So you can easily get around and just say, yeah, we do that for sure. 15:08.344 --> 15:08.544 Yeah. 15:10.104 --> 15:14.664 Well, I mean, that's another there's been rumors now where they may start doing HIPAA enforcement. 15:15.064 --> 15:18.344 Yeah, but I feel like I've heard that for three, four years now. 15:18.344 --> 15:24.104 So I have to, but you know, it's it may get to the point where you know, there's a tipping point. 15:25.614 --> 15:30.774 where something happens and then, okay, we've really, really, really got to crack down on it. 15:31.614 --> 15:31.774 Yeah. 15:32.294 --> 15:35.654 But I mean, how is that going to go down? 15:36.334 --> 15:44.054 Is it going to be like CMMC where we build it up for what, three, four years and then they kind of backtrack? 15:44.654 --> 15:45.414 They change too. 15:45.414 --> 15:47.254 Yeah, I don't know. 15:47.254 --> 15:53.894 whether they'll, they may come in and do like spot checks or they may have, I don't know. 15:55.414 --> 16:02.174 Maybe an anonymous reporting system would be something that could come up like the IRS has, right? 16:02.174 --> 16:13.414 So I don't know how that would play out, but there's got to be something because I know it's HIPAA non-compliance is just ramp. 16:13.414 --> 16:19.054 That's a whole nother topic is just rampant in the medical industry right now, especially with private practice. 16:19.614 --> 16:20.894 Yeah, definitely. 16:24.604 --> 16:27.604 how about some new Outlook search functionality? 16:27.604 --> 16:28.924 Does that sound good? 16:30.444 --> 16:32.124 I like good news every now and then. 16:32.204 --> 16:33.084 Tell me more. 16:33.404 --> 16:34.964 Let's hope it's going to be good news. 16:36.524 --> 16:47.004 So yeah, Microsoft came out and said that they're working on improving Outlook search across the board because I think, you know, it's a widespread meme and it's understood that Outlook search sucks. 16:48.684 --> 16:50.764 So they're introducing a few things. 16:50.764 --> 16:53.284 One of the things is better relevance. 16:53.284 --> 16:53.804 So 16:54.314 --> 17:04.154 Right now, and I don't know for the past year or two, I don't know how long they've had this functionality, but when you search, it gives your top or the top three results. 17:04.954 --> 17:09.754 And then below that, it starts with all the results in chronological order. 17:10.314 --> 17:14.434 So they're expanding that top three to include up to seven now. 17:14.434 --> 17:22.074 And that's based on how recent the message was, the sender, previous interactions with those messages. 17:23.114

--> 17:28.994 So they're hoping to get more of what you're actually searching for in that top seven results. 17:28.994 --> 17:32.954 And then if not, you can drill down into the chronological list. 17:34.314 --> 17:44.154 They said they're still kind of tuning how much each thing lends to the relevance factor of those of the results there. 17:44.154 --> 17:50.634 So, you know, if how recent it was takes precedence over the sender, 17:51.024 --> 17:56.544 or interaction, they don't have a solid weight for each one of those things yet. 17:56.544 --> 17:58.144 So they're still kind of playing around with it. 17:59.264 --> 18:12.064 But as far as rollout for that part goes, it's available in Classic Outlook, Outlook on Web, New Outlook, even Outlook Mobile, and then Outlook for Mac is rolling out this month in November. 18:13.264 --> 18:13.824 Oh, good. 18:14.634 --> 18:22.394 I mean, I can attest to, I mean, I'm sure anybody that uses Outlook, how terrible the search functionality is in it. 18:23.034 --> 18:30.234 And I can, I'm can get completely behind getting better search results out of Outlook. 18:30.794 --> 18:31.354 For sure. 18:31.434 --> 18:42.314 Another, I think, simple thing that they're doing is if you search like the something or it something, it's 18:43.204 --> 18:47.444 Today, searching for the, so you get search results for the word the. 18:48.644 --> 18:58.164 So they're playing around with kind of removing that word from the search, I guess, unless you put it in quotations with whatever other words you're searching for. 18:59.764 --> 19:01.524 So that should improve things as well. 19:03.044 --> 19:07.284 Yeah, I mean, it kind of, I think it's kind of silly that was in there to begin with. 19:07.284 --> 19:11.284 And that kind of explains how bad or why a search was so terrible. 19:11.884 --> 19:25.244 I mean, I guess, I mean, sometimes you do need the word in your search, but I mean, for the most part, hopefully if you don't need it, you're not searching for that word because I don't know, my mind goes to searching keywords, not necessarily whole phrases or anything. 