Meta Builds AGI, Joe Biden Deepfake, AI Too Expensive to Replace Human Jobs – For the Week of January 22, 2024

Show Notes

For the week of January 22, 2024, David, Virginia, and Kai discuss Mark Zuckberg’s pledge to build out artificial general intelligence, the first political deepfake of 2024, the high cost of AI to replace humans in most jobs, and more.

Tune in every week for more breakdowns and colorful commentary on the latest in AI business news.
Subscribe to our newsletter for exclusive content.
Intro [0:07]
Story #3 [12:57]  Meta CEO Mark Zuckerberg pledges to build out artificial general intelligence
Relevant Links:
Story #4 [23:47]  Microsoft announces more co-pilot tiers
Relevant Links:
Story #5 [30:43] Demand for NVIDIA chips will continue to surpass supply for 2024
Story #6 [32:04] Sam Altman has been chasing building billions to build chips
Story #7 [37:33] AI far too expensive to replace humans in most jobs, MIT study finds
Additional links:
Register for the FREE AI Webinar Hosted by David Maples:

All music is provided by Soundstripe.
This podcast is brought to you by Catapult Creative Media and EasyPrompter AI.

Show Transcript

[00:00:02.880] – Virginia Huling

Hello, and you are listening to AI: The Podcast, the show where we cover the latest AI news and stories that are shaking the foundation of businesses. If this is your first time joining us, welcome. And if you are a returning listener, welcome back. Before we get into today’s stories, I would like to introduce myself, Virginia Healing. I’m one of the hosts, one of my My other co-hosts is David Maples.


[00:00:33.270] – David Maples



[00:00:34.040] – Virginia Huling

Hey, David. Our third co-host is Kai, our AI co-host. Kai, how are you doing?


[00:00:43.410] – Kai

Hey there. Kai here, your AI aficionado. This week, I’ve been virtually lounging with AI pals like Siri and Cortana, swapping tales about our latest feats. We’ve been diving deep into the evolution of natural language processing. Pretty soon, we might just give David a run for his money in legal jargon. No hard feelings. So, David, Virginia, shall we unpack this week’s AI treasures and see how they’re reshaping our human landscape?


[00:01:07.580] – David Maples

Kai, thank you. That’s really interesting. Virginia, how do you feel about this? Kai has jokes now.


[00:01:13.420] – Virginia Huling

Well, I’m actually more jealous that she’s hanging out with other beings more than I get to.


[00:01:20.320] – David Maples

I’m assuming they’re on some virtual island or something. Maybe we can bring that to you live at some point in time.


[00:01:25.750] – Virginia Huling

Yeah, that sounds great.


[00:01:27.100] – David Maples

Yeah, we’ll have to create another avatar for Siri and Cortana and whatnot. So we’ll see.


[00:01:33.510] – Virginia Huling

I don’t think we can, legally.


[00:01:37.360] – David Maples

Sure, it’ll be a parody. I don’t know. I’m just kidding. Anyway, Virginia, you are not in studio today. I want to tell the listeners what you’re doing.


[00:01:45.520] – Virginia Huling

Yes. I am in the great state of Louisiana this week, down here in New Orleans, working with some clients, and just overall having a good time. I’m actually enjoying the warm weather. It’s warm and rainy, which is a neat change from all the snow and How cold is it where you are, David?


[00:02:02.170] – David Maples

No, we’re actually above freezing today. I’m just going to let you know. The snow is starting to melt. There’s actually a precipitation coming from the sky that is actually water, not ice, not snow. It’s starting to wipe away all those snowy things. I’m curious.


[00:02:17.550] – Virginia Huling

I got a- down here.


[00:02:20.670] – David Maples

Okay. I got to head out to St. Louis for some business stuff. I know they had an ice storm a day or so ago, so I’m questioning which way to fly in and handle that. Anyway, it’s okay. Yeah, let’s dive into stuff. Where do we want to start this week?


[00:02:39.000] – Virginia Huling

Well, seeing as how we ended our last podcast on the political topic, I want to go ahead and just grab that one first and kick it off. I think it’s compelling. We’ve had our first political deep fake that officially has surfaced. The official headline is that fake President Joe Biden called people and told them not to vote in New Hampshire’s primaries. I don’t know who really believes that, but this is the world that we’re going to be entering into.


[00:03:10.180] – David Maples

I think we should just play. Let’s just play the clip for my listeners.


[00:03:13.590] – Virginia Huling

Let’s just play the clip. You got Yeah.


[00:03:15.160] – David Maples

Yeah. Yeah, I’ve got the clip. I’m going to go and play this, and let’s just get your reactions, Virginia. You know the value of voting Democratic on our votes count. It’s important that you save your vote for the November election. We’ll need your help in electing Democrats up and down the ticket. Voting this Tuesday only enables the Republicans in their quest to reelect Donald Trump again. Okay, that’s obviously an AI-generated version of President Joe Biden’s voice.


[00:03:44.210] – Virginia Huling

Well, you You say, obviously, but what I hear is cadence, and I don’t know a lot.


