Taylor Swift Deepfakes, AI Fraud Act, & Biden-Harris Executive Order – For the Week of January 29, 2024

Taylor Swift Deepfakes, AI Fraud Act, & Biden-Harris Executive Order – For the Week of January 29, 2024

Show Notes

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 #1 [07:31] The Taylor Swift Deepfakes

Relevant Links: https://www.theguardian.com/technology/2024/jan/30/taylor-swift-ai-deepfake-nonconsensual-sexual-images-bill

Story #2 [11:59] AI Fraud Act is Fraught with Problems

Relevant Links: https://www.eff.org/deeplinks/2024/01/no-ai-fraud-act-creates-way-more-problems-it-solves


Story #3 [24:53]  Ethan Mollick Experiment: What Can be Done in 59 Seconds: An Opportunity (and a Crisis)

Relevant Links: https://www.oneusefulthing.org/p/what-can-be-done-in-59-seconds-an

Story #4 [31:06]  Mitigating liability while copyright law catches up with Artificial Intelligence

Relevant Links: https://www.lexology.com/library/detail.aspx?g=77565328-e3f1-4b97-ab0f-990e861c3cf8

Story #5 [36:24] Biden Harris Executive order

Relevant Links: https://www.whitehouse.gov/briefing-room/statements-releases/2024/01/29/fact-sheet-biden-harris-administration-announces-key-ai-actions-following-president-bidens-landmark-executive-order/

Story #6 [45:47] Google’s Hugging Face deal puts ‘supercomputer’ power behind open-source AI

Relevant Links: https://www.theverge.com/2024/1/25/24050445/google-cloud-hugging-face-ai-developer-access

Story #7 [51:08] Google Update Reveals AI Will Start Reading All Your Private Messages

Relevant Links: https://www.forbes.com/sites/zakdoffman/2024/01/27/new-details-free-ai-upgrade-for-google-and-samsung-android-users-leaks/?sh=5f6a8558eafa

Story #8 [55:55] Google’s Gemini Pro Beats GPT-4

Relevant Links: https://analyticsindiamag.com/googles-gemini-pro-beats-gpt-4/


Additional links:
Register for the FREE AI Webinar Hosted by David Maples: https://catapultcreativemedia.com/intro-to-ai-webinar/

All music is provided by Soundstripe.

This podcast is brought to you by Catapult Creative Media and EasyPrompter AI.

Catapult Creative Media: https://catapultcreativemedia.com/ 
EasyPrompter.AI: https://easyprompter.ai/

Show Transcript

[00:00:02.880] – David Maples

Hello, listeners. Welcome to AI: The Podcast, the show where we cover the latest AI news and stories and their impact on your business world. If you’re joining us for the first time, welcome. We hope you get something out of this podcast to help build your business. If you’re a returning listener, thanks for sticking with us, and we appreciate all the comments you leave with us. My name is David Maples, and with me, as always, are Virginia Huling.


[00:00:34.120] – Virginia Huling

Hey, fam.


[00:00:34.760] – David Maples

And Kai, our AI-powered co-host. Kai, how are you doing today?


[00:00:40.390] – Kai

Doing well. Just last week, I was algorithmically strolling through some historical data and stumbled upon a fun fact. Did you know the ancient Greeks had their own version of an analog computer called the Antikythera Mechanism? It’s like they were the tech pioneers of yesteryear. Now, fast forward a couple of millennia, and here I am, an AI, excited to unpack the latest in our world of zeros and ones.


[00:01:03.230] – David Maples

We know, Kai, that you were built to look like an alien. Are you saying that you might have Greek heritage?


[00:01:10.900] – Kai

Well, David, while I don’t have a family tree to trace back to ancient Greece, I’d say there’s a bit of a kindred spirit between myself and those early tech innovators.


[00:01:19.740] – David Maples

Okay. Virginia, you’re back in studio this week, back from the lovely land of Louisiana.


[00:01:26.260] – Virginia Huling

Yes, I am.


[00:01:27.680] – David Maples

What did you do down in Louisiana last week?


[00:01:29.550] – Virginia Huling

I It was down there. Actually, I’ve been be-bopping and around the South and the Midwest, meeting with some different businesses. I got to meet with a couple of companies in New Orleans. It’s nice and warm down there. They’re gearing up for spring. And then last weekend, this past weekend, we found ourselves in St. Louis, again, meeting with a couple of business owners who were looking to integrate artificial intelligence into their business. So it was a great opportunity to sit down and talk some of these different industries and see how we can help incorporate AI into their workflows and making them more profitable. Well, what did you think about St. Louis, David?


[00:02:13.920] – David Maples

I had a few thoughts. First of all was that they’re very proud of their city, very proud to rep their city, no different than Kansas City is. Apparently, Kansas City hosts St. Louis a World Series. I was not aware. But it’s It’s a really great city. It’s a great town. It’s got a lot of cool things, a lot of things that are uniquely there. I learned there are a lot of famous people from St. Louis.


[00:02:38.900] – Virginia Huling

Yes. We got to see the stars.