19:25.804 --> 19:26.084 Yeah. 19:27.964 --> 19:32.604 So, but they're also saying they're improving the overall speed and resiliency. 19:33.084 --> 19:38.524 So historically classic Outlook specifically has been using 19:39.344 --> 19:47.904 both cloud searching and the local Windows indexer search kind of in conjunction. 19:48.064 --> 20:00.384 So now they're trying to move more of the search functionality to cloud where there's more performance to actually do that searching directly on the server. 20:01.344 --> 20:03.664 And then if there's any sort of like 20:04.154 --> 20:15.354 slowness with your cloud connect or with your internet connection, or it's just taking too long for whatever reason, then it falls back to the local Windows index to pull those search results. 20:15.874 --> 20:16.274 Interesting. 20:17.754 --> 20:28.314 Yeah, they're saying that they're seeing speed improvements as much as 40% reduction in latency and from 4.5 seconds to 1.3 seconds. 20:30.354 --> 20:30.794 Nice. 20:30.954 --> 20:31.354 Okay. 20:32.044 --> 20:34.444 That's now available in classic Outlook. 20:34.444 --> 20:40.084 So go play around with search, see how much faster it is, if any, and hopefully it's better for you. 20:40.644 --> 20:41.484 Yeah, okay. 20:42.364 --> 20:48.604 And then I think the last cool thing that I think is interesting is in the calendar view. 20:48.684 --> 20:58.044 So you've always been able to search in your calendar, but the results show in just a regular list with the same kind of like highlighting in the word like you would in an e-mail. 20:59.314 --> 21:12.914 But rolling out in November for Outlook on Web and New Outlook, whenever you search in the calendar, it's also going to pull up results and highlight those results in the actual calendar view. 21:13.314 --> 21:19.554 So like the actual like squares of the event itself at the time, it'll highlight those. 21:20.754 --> 21:21.474 Well, that'll be good. 21:22.274 --> 21:22.554 Yeah. 21:24.114 --> 21:24.594 Excellent. 21:25.074 --> 21:26.834 Well, I needed some good news. 21:27.844 --> 21:31.124 I mean, it's good to see that new outlook is getting better. 21:32.324 --> 21:38.244 I know that, kind of when they first tried to roll it out and force everybody over to it, I don't think anybody was very happy with it. 21:38.804 --> 21:39.004 Yeah. 21:39.004 --> 21:41.284 I've been on it for about a year. 21:41.844 --> 21:49.524 And, you know, I'll say probably close to 99% plus of what I need to do is in the new outlook now. 21:49.524 --> 21:57.524 So, you know, I've gotten used to the simplified, I guess, if you'll call it that interface. 21:58.624 --> 22:04.944 But I think for the most part, 99% of the features I have are there and I'm glad they're making better. 22:05.584 --> 22:06.544 Yeah, I'd agree. 22:06.624 --> 22:13.904 I think I've been on, I forced myself on it about around the same time and I've been happy with it for a while now. 22:15.824 --> 22:18.864 Well, this is something that we finally have an answer on. 22:19.024 --> 22:23.104 You remember this over the last couple of weeks or month, maybe? 22:23.834 --> 22:32.194 where the grocery stores were sabotaged and they turned off refrigerators and freezers, which caused a bunch of food to spoil. 22:32.634 --> 22:42.034 And I think this happened in a bunch of French grocery stores, but also I think we had it at some military bases as well. 22:43.274 --> 22:52.834 the one specifically for the French grocery stores, France has officially come out and publicly blamed Iran 22:54.084 --> 22:55.844 for this attack. 22:56.724 --> 23:08.484 And it seems that they're calling out the IRGC specifically, as well as a few hacker front groups called Cyber Avengers and Cyber Insod Front. 23:09.204 --> 23:10.484 Cyber Avengers. 23:10.884 --> 23:13.964 Yeah, with a three instead of an E. 23:13.964

--> 23:15.284 That's how you know they're a hacker group. 23:17.764 --> 23:18.324 So 23:20.644 --> 23:38.324 And they actually specifically called out a few people as well in the IRGC, which, if you don't recall, they targeted about 700 food distribution sites across France. 23:38.804 --> 23:49.364 They remotely got in and they turned on some defrost settings and they caused food to spoil and fires and all kinds of stuff. 23:50.374 --> 23:59.414 where 10s of millions of dollars worth of food losses were accounted for. 23:59.414 --> 24:08.134 So now it seems like they have done some investigation and they're publicly blaming Iran for it. 24:10.894 --> 24:14.614 Well, I mean, that's good that they hopefully caught the people, but 24:16.214 --> 24:18.774 Probably not. 24:18.774 --> 24:19.094 Yeah. 24:19.174 --> 24:21.654 I'd say they're probably going to keep at it. 24:22.054 --> 24:29.054 I mean, this just goes to show, I mean, you need to don't expose your controllers to the Internet. 