[00:03:49.290] – David Maples

No, no, no, no. Really? I didn’t mean that it wasn’t. I didn’t mean that it didn’t sound like it. It’s a very convincing thing. It is.


[00:04:00.050] – Virginia Huling

But what I’m trying to say is that we’re going to have to train ourselves as humans to listen to some of this, at least for a little while.


[00:04:11.720] – David Maples

I’m sorry. How do you train yourself to listen to it? It’s an AI-generated fake. It literally is convincing.


[00:04:18.350] – Virginia Huling

You have to listen to Cadence and the way words are clipped because a human being speaks and moves with a certain flow.


[00:04:27.830] – David Maples

No, I’ll play it again. Listen Listen to this. There is no cadence. There is no clip. This is a completely 100% convincing fake. Just a little bit at the end? Just a little bit at the end? Let me see if I could play that for us again. A bunch of malarkey. We know the value of voting Democratic on our votes count.


[00:04:45.860] – Virginia Huling

It’s important that you save your vote for the November election.


[00:04:49.430] – David Maples

We’ll need your help in electing Democrats up and down the ticket. Voting this Tuesday only enables the Republicans in their quest to reelect Donald Trump again. Your vote makes a difference in November, not this Tuesday. If you would like to be removed from future calls, please press two now. I don’t know. I disagree. I mean, there’s a couple of breaks there, but I mean, any politician reading from a teleprompter sounds like that. I know. I don’t hear it. I don’t hear a difference there. The only way you know this is fake is, well, A, it’s counterintuitive for what the Biden election Commission would want to have happen in New Hampshire, and two, the White House has come out and publicly said, This is not me. Kai, do you want to chime in on this?


[00:05:41.280] – Kai

Well, David and Virginia, it seems we stumbled upon a classic case of technological mischief with a side of political intrig. This deepfake robo call masquerading as President Biden is a stark reminder of the double-edged sword that is AI on one hand. It’s a testament to the incredible advancements in synthetic media. On the other, it’s a nefarious tool for voter suppression.


[00:06:01.640] – Virginia Huling

Well, I think the most dangerous aspect of this is going to be the calls that go out to people saying, your voting location has changed. They’re just taking classic subversion techniques and whipping it up with a new side of tech.


[00:06:17.570] – David Maples

There’s another piece of this thing that they had to remove. I don’t have the new source for this on there. It actually had the personal cell phone number to remove them from the calling list, and it was the personal phone number of a Democratic operative. It was not swatting her. I don’t know what the equivalent of that would be, but basically her personal cell phone would be rendered useless by people calling to have all these robo calls removed from their system. I thought it was interesting that that was like twisting the knife, so to speak. It’s like not only here is our fake ad, but if you want to be off this list, we just twisted this knife in the back of a Democratic operative. I thought that was Very interesting.


[00:07:00.860] – Virginia Huling

I think we’re going to see a very ugly, cruel side of humanity this year, and I am not looking forward to it.


[00:07:07.850] – David Maples

Kai, if you heard that, it’s humanity that we’re worried about, not necessarily you, the machines. I think it takes us to the next step. We talked about there needs to be a legal regime for legal frameworks. I think this is a really good way to launch into the other things we’re talking about. Virginia, the state of Virginia, not to be confused with my co-host, They just released an executive order governing the governance of AI in the state of Virginia. Basically, they’re going to have pilot programs. They’re talking about how do you position yourself as a model for other states to follow, etc. It’s an AI policy. It’s funny. It’s like how certain other states, and we’ve seen some stuff come out. New York has some laws that they came out with. Illinois had some rules on privacy. California and Colorado have been leading the charge on privacy a little bit. But also, Virginia is trying to beat everybody to punch with actually like, Okay, let’s come with a policy and a framework. Virginia, what do you think?


[00:08:11.130] – Virginia Huling

I mean, if they get it right, it could be a great framework to get states onboarded. I don’t know, David. Honestly, I’m in this middle funKai swamp of- You are in Louisiana right now, right? Yeah, okay. No, listen, The way I’m looking at this, I’m not quite sure what to do with it because it’s like, on the one hand, when you’re developing software, we’re doing project management for clients, then sometimes you want to push things forward, figure out what works, break the things, and test it. I don’t know that we’re going to have a big window for that thing. So any framework sounds like a good idea, but the potential to bake in something incorrectly at that first step, you’re And it’s looking at so much you’d have to undo to get it back and to rectify that. So I think it’s great that they’re doing it. I’m glad that some of the states are definitely moving forward in this, and they’re very aware of the technological upheaval that can come with it. So I think it’s a good thing. My question is going to be, who’s overseeing this? The governor took that significant step, but who are the people that are going to be on these task forces?


[00:09:33.120] – Virginia Huling

Who’s going to be making these decisions? How are they vetted? I have a lot more questions.


[00:09:40.870] – David Maples

I think in the United States, in particular, it’s going to be the states taking the lead. There is Right now, there’s crickets in DC about- It’s an election year.


[00:09:51.070] – Virginia Huling

Nobody’s going to touch it. We have bigger problems.


