Microsoft’s Next-Gen AI Chip, Synthetic Music Debate, and an Instagram Impersonation Scandal | TSG Ep. 1007

Hey, everyone. Happy Wednesday and welcome to our Techron Gang Live. We've got a great show for you today.

There's so much going on in the tech world, it's crazy every day. It just doesn't seem to stop. The thing is, I sometimes wonder, where's Mike going to get our three segments for from, from today?

And then I, I wake up in the morning and bombarded with things that just dystopian sometimes. Yeah, amazing. Other times run of the mill some days, but today's a good day.

Let me introduce you to our gang to about what we're gonna be discussing today. We've got the woman in front of the brick wall lined up, Terry Robinson. We've got, she's joining us maybe for the first time this year.

I haven't seen you on since the New Year, Wiki Wang Wiki. Great to have you here. And of course, we have our man in the valley, John Schwartz, and our man in the Hightower.

Mike Ard. Welcome, welcome, gang. How are you?

Um, as I said, we live in interesting times. I don't know if that's an old Irish proverb. I heard it's an old Chinese proverb.

I've heard. I heard it's an old Jewish proverb, a Greek proverb. I think it's a proverb from everywhere that you live in.

Interesting times. Mike, Microsoft didn't have this on my bingo card, throwing their hat into the chip game. Yeah.

Uh, challenge to Nvidia. Well, I, I think as I understand it, this is the second iteration of their chips. And, um, what people are talking about though is this one looks a lot more serious than the first one and in terms of its capabilities and it's, uh, processing power.

But John, I'm gonna let you jump in here. One thing it does come to mind though, it's like, well, are people gonna start moving workloads away from Nvidia, the champion for other AI accelerators? What's going on?

Yes. Right. That's interesting you say that.

There's a second generation AI chip from, from Microsoft called Maya 200. So it's a, it basically can interpret it in different ways. The main interpretation, and we hear this already from some market analysts like BMP para buss, is they think that this faster, perhaps cheaper chip will reduce the reliance on Nvidia, Nvidia hardware and could lead to Marvell and Broadcom being among companies that could benefit from it and maybe wean them off of Nvidia.

Um, Microsoft put out a long detailed blog about its speed, performance, and cost. They claim it's their most efficient inference system with 30% improvement in performance per dollar compared to the current version. Um, now I'll add that not everyone thinks this is going to be a, a threat to NVIDIA's dominance.

Our own Daniel Newman, I, I spoke with him briefly yesterday and he thinks, he acknowledges that myo two hundreds a solid, uh, functional step up in homegrown silicon, and it's gonna augment Microsoft's compute. Um, but he doesn't believe it's gonna replace Nvidia or a MDI kind of fall, sorry, Daniel, but I, I just, I, I probably would disagree with that is assessment, because I do think this is a potential option or alternative to Nvidia, and I wonder if we see other companies following suits, um, which is great for customers. But, um, you know, this is a, this is a pretty bold move by Microsoft, just my main takeaway.

And it shows a less reliance on Nvidia, and I wonder if this opens the door. And Microsoft actually has been kind of taking the, the charge and, and leading the door, leading the, the charge, um, towards more efficient energy friendly pipes of hardware in chips, which is I find in a whole nother interesting development. George, I'm sorry, go ahead, Mike.

There's Two things to think about here. One is, so Microsoft is joining Google and AWS pushing people towards their own AI accelerators. And part of the reason for that is, well, it's less expensive for them, and two, it costs less to deploy these things.

They don't consume quite as much energy and they don't have a lot of, you know, graphics overhead that nobody needs for ai. So that's all the good stuff. The issue as we get to it though, is gonna be, you know, there's this thing called the channel, and this is where the salespeople are.

And what you can expect, and you know, you're already seeing is every time a customer shows up and says, I got a workload, I'd like to run on the cloud on Nvidia, the salesperson's gonna go, well, that's great, we can help you with that. But did you know that we also happen to have this other processor over here that costs less and is, uh, more readily available and we can get you on this processor in the next 30 days? Versus, well, that GPU might take us another 90 to 120 days, and how many times they have to have that conversation before all three of those cloud service providers start taking care away from.