24:29.414 --> 24:36.694 I guess the better question is, are they rolling back all that or adding protection to make sure that their stuff is not exposed? 24:36.854 --> 24:37.894 Poor security. 24:37.894 --> 24:40.214 I mean, it's just right. 24:40.214 --> 24:40.894 It doesn't make it. 24:40.894 --> 24:44.294 It doesn't make it better that they have poor cybersecurity, but 24:45.794 --> 24:50.754 I mean, it wouldn't have happened if they didn't have all these controllers exposed to the internet. 24:50.754 --> 24:53.314 So that's the 1st place to go. 24:53.314 --> 25:03.074 But the 2nd is, how do you go back now and get justice for 10s of millions of dollars of losses and fires? 25:04.344 --> 25:09.344 whatever else I that I guess remains to be seen and what France is willing to do about it. 25:09.744 --> 25:10.024 Yeah. 25:10.624 --> 25:15.824 Although that lost frozen pizza, frozen corn dogs, man, what a waste. 25:16.144 --> 25:16.464 Yep. 25:17.184 --> 25:17.584 Jeez. 25:18.664 --> 25:19.144 Bastards. 25:20.144 --> 25:20.704 I know, right? 25:22.064 --> 25:26.144 Well, let's stick to the hack train. 25:26.784 --> 25:26.944 Yeah. 25:27.464 --> 25:29.904 Because another RMM went down this week. 25:30.544 --> 25:31.024 Oh boy. 25:31.144 --> 25:32.784 Who's time enable? 25:33.904 --> 25:34.624 They're a big one. 25:35.664 --> 25:35.944 Yep. 25:35.984 --> 25:41.264 And I feel like they had stuff recently before this too, in the past couple of months. 25:41.424 --> 25:48.784 I think they've been in the news for the last couple of, yeah, I think about the last month, they had at least two incidents that we heard about. 25:51.104 --> 25:53.264 But I don't think that any of them were this 25:54.024 --> 25:55.384 impactful, I guess I'll say. 25:56.344 --> 25:56.584 Yeah. 25:56.584 --> 25:58.184 So it all started November 8th. 25:58.184 --> 26:12.104 Security researcher Sean Bales publicly disclosed more than 30 alleged zero-day vulnerabilities, which he says he found with some AI assistance, you know, as the trend goes. 26:12.424 --> 26:12.744 Right. 26:13.384 --> 26:13.944 Naturally. 26:14.424 --> 26:16.184 But 30, that's crazy. 26:17.894 --> 26:20.774 But he went straight to releasing it publicly. 26:20.774 --> 26:30.534 He posted a proof of concept of these exploits on his GitHub repo called Able to Shell to follow the trend of the two shell vulnerabilities. 26:32.294 --> 26:41.334 He said he tested these vulnerabilities against a pre-release version of in Central, but it likely affects a lot of earlier versions. 26:41.414 --> 26:47.254 So he said he did not privately notify Enable before this because 26:47.824 --> 26:56.544 He supposedly disclosed other vulnerabilities I found to enable like 5, six months ago, and they did nothing with it. 26:57.344 --> 27:00.504 He tried to reach out to them and say, hey, this is a problem. 27:00.504 --> 27:03.584 Here's the proof that this is a problem, and they did nothing. 27:03.584 --> 27:10.624 So this round, when he found all of these, he just went straight to public and that got their attention. 27:13.484 --> 27:21.964 Enabled ended up taking down their entire, all of their cloud instances up in Central for I think a total of close to 48 hours. 27:22.924 --> 27:33.164 As of November 11th, they said that they finally finished patching all of their cloud instances and they have a patch release for anybody on Prem. 27:35.284 --> 27:37.004 Man, that's scary. 27:37.164 --> 27:38.684 Yeah, I wonder if they're, 27:40.414 --> 27:44.094 will be any reports on if any data was breached at that time. 27:44.094 --> 27:47.374 I can only imagine that there had to have been if. 27:48.254 --> 27:52.094 I mean, Enable is saying they're saying no evidence as of right now. 27:52.094 --> 27:53.694 But yeah, you're right. 27:53.734 --> 27:54.934 It might come to like that. 27:55.374 --> 27:55.774 Yeah. 27:57.134 --> 27:59.134 Well, fingers crossed on that. 27:59.134 --> 28:04.654 But hopefully this get will get Enable's attention and take, you know, bug. 28:05.464 --> 28:12.024 reports, bug bounty reports a little more seriously, as should any software company these days. 28:12.024 --> 28:25.064 I know that it's getting more and more difficult to keep up with the number of reported bugs, especially when you've got AI, they can just churn through and find them at a breakneck pace, but that doesn't mean that you can ignore it because it will not go away. 28:25.544 --> 28:32.504 Yeah, but I mean, if they're there to help you, to help you find that, it's better for you to find it in a good way instead of a bad way. 28:32.584 --> 28:33.064 So 28:34.214 --> 28:38.654 take these submissions to heart and actually fix it before you get impacted. 28:39.014 --> 28:44.934 Yeah, I will say I don't really agree with the way that Sean went about disclosing it. 28:45.974 --> 28:53.574 I think he probably at least should have given them a week and say, you know, an ultimatum is better than just dropping a