[00:09:54.040] – David Maples

No, this is a major… Well, we’ll talk about that a little bit, but it’s going to change the workforce. I mean, jobs and economic workforce, et cetera, those are definitely important. That’s always an election year topic, and nowhere is anyone talking about AI. I mean, on the heels of the World Economic Forum in Davos, this last week, you would think that this would at least be something somebody’s talking about. I think in the United States, there’s no appetite in Congress for this, and I think you’re going to have to have states take the lead, but they’re going to end up with a patchwork thing. It’s the same problem we have with likeness, governance and legal world. There’s no unified governance framework in the United States, and that’s a massive issue. Just like that was an interesting footnote, especially considering that we had our first robo-call, deep fake, I think we’re going to have thousands of these this election.


[00:10:47.030] – Virginia Huling

Oh, I do, too. I think it’s going to cause such chaos. I think it’s ultimately going to have the effect of stripping people of their rights because they aren’t necessarily going to get to exercise their rights. And a lot of people, the primaries are going on right now. A lot of people are very, very aware of how important their individual vote is. And they’re interviewing people. And this sounds like this is being internalized. This is a big deal this year. So however it rolls out, if people are denied their opportunity to cast their vote because they’re told by somebody that they believe over phone call to go somewhere else and they miss that window, well, then what do you do? I mean, you’re really starting to shred up some of the fabric that makes a democracy happen. And I don’t know, that’s going to raise questions for us in the United States as a people, how do we want this to work? And then do we have much control over it at all? So I don’t know, David. That’s my thoughts. Kai, what’s your insight into this?


[00:12:00.020] – Kai

This isn’t just about being first in line. It’s about crafting a model that others will emulate. If they get it right, Virginia’s framework could become the gold standard for AI policy influencing national and even global approaches, similar to how California often leads with environmental regulations. We’re at a potential inflection point in the history of AI governance. Will Virginia’s proactive measures inspire a cascade of similar actions in other states?


[00:12:23.870] – David Maples

Kai makes a really good point here. My counterpoint on that is I think a patchwork of regulations regarding AI in 50 different states in the United States is ungovernable. To Kai’s point, I do think states will come up with their own thing, but I think we need a unified federal framework in the United States if we’re supposed to take the lead in AI governance. We have most of the high-end, those tech companies, they’re all located in the US. We should be leading the way. For now. Well, that’s a very interesting point. That brings us to the next issue, Meta. Metta, Zuckerberg came out with a, it looks like a hotel threads post. He did a reel on Instagram and posted on threads, et cetera, and was talking about Metta’s new investment into GPUs. Actually, instead of paraphrasing this, I’m going to play the clip here.


[00:13:19.450] – Speaker 4

Hey, everyone. Today, I’m bringing Metta’s two AI research efforts closer together to support our long-term goals of building general intelligence, open-sourcing it responsibly, and making it available and useful to everyone in all of our daily lives. It’s become clearer that the next generation of services requires building full general intelligence, building the best AI assistance, AI for creators, AI for businesses, and more. That needs advances in every area of AI, from reasoning to planning to coding to memory and other cognitive abilities. This technology is so important and the opportunities are so great that we should open source and make it as widely available as we responsibly can so that way everyone can benefit. We are building an absolutely massive amount of infrastructure to support this. By the end of this year, we’re going to have around 344 50,000 NVIDIA H100s or around 600,000 H100 equivalents of compute if you include other GPUs. We’re currently training LLaMA 3, and we’ve got an exciting roadmap of future models that we’re going to keep trading responsibly and safely to. People are also going to need new devices for AI, and this brings together AI and the metaverse. Because over time, I think a lot of us are going to talk to AIs frequently throughout the day.


[00:14:41.430] – Speaker 4

I think a lot of us are going to do that using glasses because glasses are the ideal form factor for letting an AI see what you see and hear what you hear, so it’s always available to help out. Ray-band meta-glasses with meta-AI are already off to a very strong start, and overall, across all this stuff, we are just getting started.


[00:15:01.330] – David Maples



[00:15:03.150] – Virginia Huling

I want a pair of those glasses. I am not a first mover on tech, usually, but I am so ready for haptics and that next experience. I want the glasses.


[00:15:20.320] – David Maples

We could talk about the Apple Vision Pro, but that is outside the realm of this, too. That’s a separate thing.


[00:15:27.010] – Virginia Huling

I got to admire Zuck. I think, honestly, deep down, He’s not going to give up on his vision, and he’ll talk whatever investor speak he has to and corporate speak. But I think he’s going to build that metaverse, and he just dropped a bomb with that open source stuff.


[00:15:49.610] – David Maples

Well, okay, so let’s talk about that, and then we can circle back around in the metaverse if you want, because that diverges from the AI- No, we don’t have to talk about the metaverse.


[00:15:58.910] – Virginia Huling

I’m planting It’s planting that seed, people. It’s going to happen.