I, Uh, you're, I'm glad you mentioned that, Mike, because Microsoft is so strong in the channel and there's this persistent shortage of, um, NVIDIA's GPUs. Mm-hmm. So in a sense, I mean, this was inevitable.

It was gonna happen. And, uh, Microsoft, Microsoft works with companies very well until it starts working against, but competing with them, and I think open AI is gonna, is already can acknowledge that. But, but guys, let me ask you a question.

Is Maya truly a GPU or is it an inference chip? Mm, it's more, it's more of an inference. So they're not going after the training side so hard, but there's not after that inference side.

Yeah, They're going after the inference market. So is Broadcom, so as you said, is Google, so is was it train, is the Amazon entry in Yes, yes. I think the inference market is wide open, quite frankly with a lot of compa competition.

But let's not get ourselves, Nvidia retains their stranglehold on the GPU training chip market, which is fi makes for a $5 trillion company that $5 trillion can open a lot of channels if, and one, they decide to go down that path. But here's the thing about Microsoft guys, even if it's not the greatest chip right now, Microsoft iterates learns, reiterates iterates again, and they get it better and they get it right. That was the Steve Balmer Yeah.

Microsoft, right. That was their method to their madness. They, they'll keep improving.

It may not be great today, but it'll get there. And I have no doubt it will be, it will be every bit a com competitor to the Google chips, which I believe Broadcom is making for them. It'll be every bit the competitor to the Amazon Traum chips.

It'll be every bit a competitor to the Broadcom chips. I don't think it's a competitor to the Nvidia chips. Well, I think really it has ambitions to go downstream my inference that are not gonna be as realized as they had Hoped.

Well, it's not gonna be a a, a toboggan run, that's for sure. Well, I have a question. 0, I was what came in to mind.

Yeah, actually maybe fun. Nevermind fun, but Whitney, you wanted to say something? Yeah, I just feel like it's very smart mo move for Microsoft, right?

Because, uh, chips mar margin is very high and they have a lot of needs internally if they use their, uh, their self, um, it can save lot of money from their side, especially for this one, right? It's delivering a 30% performance per dollar boost. I really like this number 'cause inference, uh, even though you said the competitive, um, market is very competitive, but I can see there are huge needs here.

Recently I saw there's two deal in the VC side on the inference. It's big deals. So Yeah, I I still see there's a huge market in this area.

Well, no, the, the inference chip market's going to be huge as big as the trading market, perhaps. Yeah. Bigger or bigger.

But where Nvidia has, you know, a stranglehold on that training market chip, you, you don't see a MD really is the only one that, and I didn't mention a MD in the inference. A MD is the only other one that it, you know, really has some GPU muscle. So, so let me ask you though, I mean, once they get a foothold in the inference side, won't all these companies go after the training side?

That's the next Logical? Well, no, but I think the bar, I think that's why we're seeing them all go for the inference side. I think the barrier to entry to the GPU training chips are a lot, is a lot higher than it is making these inference chips.

And I, and I think that that's where, and that's where NVIDIA has built their castle and moat. Yeah, I do feel like internally saving is huge here, right? Because they support open AI as well.

So their needs plus open AI needs, it's a, it's a good deal. Yeah. There's also a big shift going on here in terms of how these things are bought.

So the AI data science team usually drove the initial wave of these things and they like GPUs 'cause that was what they get their hands on and they usually attach some infrastructure person to that TIGER team, and that was the person who kind of helped them deal with the hardware issues. But as we deploy these workloads in mainstream IT environments, you know, the CIO is starting to exercise more influence over the infrastructure for inference. And those guys are looking at it saying, you know, why, why are we doing all these GPUs?

They're expensive. They don't run all that efficiently. They're suck at energy.

So I think you're gonna see the mainstream IT team really drive a lot of this inference conversation. And they're gonna drive maybe not two Nvidia, I think at, at, at, at, that's I think what Microsoft is thinking actually, Mike. Yeah.