proof of concept. 28:53.654 --> 28:54.854 It was 0 notice. 28:54.854 --> 28:55.494 I mean, 28:56.614 --> 28:57.454 That's fair. 28:57.454 --> 28:57.894 Yeah. 28:57.894 --> 29:04.134 I mean, I get his side too that, he tried that with previous ones and didn't get any response. 29:04.134 --> 29:09.854 So I guess this was his one week ultimatum for the other ones in addition to the new 30 ones. 29:10.174 --> 29:10.414 Yeah. 29:11.134 --> 29:21.854 I hope that trend doesn't continue because I mean, at least give some responsible disclosure notice, even if they're not going to take notice of it, even if the company's not going to acknowledge. 29:22.884 --> 29:26.244 I mean, then you have something to fall back on and say, hey, I told you so. 29:26.244 --> 29:27.044 You didn't do anything. 29:27.044 --> 29:27.764 It's your fault. 29:27.764 --> 29:31.044 Rather than just dumping something out there. 29:31.524 --> 29:31.724 Yeah. 29:32.404 --> 29:35.364 Is that, would that become a legal thing at this point? 29:35.924 --> 29:37.604 If you publicly disclosed? 29:37.644 --> 29:39.444 Probably could be, yes. 29:39.524 --> 29:39.804 Yeah. 29:40.164 --> 29:41.444 I'd be worried about it. 29:42.164 --> 29:47.524 That is a CYA thing where you tell them first and then they come back and yeah, that makes sense. 29:49.604 --> 29:52.484 Well, in the same vein, Linus Torvalds 29:52.884 --> 29:58.964 who you're, you know, is the creator of Linux, the operating system way going way, way back. 30:01.444 --> 30:12.164 He's saying that, hey, maybe we're starting to see the peak of AI prompted patch discovery or vulnerability discovery. 30:12.164 --> 30:14.404 I don't know if you saw this, but 30:18.774 --> 30:37.454 Let's see, was it OpenAI, I think, did a initiative where they took the Linux kernel and they trained their model on it and went and found some vulnerabilities in just the kernel, which is like the very, very core. 30:37.454 --> 30:40.094 It's a very small portion of the operating system. 30:40.414 --> 30:43.294 It's what does the lowest level tasks. 30:45.534 --> 30:46.814 And they found 30:47.774 --> 30:56.334 948 working local privilege escalation vulnerabilities in just, I think, a month or so. 30:57.374 --> 31:02.174 And so that's something that is, first of all, that's a crazy number. 31:02.734 --> 31:11.214 Second, I guess it's good that these things are getting found and then OpenAI is spending money to figure out patches and how to get those things fixed. 31:12.734 --> 31:16.654 But I think we, or at least Linus thinks 31:17.044 --> 31:26.204 that maybe we're at the peak of this and AI has started to discover the bulk of vulnerabilities in software. 31:26.204 --> 31:30.164 And once they get patched, maybe we'll start to see that number come down. 31:32.444 --> 31:35.844 And so I think this is a big maybe, right? 31:36.164 --> 31:44.124 Because as the models get smarter and smarter, they're going to start to discover more sophisticated ways to break software. 31:44.124 --> 31:44.404 And 31:44.934 --> 31:53.334 But on the other hand, if we start fixing the things that are blatantly wrong, maybe fewer things will be started in the future. 31:53.334 --> 31:55.334 So I would think so. 31:55.334 --> 31:55.574 Yeah. 31:55.654 --> 31:58.054 I don't know if it's like the peak. 31:58.054 --> 32:02.134 I feel like maybe if we are at the peak, it's going to be a semi-long lasting peak. 32:02.614 --> 32:02.814 Right. 32:02.894 --> 32:05.094 But plateau and then trail off. 32:05.254 --> 32:05.734 Right. 32:06.454 --> 32:06.734 Yeah. 32:06.734 --> 32:10.134 I mean, I think this is something I was talking about with you just briefly yesterday and. 32:10.614 --> 32:27.254 So I, of course, have no basis for it other than, the thought experiment of, hey, this just can't go on forever in, at least, in the model of how software works in my head. 32:27.254 --> 32:27.494 But 32:28.694 --> 32:30.054 Only time will tell. 32:30.534 --> 32:49.014 But what all these vibe coded stuff like is that going to immediately be patched for any of these found vulnerabilities or is it going to be using old, old code that is still vulnerable and then other AI is going to scan that, find the vulnerabilities and we're just in this like this circle. 