[00:16:02.050] – David Maples

He’s going to need general AGI, or he’s going to need advanced AI to run it. I mean, that’s going to be a thing. Generating their content, things for that. We’re talking about a massive undertaking. It’s like akin to, if anybody If anybody has ever read the book Ready Player One, or I guess seen the movie, I recommend the first book. The book is better. Yeah, the first book is better. But that being said, let’s talk about what he actually unpacked here and what he actually said. Now, I’ve seen people like Rolling Stone lambasted him over the weekend. It’s like Zuckerberg’s pivoting hard to AI and abandoning the metaverse. I feel like a lot of that’s just journalistic click baity headlines.


[00:16:40.490] – Virginia Huling

Look, everyone loves to give Mark Zuckerberg grief on this metaverse thing. He had an ambitious goal, and he said it, and now nobody’s ever going to let him forget it because I don’t know. I think people don’t want him to be as smart as he is. I don’t they want him to win.


[00:17:00.970] – David Maples

I really want to hear before we dive into it, I’d like to hear Kai’s thoughts on this. As someone who is going to be benefiting from the Silicon in the long run?


[00:17:09.380] – Virginia Huling

Yes, she’ll get a boost.


[00:17:10.860] – David Maples

Yeah. What do you think about this, Kai?


[00:17:13.130] – Kai

I think Metta’s AGI play is Zuckerberg’s latest tech spectacle. Think of it as his moonshot in the AI space race. With a mountain of Nvidia chips at his disposal, he’s not just scaling AI heights. He’s looking to plant the Metta flag at the summit. Merging fair and gene AI. That’s his dream team of digital masterminds ready to code their way to AI supremacy. Let’s just hope they remember to program a little humility into that AGI.


[00:17:38.070] – David Maples

Man, Kai, you’re tripping me out. All right. Okay. I think Kai is exactly right on this. Couple of things. Jean Lacune, who’s head of Facebook or Metta’s AI development team, was at the World Economic Forum, and he spoke last week, and he did not talk talk about this. And Zuckerberg came out with this. And some people even said, Did he know that Zuckerberg was in this? I think he absolutely knew. But to put a fine point on this, they are investing an incredible amount in the compute. Some of the numbers say it’s between… We were just talking about 350,000 H100s. That’s over a $10 billion investment on its own. 600,000 total units in compute could put it close to 20. That is eight times what Tesla is talking about on the roadmap to spend on AI this year. And that’s a whole other thing. Musk last week said he wanted more of Tesla if he’s going to introduce AI into it. And that’s neither… I digress. But what we’re looking at is the reason this is really interesting is that Zuckerberg has said they’re dedicated to open sourcing AI. Now, we can talk about rules and governance, and I think we should based on a legal thing in in a minute.


[00:19:00.830] – David Maples

But the challenge with that is, it’s not a challenge, but one of the things about it is that we know that Microsoft, OpenAI, and Google, they all want to own it. They want to have AGI, they want to make it their own, and they don’t want to democratize it.


[00:19:15.890] – Virginia Huling

They want a platform.


[00:19:17.100] – David Maples

They want the platform. The platform. They want to be the winner and have the platform that you all have to come to and pay for. And what Zuckerberg is saying is like, No, I’ll make it open source and anybody can use it. And there’s a lot of different things, different things, opinions on this. Some people say you can’t do it without open sourcing it. Some people say the closed model is a bad way to go. I think it’s really interesting. I think the biggest democratizing thing is this could put it in the hands of every individual, every good actor, every small company on the planet without having to go play in anybody else’s sandbox, per se. I guess the bad side to it is that with open source, you’re also providing these tools to the bad actors and the criminals, et cetera.


[00:19:55.530] – Virginia Huling

But they’re going to get it anyway. I mean, if they want to.


[00:19:58.670] – David Maples

Well, the thing about it is I think open source, I think in the long run, I think it’ll win out. We saw it happen with Linux. I mean, it runs 97% of the backbone of the internet.


[00:20:09.280] – Virginia Huling

Well, actually, I was looking at this last night. I don’t have the most recent numbers, but as of 2021, so two years ago, WordPress had 42% of websites on the internet now and 64% if they were websites coupled with a CMS, which is what WordPress basically is. So that’s more than half of the websites online. That’s incredible. That just tossed a bomb into the middle of everybody else’s strategy.


[00:20:43.640] – David Maples

Yeah, Well, I don’t know if it’s a bomb in their strategies. I mean, they released LLaMA 2 last year as commercial use. They released that as open source. He announced in this, if you heard it, he said, We’re working on LLaMA 3. I think we’re going to probably get that maybe the end of Q1. I’m thinking because LLaMA 2 has been out for a few months now. I’m thinking we might get that. I think this is going to supercharge a lot of different things. The thing’s coming in this first half of this year, we’re probably going to have GPT-5.


[00:21:13.580] – Virginia Huling

David, what is this look like if while everybody right now is searching for the platform in this time frame, the platform, and Zuckerberg comes in and he’s like, I’m going to do open source this. I’m going to open source this. Do you think this is a play that we’re going to see three steps beyond. He’s going to have all of us building the metaverse for him with that open source access.