He hit on offer an alter on alternative, um, to the IT guys. Sorry. Well Did, they're all thinking that way.

But you know, there's different, there's different, there's different categories of players here, right? You take Microsoft with Azure, you take Google with Google Cloud and Amazon of course with AWS they're thinking to themselves, I already have the house, I already have the real estate. Why should I put someone else's chips into my house?

I'll put my own chips into my house. I have more control. I I I make it harder for them to switch out to a different cloud competitor and I make more money.

So I do that. Then you have a different player, the a MD players, the Broadcom players who say, look, training is kind of had its day training's great for training these big, massive models, but the future doesn't belong to the frontier model. The future belongs to people using AI and apps and stuff using inference.

We could sell these chips to everyone. We could sell 'em to cloud providers even though they have their own now. But we could sell 'em to people who are in the private data centers, in the colo centers to the server market in general, who want to make inference boxes.

And I could sell it to everywhere. Now, Microsoft has a foot in both. 'cause they have that channel where they may be what, they traditionally haven't sold chips, but they've sold more laptops.

They haven't even really sold Microsoft built servers that I know of. Um, but Microsoft could potentially play there, but certainly a MD and Broadcom will have broad, no, no pun intended. We'll have broad, uh, support in that marketplace.

And then you'll have Nvidia saying, Hey, we're already doing the training in the GPUs. Who knows AI better than us? Our inference chips have a tighter integration.

So when you do need to cross over from training to inference in both, you want an, you want an Nvidia stock? And that's, that's how it works out, I think. Yeah, just there'll be one other thing at play here.

Um, these enterprise agreements that the cloud service providers give folks are all based on volume and they're gonna give them points for using their stuff more so than, you know, other third party GPU services that they may be offering on that same cloud platform. Yeah. I'm just curious right after, after like, um, Microsoft has their own trips, do they still try to buy from Nvidia?

Of course they do. They have to. They need the GPUs as well.

That's the beauty of this system. Everybody, you know, it's all Switzerland, everybody's going to, it's no exclusivity like that. But, um, here's the other thing.

Where, where's Microsoft? Where's the fab making these chips? Are we using Intel?

Microsoft used to have a pretty tight relationship with Intel, didn't they? Mm-hmm. Wouldn't Intel be happy about that?

Yes, Wintel baby. That's right. I think they will spread that around a little bit and they'll be, you know, Intel, Marvel, Broadcom, all these guys will be sources for chips.

Well, I think so, but I specifically Microsoft, but it, it's interesting. 18 or whatever it is. I'd like to know that there are fabs that can produce that quality of chips that maybe don't reside on the island of Taiwan.

And, uh, you know, are they based in the us? Are they based in Europe? Where, where are they based?

Well, They'll be coming outta Arizona sometime next decade, I think. February. Yeah, sure.

Like the 2030 timeline for all this stuff. Mm-hmm. All right.

Hey, we're about outta time for this segment, guys. Let's move on. Mike, we're talking about the music man or music men, or maybe a music bod.

I don't know. Well, we, we've talked about this, uh, emergence of so called synthetic music or AI music, whatever you wanna call it, on a couple of shows previously. But we have an article now up on digital CXO that kind of talks about the phenomenon in general.

And a couple of things going on. It's clear that, you know, people are buying this music and there's, uh, folks who seem to enjoy it or get some value out of it, but there's also other folks who are kind of basically saying, Hey, this is not a human experience. And they believe that music is about an expression of our humanity.

And I'll tell you this, probably last week, I was, um, a local brewery here, had a record night. People bring their records and they get the place. They a small selection of them.

There's kind of like this raffle. And, um, we listened to, you know, albums for, I don't know, maybe eight albums for the one side each for the, all these people brought. And it was this human experience.

And none of those people, you know, they're hardcore music fans. They're not ever gonna bring something synthetic, that's for sure. So, I don't know, Terry is in two different markets here for music.