32:49.334 --> 32:49.934 Endless loop. 32:49.934 --> 32:52.454 Yeah, I think you're right for 32:52.864 --> 32:58.144 Applications, I think the runway for this type of thing is a lot longer. 32:59.264 --> 33:05.904 But I would say hopefully for core systems going up, let's get Linux kernel, right? 33:05.904 --> 33:06.464 Let's get it. 33:07.184 --> 33:07.824 Start the bottom. 33:08.184 --> 33:21.024 And then, yeah, and then let's get to the actual operating system and then it's windowing interface and then, work our way up from there to applications, which, of course, everything is going to be 33:21.954 --> 33:23.234 Under the under the AI. 33:24.194 --> 33:26.994 for at least the foreseeable future. 33:26.994 --> 33:34.914 So yeah, I'm hoping that at least for core systems like that, maybe we've hopefully seen the peak of it. 33:36.834 --> 33:39.154 The rest of the year is just going to be anything to shell. 33:39.714 --> 33:41.554 Yes, that's probably true. 33:41.554 --> 33:41.794 Yeah. 33:42.834 --> 33:49.234 Well, this was something I thought was curious was that VMware 33:50.294 --> 33:57.334 is reintroducing their low end server virtualization platform and they're calling it vSphere Express. 33:58.054 --> 34:02.774 So this is for systems that run on up to 128 processor cores. 34:03.334 --> 34:17.814 They really didn't fit into even like the essentials, the old going way back to the old VMware, VMware essentials where it would be like you're limited to three nodes and six processors, six physical sockets. 34:20.014 --> 34:32.094 this they're actually limiting it to 128 processor cores and they're also limiting down the number of features that it has. 34:32.574 --> 34:40.374 So

similar back to where you just had the essentials are standard with an X number of servers. 34:40.374 --> 34:42.814 So instead of forcing 34:43.264 --> 34:49.984 into the high end products where you have vSphere Foundation Enterprise or vSphere Cloud Foundation. 34:52.064 --> 34:54.784 This is a more basic entry level product. 34:54.784 --> 34:58.544 So the way they announced it was like, oh, this is a new thing. 34:58.544 --> 34:59.744 Well, it's not really. 35:00.144 --> 35:05.344 They took it away, realized everybody threw a fit about it, and then now they're bringing it back. 35:05.344 --> 35:08.144 So, you know, I think on the one hand, it's 35:08.764 --> 35:19.324 good because people that are stuck with VMware, this gives them a way to have a lower cost subscription to stay with VMware. 35:20.044 --> 35:31.164 But on the other hand, I think Rotom has burned so much goodwill with VMware users that I don't know, this may be too little too late for anybody to accept. 35:31.724 --> 35:32.364 Absolutely. 35:32.764 --> 35:34.444 I think so. 35:34.444 --> 35:35.644 I think it's too late. 35:35.644 --> 35:37.244 I think there is an argument that 35:37.724 --> 35:48.364 if you are actually stuck with VMware and you don't have a good way to migrate off of them and you have to get something, this is I think a valid answer and a good answer. 35:48.404 --> 35:50.204 I'm glad that they're bringing something back. 35:50.684 --> 35:53.004 I mean, that's true. 35:53.084 --> 36:04.204 I think if I was in a position where I was not able to, if I was currently coming up on my subscription end and not able to migrate off in the near term, 36:04.684 --> 36:13.244 that may be a way to extend your runway out in a more cost-effective manner than having to go to the more expensive packages. 36:14.524 --> 36:19.964 I would still be after my next renewal planning on being off of it completely. 36:19.964 --> 36:25.804 I think just the trust in VMware or in Broadcom specifically 36:26.574 --> 36:32.254 as a vendor is just so completely diminished that I don't I think this is way too little too late. 36:32.734 --> 36:33.294 I agree. 36:33.454 --> 36:40.414 But is this like who's to say they're not going to start completely backtracking and get back to the old ways? 36:43.164 --> 36:46.044 I don't think so. 36:46.844 --> 36:55.084 I can't, I don't see a way where, especially I don't see a way where they would go back to perpetual license with maintenance renewals, things like that. 36:55.804 --> 36:58.684 I think it's, I think they're permanently stuck on subscription. 36:59.404 --> 37:01.564 Yeah, I agree there. 37:03.224 --> 37:11.864 that this is a cash extraction enterprise and not a long-term investment endeavor. 37:13.024 --> 37:21.064 Because I mean, the way that they are looking at this is that they even came out and said, hey, we're going to focus on our top 600 clients. 