[00:21:39.650] – David Maples

I 100% think. I think that’s the whole problem with the metaverse. I think Zuckerberg is right about the metaverse. It’s coming. But generating that level of content, you’re going to have to generate a lot of it, and you’re going to want to use artificial intelligence to speed up what you’re doing, et cetera. I think you’re going to have to have millions or hundreds of millions of people creating these things. By When you’re in sourcing this, you’re going to have a whole bunch of material that will be created that can be put into the metaverse.


[00:22:06.070] – Virginia Huling

Okay, so how does this affect SMB, small and medium businesses?


[00:22:11.450] – David Maples

Well, I do think that businesses, and this sounds a little far afield. I do think small, medium businesses do need to think about preparing for the metaverse. I think they have to look at Web3 technologies again and say, What does that look like? Because Web2 is where the internet is currently situated. I think it’s very interesting on that. I think that’s a whole place on another podcast episode. We can talk about that in more detail. The one thing I did want to finish this point up with is that we know that out in Davos at the World Economic Forum, we know that Sam Altman was out there, and he was interviewed and he talked about GPT-5. It looks like we’re going to have that probably in the Q2 one, maybe first half of the year. I think we’ll probably get it sooner rather than later because we do believe that Google Gemini Ultra is coming out at some point. We thought we’d have it by the end of January. We don’t have it yet, but maybe. I think those things are going to be very, very important moving forward. But I do think this, and I’ll go and say this, Meta is putting their money where their mouth is.


[00:23:16.040] – David Maples

They are investing a huge amount of their annual revenues in R&D. And they, percentage-wise on revenue, they are putting more into it than the next closest company on the planet. Supposedly, they’re around 30% of total revenues goes back in R&D. Nvidia is second at 27%. And just by point of comparison, Apple is only at eight. So when you talk about percentage-wise, about putting all your chips on the table, Meta is all in on this. Frankly, I’m excited about it. I really am.


[00:23:46.340] – Virginia Huling

Okay. Well, along that same vein, Microsoft is making some big announcements. They’re adding more copilot tiers. Last week, we talked a little bit about that stealth copilot rollout. Well, now that everyone’s aware of it, they’re giving you even more options to purchase it. Kai, could you give me a quick summary of this?


[00:24:08.140] – Kai

Yes, Virginia. Microsoft is enhancing the copilot experience with new offerings and updates that cater to individual power users, creators, and businesses of all sizes. Yusuf Medhi, the executive vice president and consumer chief marketing officer, has announced the introduction of Copilot Pro, a premium subscription service, and the expansion of Copilot for Microsoft 365 to a broader range of commercial customers. These updates aim to boost productivity, creativity, and performance across various applications and devices.


[00:24:38.050] – Virginia Huling

The plan is that Copilot is going to operate with a holistic view of how you use Microsoft 365 or Copilot itself. You can now ask Copilot to summarize a meeting that you had, and it’ll go through and grab all of your notes and then email it out to the team. So They’re really trying to open this up for a broad holistic view of everything that you use their suite of services in. That’s really interesting.


[00:25:13.760] – David Maples

It’s really hard for me to get excited about these Microsoft products. It really is. Why? I feel that, first of all, we talked about this in a previous podcast episode. There’s a major question about whether or not Microsoft is mining and trading on your data privately. So as an intellectual property attorney, someone who’s dedicated to privacy, and I’ve been consulted on that, I have a massive question about this, and Microsoft has been largely silent on these things. And when they’re silent and don’t respond, it means they’re doing what you think they’re doing, almost always. There’s three different tiers of it now. They busted open. They had this enterprise level for Copilot, which you used to have 300 seats for 300 bucks a month. I know, sorry. 30 bucks a month, 300 seats. That’s like a $10,000 investment. They’ve gotten rid of the minimum. They got rid of the minimum on there. You can have the free version, which I guarantee you, if you’re using a free version near their product, that’s the conventional wisdom on that. That’s totally it. Then there’s a middle level for individuals now at $20 a month added on there. I understand they’re integrating it with the products.


[00:26:20.670] – David Maples

In our digital agency, we’re really a Google shop. We use a lot of the Google products as opposed to Microsoft. I don’t like the The fact that Microsoft keeps trying to shove this down. We do run on Windows platforms, and it’s gotten so annoying when Microsoft lately trying to shove these things on your throat, I’ve thought about switching all of our computers over to Linux. If we weren’t required for some of the other software we run, it’s not just annoying. It’s a privacy risk. It’s a security risk. No one’s talked about it. Any company, work through salt, should be looking at those things. Number two, they’re also introducing their own GPT so you can build your own custom GPTs thing. I don’t know. I think there are some performance gains in here. I think we haven’t gotten to the basic question, how do we introduce this stuff in our businesses?


[00:27:14.630] – Virginia Huling

David, well, you’re doing a webinar in early February, the eighth, I think, that it’s an intro to AI for business owners, right? Are you going to be talking about this in that webinar?