Is there gonna be May, maybe? Uh, it's, I, I read this article and I was so disheartened because I, I guess partly because people didn't realize it was synthetic, you know? I mean, that, that's, that's one thing I'd like to think that I would, I mean, I love music and I'm into music, and I like to think about the human aspect.

I wanna know what Robbie Robertson of the band was thinking when he wrote the weight. You know, what inspired that? So what inspires AI to pin music, prophet, whatever, somebody wound it up and let it go.

I just, I think the ai, now, I didn't listen to all these songs, you know, or whatever, but I think the issue with AI too is it has no, it has no soul. I mean, when you read the, when it writes something, you can usually spot it, even if it's pretty well written, right? Because it uses things like this quietly happened, you know, or whatever.

You have these, these words. Um, but it, it, it lacks soul. It lacks that human experience.

And, um, I just can't help but thinking that would be the same with, with music. And I don't understand why so many people thought these artists and these songs were like real, I mean, I guess the songs are real. They're just not made by, you know, real people.

And, um, so that was, but, but I also think, you know, what's, what's happened here is that the one song, and I think it's appropriately named, right? I know you're not mine. Um, that seems relevant.

Um, that was Sweden's biggest hit so far this year, right? But I think the issue, you know, like with the Swedish government, everything else is the deception here that people don't know what it is. And I guess they've, they've banned the songs from its, uh, uh, from their official listing, right?

The Government of Sweden. Is there an obligation, I, I'm sorry, Sorry, is an obligation when, when you play this music, when whatever, wherever it's played to, to point out that it's not written by human. I think we should put a scarlet letter on its head or something.

Or do you Just have to read the minor notes to find out? You know, You guys, Hey, don't, don't blame the instrument. Blame the musician Right now.

I, so I, I have a totally different take on this. I had a totally different take on it. First of all, my God bless you, that you found a brewery that lets you to come in and play a half a side of an album at a time.

7 in my youth, where we used to listen to half side of an album at a time. But here's the thing, it's not fake. It's real music.

Well, it's real music. And you wanna know what a musician was thinking when he gave a prompt to an AI to create a particular song and what lyrics to include and, and put it all together. What better way than to have the actual prompt and to see what he actually told it to do, or told it to?

Do I have be A musician? Do I have to be a musician? Musician?

No, you don't. And that's the beau of, and that's the beauty of synthetic music. The same way we're all going to be co-developers.

We all could be creative. Now, I took guitar lessons for major parts of my life that my fingers bled, and I still sucked. Oh, Yeah.

I have No rhythm. Guitar, piano, all of that suck. But, But with ai, I could actually make beautiful music.

I can make beautiful music, but there's going to be a difference between the music maybe that I can make with AI versus what a musician Can, Can someone understands music can make, Can I, I'm sorry to interrupt. Um, no, Ellen. Um, but, so I talked, so a couple years ago I went on a podcast about this topic, and there was a rap artist, uh, and there was a guy who used to tour with Pink Floyd.

He played saxophone name Scott Page, and we talked about this. And he, he mentioned that he was listening to music that he thought was fantastic. And to his surprise, it wasn't written by a human.

It was written by a machine. Um, he had two takeaways. His number, his number one takeaway was the music itself was, was, was really good.

The second take away, the lyrics were terrible. He said he could tell it wasn't written by a human. 'cause there was a total disconnect between the music and the words.

So it was kind of like half a song. But that's, that still opens up the door for, But that was years ago. This was like, this was like, uh, 18.

No, it was two years ago. Yes. That was a while ago.

So think about how you were using ai, gen AI to write two years ago versus how people use it today. I don't mean not at all. I mean, his, his idea was it will get better, but yeah, I'm sure it has.

So That raises an alarm, though, among artists. And there are a lot of people signing petitions who are really p****d off about this, including a lot of well-known musicians, maybe because they don't want to compete, but maybe, But John, you know what? There's a lot of, there's a lot of co-developers p****d off about it too, and there's a lot of writers p****d off about it, too.