37:22.304 --> 37:24.184 Anybody smaller than that, we're just going to dump. 37:24.184 --> 37:28.024 We're going to get rid of all of our distributors, resellers. 37:28.444 --> 37:33.404 the entire sales channel just evaporated within months of the purchase. 37:36.364 --> 37:47.484 And then they get rid of the things like, individual use where they're free ESXI, the free virtualization platform, that went away. 37:47.884 --> 37:55.724 And then they got rid of the VMUG Advantage program where somebody who was using 37:56.814 --> 38:03.654 I mean, they would pay $200 a year and they could use any of the products for learning purposes on their own individually. 38:03.654 --> 38:13.294 And they changed that and said, in order to get access to that, now you have to be certified in our program in order to have access, which, I mean, that's just crazy. 38:13.294 --> 38:16.254 How do you get certified if you can't practice with anything? 38:17.534 --> 38:23.134 But then also, where does your pipeline from entry level kind of on-ramp clients go? 38:23.884 --> 38:32.844 if there's nobody using the entry-level product to grow into the big boy products as they expand. 38:33.564 --> 38:38.924 So it's like they've just burned every bridge in all directions. 38:39.084 --> 38:47.244 And I just don't see a way back for VMware, Broadcom in any meaningful capacity. 38:50.604 --> 38:51.404 I know that 38:51.824 --> 39:03.664 It feels like that's an emotional, reaction to all this, but I think it's also logical in that they've done the same thing with, computer associates when Broadcom bought them. 39:03.664 --> 39:07.104 They've done the same thing with Symantec when they were bought. 39:07.824 --> 39:14.784 They did the same thing with, what was it, Avago, any of their products like that. 39:14.784 --> 39:20.864 It's, they are maximal cash extraction, do as little 39:21.924 --> 39:28.084 research and development, fire all of the salespeople to reduce cash expenditure. 39:28.244 --> 39:31.044 Like this is, that's, this is just Broadcom's MOO. 39:31.764 --> 39:36.884 Well, I mean, their response right now by releasing this vSphere Express seems kind of emotional. 39:36.884 --> 39:39.204 Like, are they panicking right now and realizing that? 39:39.204 --> 39:41.204 I think it's an, I think it's internal panic. 39:41.204 --> 39:41.524 Yeah. 39:42.404 --> 39:45.764 I think they're like, oh shoot, we went too far. 39:45.764 --> 39:46.804 How can we correct? 39:48.794 --> 39:49.794 But it's too late. 39:50.074 --> 39:50.754 It's too late. 39:50.914 --> 39:51.114 Yeah. 39:52.434 --> 39:53.474 You done screwed it up. 39:57.234 --> 39:57.474 Yeah. 39:57.474 --> 40:04.794 So I think it was a fairly large miscalculation on their part, at least in the long term. 40:04.794 --> 40:08.114 Now I know that the revenue year over year is just massively increased. 40:08.514 --> 40:17.874 Obviously, when you've got a capture audience of people that cannot leave immediately, you're going to have some of that, especially when you're increasing prices. 40:18.544 --> 40:21.744 5,

10, 20x over what they were paying before. 40:23.824 --> 40:32.704 But that's not a long-term sustainable, in my opinion, long-term sustainable business model. 40:32.704 --> 40:37.824 If you want to keep people on your product or gain new users, I can't imagine anybody 40:39.514 --> 40:44.554 who has decided that, hey, maybe we should look at VMware. 40:44.634 --> 40:47.954 I don't think any of that has happened since the Broadcom acquisition. 40:47.954 --> 40:53.354 And I cannot imagine anybody has decided to migrate to VMware platform. 40:53.914 --> 40:54.114 Yeah. 40:54.634 --> 40:59.434 And I'm sure those big players that they're betting on, I'm sure they're still planning to migrate. 40:59.474 --> 41:04.874 And, you know, years down the road, yes, but at some point I'm sure they're thinking about that. 41:05.844 --> 41:15.684 and especially when, back in, August, VMware had some tools called VDDK, which was a migration tool. 41:16.564 --> 41:21.684 It's like a way to get your virtual disks out of VMware and onto other platforms. 41:22.644 --> 41:30.724 They pulled that and they pulled it from public, they pulled it from their clients, and only partners now have access to 41:31.954 --> 41:36.514 that tool, which is crazy because there are barely any partners any longer. 41:36.834 --> 41:36.994 Right. 41:36.994 --> 41:38.114 Nobody's a partner now. 41:38.314 --> 41:38.434 Yeah. 41:38.874 --> 41:39.194 And so. 41:39.794 --> 41:41.074 Such a jerk move. 41:41.394 --> 41:43.474 I know they're saying, well, people are leaving so fast. 41:43.474 --> 41:47.954 Let's pull the tools that people are using to do it and which is forced now. 41:49.354 --> 41:57.954 There's open source projects where there's one called Open VDK that somebody has released open source. 