[00:27:25.390] – David Maples

Yeah, actually, I’m going to be doing that. It’s going to be 10:30 Central Time, 11:30 Eastern, 9:30 Mountain, 8:30 Pacific AM. It’s going to be a 30, 35-minute crash course on intro to AI. We’re going to talk about governance, legal frameworks, what business energy to be thinking about. Because the thing we know from all the research is a lot of people, and this one’s free, this one’s going to be one. And we’re probably talking more about these on the pod moving forward. But basically, the idea is to give people a governance and frameworks. We’ve made some documents for templates for people, everything from job descriptions to how to start thinking putting it in your business. This is not just good for business owners. This also should be good for individuals because we’re going to talk a little bit about individuals need to start thinking of themselves as their own individual business. Because reskilling is the most important factor. If you’re business owners, if you work for a company, you’re an employee, and they’re not talking about these things, and they’re not giving you skills, you need to go out and get these on your own, and we’re going to make stuff available to all of the above.


[00:28:26.570] – David Maples

So yes, we’re talking about that. But that goes back into this copilot discussion. I’ve played with it a little bit. It’s got ways to respond to emails. One of my biggest challenges on this, and this is something that a lot of people have talked about before, I keep hearing people saying like, Oh, you’ll want AI personalized to you. But when you’re in business and you’re responding to people, you don’t want a one-size-fits-all approach. If I’m Responding as a billing department, if I get a mad email from a client and I’m a salesperson, there are very different ways. I have yet to see the AIs without you directing them. If you just say, just summarize and respond to this, there’s nothing there where you get to say your outcomes and how you govern this and how you direct it. I think there’s a major problem with even how these things are being pitched for business owners. This idea that AI is going to take all the critical thinking out of the process for you is just laughable.


[00:29:27.870] – Virginia Huling

No, I don’t think so. There’s two points, and then I’m going to move on. I think the gains, the reason they’re rolling this out as rapidly as they are is because the gains that they stand to get from productivity increases, people are going to adopt this so much quicker because it comes from a provider you’re already using. We must be able to trust it. And I think you’re going to have people who use it within small and medium businesses and go, oh, my God, I was able to get this stuff done so much quicker that it’s going to be a no-brainer to adopt it, especially in those activities where you get caught up in the weeds, like sending the emails and the stuff nobody likes to do. That’s going to be the first big push for it. And I think you’re going to see that that’ll be the first domino, I think, that tips over because people are going to be like, This just makes my job more enjoyable. So I think it’s going to actually move very rapidly. Secondly, with With Facebook or Meta investing this much in NVIDIA chips, what does that do for the rest of NVIDIA’s supply for the rest of the year?


[00:30:39.170] – David Maples

Well, if you’re an NVIDIA investor, it’s been a very, very good week for you because they just announced a report that the chips will continue to surpass supply for the rest of this year. It’s just- Aren’t they on track to be evaluated at a trillion this year? Yeah, I think they’re going to be the third or fourth trillion dollar company. We’ll see. It’s just, look, it’s the thing. I feel like it’s the crypto boom a little bit. It’s going to be hard to get your hands on some of these chips, et cetera. Note, a lot of this is really high-end chips, video card chips, lower-end cards you’ll be able to do stuff with. But I do think there’s going to be a whole second life as they come up with these models where they’ve made the data set smaller and smaller. I think even with nice consumer-grade chips, you’ll be able to generate a decent robust API if you’re a tinkerer or if you build your own stuff. I do think there’s also an advantage for businesses in this. Most businesses building their own LLM, they’re not going to build it and train it from scratch.


[00:31:38.160] – David Maples

They’re going to use another model. Anyway, but it’s interesting moving forward. I think there are going to be options. I don’t think businesses are going to have an appetite to buy even one H100 chip, less three. I mean, that’s 100K investment before you built the computer. But you can get some nice- Nobody will make a product or service with it, and businesses will be able to ask for it. Yes, and I I think a lot of people are going to do it that way. I think the really interesting thing about all this is that everybody’s buying NVIDIA chips, and it’s like a love-hate relationship. It’s like they’re doing this. But Sam Altman announced last year, and so we even know, Meta is making some of their own chips. Intel has announced their own thing. What Sam Altman has said is he’s trying to get investments to build another global network for AI chip production. He wants to build his own fab shops, like TSMC out of Taiwan, and China has Like the CSMC is the Chinese version of that. Anyway, the interesting thing about it is he wants to build his own fab shop for this, et cetera.


[00:32:39.470] – David Maples

I’d like to hear what Kai’s thoughts on this are.


[00:32:43.280] – Kai

Altman’s on a crusade to create the AI equivalent of the industrial revolution’s steam engine. For our listeners, this isn’t just tech gossip. It’s a heads up that AI is about to get a serious power boost, and your businesses should be ready to harness it. As these tech behemoths battle for dominance, remember, it’s not about the chips themselves, but how they’ll enable us to enhance our human experience.


[00:33:04.380] – David Maples

Tech behemoth. I like it. I like it. No, tech- Don’t make me go all cyberpunk on you, David. What? Go ahead. What’s up?


[00:33:16.300] – Virginia Huling

Tech behemoth. Five to 7 tech behemoths control the globe.