There's gonna be a lot of doctors p****d off about it. I, I know doesn't make a difference When You look industries, You look at the plight of a, a musician and, you know, I'm engaged to one, so I'll, you know, talk about that a little bit. It's, um, you know, they, they kind of sweat blood and go through a whole lot of stuff in their evolution as musicians and whatever.

That's not, I, I guess I can understand a little bit of resentment that somebody can come in with a few prompts and turn out, you know, a piece of music and have it sell millions. I mean, that's the thing that's remarkable about this. I mean, just millions of downloads or whatever, um, for, you know, these, these songs.

Whereas, you know, musicians who have been working sort of their whole lives don't get that kind of recognition, or their songs don't get hurt. Hey, Terry, what about the poor coach drivers who used to drive the horse and buggies for everyone? Well, I felt sorry for them, but their time is Well.

Well, I tell you, it's gonna, it's gonna ruin the live music experience though. That, well, it could be like a Star Wars cantt kind of thing. Well, you got hologram of like the blue elephant thing playing, and you know, who knows what goes on.

Well, we're going Down that post performing around. I have one question. So in Alan's view of the future, Terry, does that mean like the liner notes now come with the prompts?

Is that how this has been? Basically, yeah. Why not?

Why not? But you Know, like, um, I just, Well, no, let's what she Talk, if I'm a, if I'm play the musician role, right? I play volume when I was little.

It take me like three years to make noise, make it to music, and take maybe like for other people take 10 years to, to, to, to make the sound, right? Like for ai, like, people does not have experienced maybe 20 minutes, they got sound. How do you feel about that?

Like, it's kind of like when I, when I, I'm a content creator, right? I try smells Like Napalm in the morning. I mean, Victor, Yeah.

I like a very insightful company analysis. Take me like maybe a month to do a research, and then I go one article, and then somebody just, uh, read a pop within one hour. Maybe they can get a better report.

So like, psychologically, logically, this is a good way to go. But emotionally it's really hard. Emotionally it's hard.

It's you, you're right, Wiki. And I think that's the issue. And this isn't just musicians, as I said, it's across the board, but, you know, well, You know what, what would, Hey, Ellen, I'm gonna date myself, but I you, I think you and Michael appreciate this.

What would Don Kirschner have thought of this? Don first would've loved this. The Archie.

Aren't you All of them? Barry Gordy, John Kirschner, they all, who needs those stinking artist? They're temperamental.

They got issues. You've Got, I asked, I asked Will I am about this, about this, this comparison. He, he laughed at me and told me to shut the F up.

Uh, when I, when I brought it up, I was like, but doesn't that kind of concern you? He goes, no. If it's music, it's good music.

People will love it. So move on. Yeah.

I think maybe in the future, right? Uh, maybe it is not good comparison, but if, if you see the food, right? We have the organic food and we have the regular food.

Maybe in the future, the, the, all the art or like cut would be Organic. Those have Like a mark there. Okay.

This is the fuel Organic. Can, can we bring back American AI bandstand? Come on.

That'll work. That'll work. Dick Clark will Forever 35.

Yeah. Holograms with regional music. Um, I mean, it's happening already.

It happens, you know, uh, you see the projection of a Freddie Mercury on stage or Prince? Sure. Yeah.

Michael Jackson. Michael Jackson's Good. Yeah.

I mean, it could be the new MTV, right? Maybe they'll bring that back as an AI thing, right? Why not?

I know, I, years ago, before this stuff even really started in earnest, I did, I saw a, a show with, uh, Elvis on stage. And the reason I went is because mm-hmm. James Burton, the guitarist for Elvis, it was actually his show.

And then he had the Elvis hologram stand in, you know? Yes. Doing whatever.

And James Burton's from my hometown. So, you know, we always go support the hometown guy. But, um, I just, you know, I don't know.

I, there's just something about this that ruffles me. Um, You, you know, Jerry going, going back to that, that Elvis thing, I saw something on YouTube where he, where the, this was the experience with the guitarist. Yeah.

And there were people in the audience actually crying. Oh Yeah. No, no, no.