42:01.054 --> 42:06.414 which allows the same functionality for VMs to get migrated to another platform. 42:06.894 --> 42:10.854 And funny enough, it's way faster. 42:10.854 --> 42:19.774 It's more than twice as fast as doing a real apples to apples comparison on migrating a disk over twice as fast is impressive. 42:21.054 --> 42:25.934 So that just goes to show how much effort was being put into these tools to begin with. 42:26.974 --> 42:28.894 Well, I'm sure that their tools haven't been 42:29.484 --> 42:30.924 developed on in years. 42:31.404 --> 42:32.364 yeah, yep. 42:34.884 --> 42:36.004 Broadcom is definitely not. 42:36.004 --> 42:37.564 And is this Open VDK? 42:37.644 --> 42:40.364 I'm sure it's mostly AI. 42:41.404 --> 42:42.284 You know, I'm not sure. 42:42.284 --> 42:45.004 I think he had been working on it for a little while. 42:46.364 --> 42:55.644 I can only imagine that really any software development going on these days has some sort of AI going into the development life cycle. 42:56.764 --> 42:58.124 But I'm not sure that it's 42:58.624 --> 43:00.704 specifically says one way or the other. 43:03.344 --> 43:04.784 But if you're looking for... 43:07.424 --> 43:11.824 Yeah, and the interface looks pretty bog standard AI generated. 43:13.184 --> 43:23.184 But yeah, if you're looking on a way to get off of it and on a Proxmox or some other platform, open standards platform disk, this may be the way to do it now that VMware has pulled those tools. 43:24.224 --> 43:26.904 There's even a V2K migrate utility. 43:26.904 --> 43:27.744 That's nice. 43:31.194 --> 43:31.674 Very cool. 43:35.114 --> 43:46.794 in the same vein of industry behemoths, AMD is on track to overtake Intel as the largest x86 manufacturer. 43:47.434 --> 43:53.354 And I think that's largely driven by data center CPUs. 43:54.904 --> 44:12.664 And what they're saying this last year, AMD's revenue is up 97% year over year in the data center market, which is just insane to double your market share or not double your market share, but double your revenue in one year is crazy. 44:12.824 --> 44:23.664 And they're saying they are looking to or expecting to jump from 13.4 billion in revenue to 44:23.874 --> 44:26.034 17 billion of revenue next quarter. 44:26.754 --> 44:37.714 So yeah, that's a pretty big jump and just goes to show, I mean, Intel forever. 44:37.714 --> 44:43.314 Nobody thought Intel could fall from the top position and it just seems that they have. 44:45.754 --> 44:48.834 They had a luck for their money back in what Athlon XP days. 44:49.234 --> 44:50.034 Yeah, that's true. 44:50.074 --> 44:50.754 That's true. 44:52.114 --> 44:55.434 they just couldn't keep up with when bulldozer CPUs came out. 44:56.354 --> 45:02.274 Kind of that era is just terrible performance and they couldn't compete in the high end, at least in the desktop market. 45:02.834 --> 45:08.554 And so yeah, that just took them a while to come back with the Zen line of CPUs. 45:08.554 --> 45:10.994 And this has been a good number of years. 45:11.314 --> 45:16.674 Zen 6 is just released this week, yesterday or so, they've formally announced. 45:17.314 --> 45:19.394 I think large data centers have been on it for a good while. 45:19.924 --> 45:35.044 testing out, finding the bugs, looking at the performance, but 256 core processors are now in the market, which is just insane, or for CPU, for server CPUs. 45:36.884 --> 45:45.764 But they're saying though they lost, sales on the gaming segment, 45:49.084 --> 45:51.244 I think they were down 8% in that segment. 45:51.244 --> 45:54.484 And I would say that's probably due to the price of memory. 45:54.484 --> 46:02.084 Just everything is so much more expensive that those buyers are going to be way more sensitive than the data center segment is. 46:04.164 --> 46:08.964 But despite that, their sales are just way up. 46:08.964 --> 46:15.844 And I think AMD is at least expecting to surpass Intel next quarter is I think what they're saying. 46:17.254 --> 46:20.534 That'll be an interesting read if that happens. 46:21.014 --> 46:21.174 Yeah. 46:21.574 --> 46:24.934 Are they ever going to get in the SOC market for like phones? 46:26.134 --> 46:27.734