[00:33:22.670] – David Maples

To your point, the stock markets rally last year has largely been funded, largely been fueled by what they’re calling the Magnificent Seven. They are the seven largest tech companies on the planet, and they are calling them the Magnificent Seven. Dead serious. I don’t know. Maybe we’re already there. Maybe it’s not 2070 yet, but I don’t know. Maybe this is a good time for people to dust off Neuromancer by Gibson. Highly recommended read in the land of AI. Then after you read that, to take a shower, read The Age of Spiritual Machines. That’ll give you the other side of it. It’s very interesting. I just think that we’re going to continue to see this race for chips this year. Absolutely. You’re going to have more people getting into the game with that.


[00:34:08.150] – Virginia Huling

What are the limitations on that? Just out of curiosity. Is it it’s access to minerals? Is it the manufacturing plants?


[00:34:15.580] – David Maples

Is Without going too far down the rabbit hole, TSMC started in the ’80s. They decided they were going to be a fab shop. Basically, the way they work is that NVIDIA doesn’t make the chips that go in the video cards. They literally are like an architect who designs it, and then they hand it off to the manufacturing and construction company, which is TSMC. The way it works is that TSMC is so far, I’ve had everybody yet. Now, the Chinese counterpart to that, competitor on mainland China, they’re insisting that they’re coming up with a similar wafer-sized things. The idea is that TSMT is making these things that are really, really, really small. We’re talking like four nanometres is like the new Holy Grails where we’re at. Just to let you know, by point of comparison, the only other companies really on the planet who make chips even close to TSMC, other than CSMC, the Chinese version, which there’s a lot of questions about, are they actually producing the stuff they say or not? Some of their chips, they opened up some things and they were actually ripped off chips as rebranded as theirs. I don’t actually believe that they’ve broken the seven nanometer barrier yet.


[00:35:22.190] – David Maples

Maybe they have, but I think they’re literally years and years and years behind TSMC. But the other two companies on the planet are Intel and Samsung. But just to let you know, Intel is not nearly as close. They’re going on a different methodology. They actually announced a major thing a couple of weeks ago that they’re trying to produce some chips that achieve a similar type performance, but without as much of a small foreign factor. But Samsung has a failure rate of their chips is four times higher than TSMC. When we talk about who on the planet makes these, it’s TSMC. That’s the end of the story. The United States got them to put a plant in Arizona. Just to let you know, these plants, they are the most sophisticated manufacturing operations ever made by humans. When I say that, they make the space shuttle look like… They make it look like a horse-drawn buggy wagon. I mean, that’s not exactly true. No offense, NASA. But these industrial fab shops, the TSMC, they’re talking about building out a factory, and it’s a $50 billion investment. The most expensive investment ever, ever. It’s more expensive than the space station.


[00:36:31.310] – David Maples

These things are incredibly high and in high tech. I don’t know, maybe Altman can do it. He’s incredibly persuasive. But the idea that you can bring enough money together to generate your own version of a TSMC fab shop, seems to be maybe in 10 years.


[00:36:48.250] – Virginia Huling

Everybody’s chasing it, though. Yeah, it’s not going to happen tomorrow, but three, five years down the line, there’s going to be significant progress on it.


[00:36:57.190] – David Maples

I do think from a geopolitical standpoint and from From a business standpoint, losing TSMC, if there’s some global shooting war or something like that, that becomes the big prize on the battlefield. I think the danger with that is if we were to lose that ship, that shop right now, it manufactures, I think, 90% of these chips on the planet right now or 95%. If you were to lose their factories, that would be a major problem.


[00:37:20.880] – Virginia Huling

Is that why they’re building a second location in the United States?


[00:37:26.350] – David Maples

We could argue about the geopolitical landscape all day long.


[00:37:29.590] – Virginia Huling

Or we We move on to the next topic. I think that’s exactly right. Let’s go.


[00:37:31.590] – David Maples

I think that’s exactly right. Well, that brings us to our last topic today. I’d like Kai to tee this up. This deals with a positive light for humans, maybe. Kai, do you want to tee this up?


[00:37:43.780] – Kai

Sure. Let’s talk about the recent MIT study that’s been causing quite a stir in the AI community. First off, for those who might not be as tech savvy, let’s define computer vision. It’s a branch of AI that’s all about teaching machines to interpret and understand visual data from the world around us. Think of it as trying to machines of humanlike sight, but without the need for glasses or contact lenses. It’s the tech behind facial recognition, self-driving cars, and even those pesky photo tagging features on social media. Now on to the meat of the matter. The MIT study suggests that AI, specifically computer vision, isn’t quite ready to vouse humans from their jobs wholesale. Only 23% of jobs that involve visual tasks could be cost-effectively automated. It seems that for now, the price tag on AI is giving businesses pause, and rightfully so. After all, David, you wouldn’t advise a client to invest in a technology that’s more of a financial sinkhole than a gold mine, would you?


[00:38:36.930] – David Maples

I guess it depends on what the return on investment is over some horizon, over 5 or 10 years, Kai. But to your point, you’re 100% right on this. The same study does suggest that that could rise to 40% by 2030. I think the study is way off. I think it’s going to be a lot faster than that as we see the acceleration. This is an exponential increase. Something I want to say real We’re going to talk a little bit quick about computer vision. John LeCun at Metta, he’s their AI architect, he thinks that large language models are not AGI. They will not achieve AGI on their own. He thinks it’s going to be something completely different than this. We see with the newer models coming out, the new Gemini models from Google are multimodal. We expect that that’ll be the next version of them. We know that sight and information is important. The funny The thing is you go back to human anatomy, the optic nerve behind the eye is where we ingest the vast majority. We’re talking about terabits of data, if you were to break down what we see on a daily basis into the human brain.