They have the emotional reaction, you know, uh, to to, to that. Uh, it's, but you know, it's also kind of crazy. I had tickets to see Elvis two months before he died at LSU.

And I sat in the parking area watching all the women with their boons and jumpsuits go into the thing. And we were in there two hours and they can, they canceled. They said he had left the building 'cause he was ill.

Um, and then they rescheduled. So let me get this straight. Did they really say Elvis has left the building?

Yes, he did. You know what? That's it.

Drop the mic. We're done. Let's move on to the next thing.

Elvis has left the building. Mike, we're gonna talk a little bit about fake shimmy. Yeah.

So I guess you had some incident where somebody is taking your identity and is impersonating you on the social media accounts. I think in this case it's Instagram, but yeah. You know, is this likely to, uh, proliferate and I guess you got a, um, shall we say, a suboptimal response from meta about what to do about all the stuff?

So, um, you know, walk us through this. Yeah. So I'm not the only one this has happened to as a result of this.

I've heard from many people that they've had similar plates. Evidently someone got the bright idea on Instagram to open an account under the name of Shimel, do Allen and Claim they were Alan Shimel. And I don't know if they didn't realize that I have like a cyber or a security background, but they went out and friended about 85 or 90 of my friends on Instagram, many of which were security people, What Wrong people to try to scam, right?

So I, someone immediately wrote me and said, Hey, someone's impersonating you. So they gave, I said, gimme the account. I went on the account, I clicked the three dots.

I made a report, this account impersonation of myself, scam. Within 10 minutes I got an Instagram automated message. This account doesn't violate our yeah, uh, community standards, but if you feel really strongly about it, here's a suicide prevention hotline.

I wasn't quite at the point of ending it, but I appreciate the suggestion. Um, so I did what any good cyber person should do who ha was also happens to be the publisher of Security Boulevard. I wrote a real long p not a real long piece, but a real loud plea piece about someone is impersonate impersonating me on Instagram and meta doesn't give a s**t because they obviously don't, because there was nowhere in that response from meta to, to appeal to reach out to a person, to have someone look at it.

But that's not all I did. I then reached out to, I, the nice thing about Instagram is I could go into this impersonating account and see every person he followed and every person he followed back or that followed them back. He didn't set it to private.

No. Well, you know, this is what you're dealing with. So I reached out to all the people that followed them back and said, Hey, in case you wanna know, I don't know if you wanna follow them, but that's not me.

And it's probably a scam, and I'd appreciate it if you blocked him and reported him. Then I reached out to everyone. He was, they were following, I don't know if it was a he or she or them, whatever.

I reached out to everyone they were following and told them the same thing. And it turned out it wasn't just following and following back. Once you followed them back, they would message you with a link to something with the Bill and Melinda Gates Foundation to some people, other people.

It was another kind of, they, it was all fake links, right? Spam, you know, phishing kinda links. Now, a lot of my security people sent me back snapshots of the, of the interaction where they would say, oh, shimmy, it's great to hear from you.

My, my brother Bob says hello, and this guy didn't know who he was dealing with. And they'd write back and say, oh yeah, great, good to hear from Bob. Tell Bob I said hi too.

And then, you know, my security friends would, would probably use language I don't even want to use on here and, and tell them where to go. Um, but evidently the way the algorithms, algorithms work, when enough people report the account out, report the scam. But here's the thing, they don't even let you send a a, a screen capture of the scam, but you just gotta report their scamming.

But when enough people do it, the algorithm, I guess, kicks in. Or maybe I did see one of my friends on LinkedIn tagged a whole bunch of meta security people with my LinkedIn post around the, uh, article. I don't know if that had anything to do with it either, but whatever it is, the guards of the algorithm or knowing someone and who, you know, about two days later, I got a message from Instagram that Shimmel Allen has been removed for violating community standards and thanking me for reporting it.

So there you go. Now, here's the thing. I'm a security guy.

I'm a media person. I, I was able to fight this. I was able to mobilize, I was able to get my friends involved.