Yeah, that's a really good question. 46:27.734 --> 46:42.694 I would wonder that they may have to come in and license, get an ARM license to be able to do that unless they want to get into something really obscure and go the RISC-V architecture. 46:44.774 --> 46:47.654 Yeah, we'll see. 46:48.054 --> 47:04.214 I mean, I know that they're kind of got their hands full with the AI market, GPUs and data center systems to fulfill those, that demand. 47:05.334 --> 47:10.534 And I'd be willing to bet that it probably not for this foreseeable future while they've got 47:11.654 --> 47:14.614 The gigantic cash cow that is AI in front of them right now. 47:15.094 --> 47:16.494 Yeah, that makes sense. 47:16.494 --> 47:19.974 That's where all the money is going to go to mobile phones and stuff like that. 47:20.614 --> 47:20.894 Yeah. 47:21.334 --> 47:23.574 I mean, they've been expanding into other segments. 47:23.574 --> 47:30.294 I think they've just recently gotten into the networking, the high end networking part of it. 47:31.894 --> 47:37.814 And they acquired Xilinx, which is a FPGA provider. 47:38.294 --> 47:39.094 And they've got 47:41.814 --> 47:47.734 kind of some bigger systems around doing machine vision and things like that in those segments. 47:49.334 --> 47:54.854 But I think it'll be interesting to see what happens over the next year or so. 47:55.654 --> 47:58.294 But I think they've been in a, they're in a really good position. 47:58.294 --> 47:59.054 I mean, they deserve it. 47:59.054 --> 48:04.214 Their products have outperformed and been cheaper than Intel. 48:04.854 --> 48:11.254 And so I think they've put themselves in a really, really good position and hopefully they can maintain 48:13.054 --> 48:15.214 maintain their upward trend. 48:16.134 --> 48:18.174 Yeah, I'm rooting for them. 48:19.694 --> 48:26.814 Yeah, I just wish we could see their products more easily available in commercial systems for desktop. 48:27.214 --> 48:39.134 I think that's one thing that I've had, there's just not as many laptops or desktops for consumer or workplace than there has been Intel even. 48:39.844 --> 48:43.364 through this uptick and uptrend in their market share. 48:43.364 --> 48:50.644 So that's one thing that I personally would look forward to is hopefully a lot of this stuff trickles down into the consumer market. 48:51.204 --> 48:55.684 Well, and are you talking like OEM builds or like build your own kind of stuff? 48:56.004 --> 48:56.804 OEM builds. 48:56.804 --> 49:02.404 I think AMD has been very, very strong in the DIY market. 49:03.144 --> 49:15.544 for a number of years, just because they've been on top, I think probably the last three or four years, they've kind of been the top dog for desktop CPU performance and in laptop as well. 49:15.544 --> 49:22.264 There's just not as many of them as have been Intel systems. 49:22.264 --> 49:27.544 And I think part of that, I don't know if you remember this from a few years back. 49:27.624 --> 49:32.904 I think this may be going pre-COVID, maybe 2018 or so. 49:33.524 --> 49:48.004 it came around that Intel was paying Dell under the table to exclusively use Intel CPUs in their products. 49:48.644 --> 49:58.364 And there was a huge multi-million dollar, I think it was $100 million judgment finally after everything, after all the dust settles that AMD won against Intel to 50:04.444 --> 50:11.164 settle that for those underhanded payments to Dell. 50:11.484 --> 50:15.964 And that's why we didn't see those systems from that OEM for a while. 50:16.844 --> 50:22.924 And I wonder if it's just taken this long for them to get the development for an AMD system 50:23.534 --> 50:26.654 back through the complete engineering life cycle. 50:26.974 --> 50:30.974 And now we're starting to see them more and more, at least from that, at least from Dell. 50:31.374 --> 50:31.854 So. 50:32.814 --> 50:33.014 Yeah. 50:33.214 --> 50:36.574 So really it's a more of an OEM problem, if anything. 50:37.134 --> 50:37.454 Yeah. 50:38.334 --> 50:48.374 And you know, part of it may have been, you know, can AMD manufacture these CPUs in enough quantity to be able to meet the demand of a large OEM like that? 50:48.374 --> 50:51.494 And that's why it maybe stayed in the 50:52.244 --> 50:56.324 smaller segments like gaming desktops and DIY market. 50:56.324 --> 51:05.724 And so that may have been part of it, but underhanded payments from Intel to keep their products exclusive surely didn't help. 51:06.724 --> 51:06.964 Yeah. 51:09.204 --> 51:12.884 But anyway, well, thanks for joining us for, shoot. 51:17.364 --> 51:19.924 But anyway, thanks for joining us for another shift. 51:20.314 --> 51:24.474 One thing we've learned over the years is that technology works best when you've got the right team behind it. 51:25.034 --> 51:34.234 And if you've got questions about your IT strategy, security, or business technology, connect with us at harvestitsolutions.com or through the link in the description. 51:34.714 --> 51:35.914 We'll catch you on the next shift. 51:36.954 --> 51:37.234 See you. 51:37.914 --> 51:38.314 Bye.