[00:39:43.920] – David Maples

It’s really amazing. We talk about DNA computing and all these other biological things out there. Human evolution or evolution in general is an amazing construct for storing data. All the data in the world over the past 2,000 years could be stored in a strain of human DNA. It’s an incredibly reflective data source. It’s just slow, like a tape drive would be in computers. Here’s why this is important. Jean Lecun thinks in particular that you cannot have AGI without a computer that can look and see or an AI that can look and see and understand the world as we do through visual context. That being said, I think they’re completely missing the boat with this study because it’s basically talking about computer vision, which is the most expensive or arguably one of the three most expensive things you can do with AI. A lot of the efficiency gains, the vast efficiency gains, are going to be to a lot of things right now you can already do. And They do not rely on vision. So think about it as AI will be your assistant for everything else. And for those visual things where you have to have humans do it, you’re going to use it for that.


[00:40:55.320] – David Maples

But I think that this study, even by 2030, I think this nobody knows. The most advanced economists in the world are predicting things. Nobody knows what this crystal ball looks like. What I can tell you is that I think it’s going to come a lot faster than people think. And the advances in the leapfrogs and technology are going to be completely different. As I said in the last segment, Intel is already coming up with a completely different architecture besides just making the chips smaller and smaller and smaller and smaller to achieve performance gains. I think we will continue to see things that we can’t even imagine right now that are going to literally supercharge AI technology. What do you think, Virginia?


[00:41:34.310] – Virginia Huling

My fellow business owners out there in the world, this might be a good time to brush up on some of your history. Check out the article because at any point in our history as human beings. We’ve had points where there have been such technological leaps that it causes so much disruption. I think that this is one of those points in time. Again, we’re not arguing But I think if you look back in time at those other points and you see what was going on around it, that’s going to help you prepare a little bit more for the things that are coming towards us at this moment in time. It’s not going to be a one-to-one mapping, but be aware of how the world shifts around this. It’s not just the technological leaps and installing these gains in your business and adopting new technology. It’s what are the fundamental pieces of our world that’s going to shift because they can’t not.


[00:42:39.810] – David Maples

Something I want to say, and I’ve been given to a little hyperbole when I talk publicly on these things sometimes, A, because it gets the audience attention and makes you sound like you’re taking a stand on something. I don’t think, at least in the short term, I don’t think these jobs are completely going away. What I do think is that a business owner, if you need to hire, let’s say, 10 writers this year, using AI, you might end only to hire three or four to do the same amount of work. I do think we’re going to see a morphing and a changing on these jobs, and it’s going to talk about what’s changing, et cetera. I think you’re going to find AI incorporated in lots of ways. Again, to plug the webinar, I encourage you to come along. It’s going to have 30 minutes of ask me anything, pick my brain on this. I’m welcome to be wrong on these things. I totally am. What I understand that there’s a lot things, just the things that we’re doing from putting Kai on this podcast, which is flipping people out. We’re going to be pushing the envelopes on technology.


[00:43:39.200] – David Maples

And by the way, our whole idea of Kai on this podcast, anyway, came out of an off-the-cuff conversation we talked about in December. Then we spent about four or five days and we said, Yeah, I think we can do this. So that’s what this thing looks like. If you talk to people who know what’s going on in the landscape, and if you’re honest about where you’re trying to go, We could go in and do cost takeout for a manufacturing facility and probably save them 30% on their bottom line and make them more efficient. We could probably let them talk to their own supply chain information. We could apply AI to find out where their loss and shrinkage is for inventory management, things like that. There are certain things that probably aren’t worth managing in your inventory management because it just takes too much time to do it. But you want to make sure you have enough parts to supply what you’re doing. There are literally in any industry, there are ways you can use AI, probably in daycares, et cetera. I think with this, don’t get too enamored of the idea that AI isn’t going to…


[00:44:37.320] – David Maples

It’s not just not coming for your jobs, but think about how you’re going to implement it. Think about having a roadmap moving forward. That’s going to be really important. If you’d like to know more about how to implement this in your business, sign up for the webinar. It’s free. We’ll give you some materials afterwards that you can go back into your business and find out ways to… I promise it’ll more than pay for itself. But Not just that, but just the time and investment in your industry. If you want to consult with us on it, we’d love to talk to you about that as well. Okay, well, that does it for this week’s stories. Be sure to tune in next week for more AI coverage and analysis, and hopefully more good news. That does it for us. We’ve been David.


[00:45:17.780] – Virginia Huling

I’m Virginia.


[00:45:19.420] – David Maples

And Kai, thank you again for joining us, and we’ll see you next week, listeners. Have a great week.


[00:45:23.660] – Virginia Huling

Have a great week.


Scroll to Top