What, what about the average person John Terry Wiki? What's the average person gonna do when this happens? You know, I, I was gonna say that you have this resolved in a record amount of time compared to all the examples of friends I know of because, um, I had something happen to me where there was a major breach, uh, face back when it was Facebook about 2018, a couple million people were affected.

So we had to read log in, and it was a, it was a pain in the ass. So I decided just to create my own, my own new profile. But they were responsive to me only because I was part of the media, right?

I I, I've heard from people who used to work within Facebook or slash meta that up to 21st sense of their profiles were fake. Um, like And they don't care. They just don't care.

No, they 100% don't care. And usually it's, you know, the person who's been, you know, imitated or whatever, that gets screwed on this. I had one friend had a similar experience to yours.

Alan didn't really know what to do, but reported it. He ended up losing his account, but the fake account stayed. Explain to me how, you know, that makes any kinda sense, Guys, if, and this is a message, it's not just meta.

I've heard horror stories with all of them X Google X, right? It's to all of them. If you, you can't talk about caring about security and caring about your user's experience and scamming and all of this.

If you don't take this stuff more seriously and do a better job, it can't just be handled by some algorithm and some AI bot and some nonsense community standards that no one even knows. It says you wanna find out more and you click more. I was expecting to see what community standards there are here that I need to know about.

And instead I got the suicide prevention hotline or reach out to a friend to help you feel better. Oh my God, are you Kidding? You Actually, yeah.

Sorry, repeat. Actually, I had a similar situation, but it was, um, LinkedIn. But even worse is after they use my name to make a profile, they post a phishing post.

They put information like, oh, um, some like famous security company is hiring. We're hiring this type of people. And if you are the, uh, fa fast 20, uh, 20 people, you get direct talk to me.

He left his, uh, email on that post. When I look at it, I'm like, oh, this is, this is kind of like a phishing, right? Because like normally for the famous company, they will not leave their like a Gmail account to do the recruiting.

So I immediately make a post on LinkedIn and act like, uh, hash the link in and maybe a couple of year, uh, days later, that page is gone. But I still feel it is terrible. 'cause like people use my name to try to fishing people to get their resume, their contact information.

Something terrible can happen under my name, right? It's kind of like dis destroying my reputation. Oh, absolutely.

Yeah. Even worse is like if I, you know, currently on the social media, if you make a post, not everyone can see only the part of people they can see depends on the algorithm, right? So like, there are two, like three things.

I think it's very important from security perspective. It's number one, after this happens, how can I let everyone, um, care about know some people they don't even connect to me, but they get impact, like this case, right? Mm-hmm.

Um, maybe some like PLA platform need to think how to let people know this. Uh, I saw actually in uh, acts, they did a very good job. It's like if, if some, some of the fund, this post is really bad, they can make a comments like warning people, this is not real.

This is fake, right? The comments cannot be deleted by anyone. They must, which Is good.

At least people know this is like fake information, this is fraud or something, right? And then the second one I see there are a lot of like Asian social media. They are trying to do like, uh, at least like a two way authentication.

Uh, after you log into your account, they link to your phone or something. At least the people know like a p at least, uh, it can validate your you. Yeah.

Mm-hmm. Yeah. And some, some, some social media, like they even ask you to upload a picture.

You show your identity in the picture. But it's, and, and I know it's a little bit hard to manage on the privacy part. Um, and then I think maybe there's some way like the, the identity part, we need to sync a solution, but it's, it's always like difficult part from the cybersecurity perspective.

Yeah. Terry, how crazy can all this get? Because theoretically I could use the fake shimmy to maybe log in and bypass some two factor authentication.

And then the next thing you know, I got myself a deep fake imitation of Shimmy says, and away I go. Well, yeah. So it can get, I, I think really bad.

I mean, you can imagine just all the consequences of this. And as, as Wy said, there's no real, you know, process for identifying people and that they are who they say they are. This account is legit.

Oh yeah, you can get your blue check or whatever it is. I don't think that's super difficult to get these days on most of these, uh, accounts. I don't know.

You know, I don't know if anybody wants to really try on a fake account to, to see, but maybe they did, you know, maybe they can. Um, I think that, you know, uh, these social media platforms are so concerned with, they want these giant communities of people, right? And they want, I mean, that's their business plan is to reach all these, these people with, you know, whatever their advertisers are hawking or whatever, they're hawking.

And I don't think they pay any attention to any of this. That's where the, the fault lies. I don't think there's any serious attempt by really any of them to, and haven't, Haven't x in, in meta the last couple of years, at least up until recently been just slashing away at the content moderators and the people in the securities.

Well, yeah, Yeah, exactly. So they're, they get, they're getting rid of all of, of all of that. And then they've of course bent to pressure, like in the meta they've bent to pressure from the administration, right?

The, the Trump administration. So I think that, that that security and resolving all of this have, I mean, look, nobody, everybody has always complained that if you have a problem with a social media account, you can't get it, get to anyone, right? I mean, that's always been a big complaint.

It's very difficult to get to a human being to help you with, you know, whatever's going on. I think now it's gotten worse. And I think they may be actively just, I, you know, thumbing their nose insecurity.

I don't know. I mean, I tend to, when I get these kind of things or when somebody is, do I tend to mess and do a little bit like your friends Alvin and I mess with the, with the attackers a bit or the, the bad, the bad guys, you know, um, uh, a bit. And I had somebody try to do an account takeover recently on Facebook and you know, the guy accused me of not wanting to help my friend, you know, he is trying to send a code to me so he could take over my account.

I said, oh, I do anything to help her. But account takeovers are hell and uh, to, to undo. And so, um, we're not gonna, you, you're gonna have to find somebody else to scam.

But I don't think, I mean, uh, you know, I don't think there is a real process for making sure that this, these kinds of things don't, don't spiral. I don't think there will ever be a process. I don't Even more, I don't think it's the nature of these platforms and how they came out.

It's about, it's about the money. Look, meta doesn't wanna go back to the advertising community and say, oh, by the way, we discovered that 20 20%. 20%.

Yeah, Advertising community's gonna go, where's my refund for all these fake things? Exactly. No, it messes up their business model.

But, you know, yeah, I, I just more joke somehow. Like, it's very, it's very funny. Like one famous people saw there's a fake account on social media for hi with his name.

So like in the, like in person event, he met his friends, he said, okay, I saw your profile talk to this fake account, so don't talk to him because it's fake. And then the other person said, okay, I've never been in that social media platform. So it's like a fake account, talk to fake account.

Wait, Well, and, but there is that, when you get that level of fake accounts, you're get that spot traffic and, but, but that feeds into the algorithm, right? Right. That's how all these Russian bots and all these things do.

They, they talk to each other. They like each other's posts, they repost each other's posts, and the algorithm lives off of that and blasts it out. And that's how it winds up on your feed.

Wait, are You telling me that I am not best friends with George Clooney Online? You are Mike, because he, he goes to the brewery to play a half a side of album with you on Thursdays. Alright, we got, Yeah, but Mike, like, you know, like when, when, when I meet somebody, like on the social media, especially in the cybersecurity side, when we talk to each other, I can see people try to test, ask me some question.

Like we talk in person and then they, they start to talk to me. Yeah, no, they verify. They, alright guys, we're at, we're at 40 minutes.

I gotta pull the plug. When we do live shows, we're doing 'em 40. So what a great conversation.

I'm grateful that Meta did shut that account down though. Uh, synthetic music is music. It's real, it's not fake.

It may not be your cup of tea. I don't know if it's truly soulless, but um, and then of course Microsoft is making inference chips. You've heard it all here.

Techstrong gang. We're gonna have Techstrong tv. I think we've got, uh, tech Field Day today.

So stay tuned for that today. And, uh, we'll be back tomorrow with More Gang. Also a shimmy says tomorrow till then is a real shimmy.

The real shimmy, You know, shimmy. Yeah. Until then is Showman, we're out.