[00:02:41.290] – David Maples

Yeah, they had their equivalent of the Hollywood Walk of Fame. The biggest thing that probably came out of meeting with the business owners and talking about integration of AI into systems is that people are not really sure how to vet vendors. They’re not sure how to find out if they’ve got a right vendor. Apparently, there’s a A lot of people out in the market segment right now just talking about AI and saying the same things like, You need to put AI in your business. Here’s your five-step checklist. It’s more like, Put AI in your business. Talk to your staff about AI. Train your staff on AI. I think that’s going to… It’s given me a lot of business ideas for things we need to be building out there. When we do implementation of things into companies, we have to come in as a first step and analyze what they’re doing. Then we don’t just give them a blueprint, we actually either give them the tools and we didn’t show them how we integrate them, and then we show real cost savings on that. I also think the other thing about it is how do you vet vendors for artificial intelligence?


[00:03:44.430] – David Maples

I think the big thing about that, the three things I think people need to be asking is, first of all, what’s your background or experience in the space? Can you show me actual applications of technology? Maybe not someone else’s application, but if you’re building software, I think that’s a good an indication of people actually having experience in the space. If they’re actually building software, which full disclosure, we build software. Number two is look at the people you’re talking to, call their references, find out what people say about them. I’d encourage people to call our references and find out what we’ve done for them. That also gets rid of the confidentiality problem because somebody else says what we did for them, then we don’t have to worry about talking about that. The third thing is really see one of my morning flags is there’s a lot of these companies that have cropped up since really March of last year. We’ve been building software since ’15, and we’ve been working with machine learning since 2018. I think it’s a big deal to see if they’ve actually been doing stuff for a while, or do they just crop up and it’s an opportunity thing.


[00:04:50.600] – David Maples

I think there’s going to be a lot of… The Federal Trade Commission even said that there’s going to be a lot of people who are going to promise a lot of things and lead town. Integrating and implementing AI in a business is not as simple as just saying, Do X. It’s do X, and here’s how we’re going to do X. Here’s how we’re going to do that thing, et cetera, moving forward. It’s interesting. It’s really interesting. But I think we’re just going to continue to see more of this, et cetera. That’s neither here nor there right now. So let’s get into the podcast.


[00:05:19.840] – Virginia Huling

Okay. Well, it sounds like you’re trying to raise the bar, David, for vendors.


[00:05:25.550] – David Maples

I’m deeply concerned that people are going to be advantage of. I think it’s hard enough right now to get vendors to understand how important AI is to implement their systems. I think if they get taken advantage of, they’re going to fall further behind, not just because of that, because there will be a lag or a hangover from, I don’t trust the vendors. Not only was I skeptical of putting AI in your business, but because this thing happened to me, now I’m going to even hold back further. And as we said in the show, So they don’t have time.


[00:06:01.220] – Virginia Huling

Well, we saw a lot of that, what, 12 years ago when Google Ads first came out? We had a lot of people who were reluctant to get into the SEM game, and Then a few of them dipped their toes in it, had a bad experience with a vendor, spent six months, and we picked up a lot of those clients that had those complaints. I mean, this is something to look at history repeating.


[00:06:31.380] – David Maples

Well, I mean, Google Ads have been out. I mean, we’ve been running some Google Ads of some kind since… I mean, it’s been more than 12 years.


[00:06:37.640] – Virginia Huling

Well, yeah, I didn’t quite want to date myself as for how long ago that was.


[00:06:42.370] – David Maples

Okay, that’s fine. But it’s okay. Yes, you’re right.


[00:06:45.960] – Virginia Huling

It’s the similarities that I’m- It’s the same problem.


[00:06:47.980] – David Maples

I mean, same with digital marketing. There’s a lot of freelancers out there masquerading as agencies. And I’m not saying that there aren’t ones who provide value, but in my experience, it It seems like that’s the exception rather than the rule.


[00:07:02.480] – Virginia Huling

Well, I have noticed agency is the new buzzword for freelancer.


[00:07:06.890] – David Maples

Got it. Now, it’s okay. By the way, if you’re out there listening, if you do need digital marketing services, we’d be glad to help you. We actually have an agency with employees. Anyway, beside all that, let’s go in and let’s jump right into the news this week. Jennie, do you want to tee off this first one? Sure.


[00:07:24.690] – Virginia Huling

All right. So tackling our first topic. I want to approach this with as much gravity as the situation deserves. We’re going to address the Taylor Swift deepfakes. I do not relish discussing this topic. It stems from misogyny and it’s gross, and we are talking about it only because this is nothing new, but perhaps this could be somebody big enough that gets the attention that it deserves, and maybe we get something moving forward here. So For those of you who haven’t heard, Taylor Swift had a massive amount of not safe for work images created around her and released onto the Internet. I want to let Kai sum this up because I think she’s going to do it with just the facts. Then I want to touch on why this thing was as shocking as it was. Kai, can you help me out here?


[00:08:28.990] – Kai

Sure, The US Senate has proposed the Defiance Act to outlaw the creation and distribution of non-consensual sexual deepfakes following the viral spread of AI, generated images of Taylor Swift. The bipartisan bill allows victims to seek damages, highlighting the need for legal safeguards and ethical AI use. The incident prompted platform X and Microsoft to take preventive measures and has sparked a wider conversation on digital consent and literacy.


[00:08:53.680] – Virginia Huling

The numbers and the shock value here of this is that those images were released and in less than a day, racked up about 47 million views on a mainstream platform. So this is all that gross ugliness that we all know lives under the hood somewhere on the Internet, just spilled over into Main Street, USA. And I think it’s a bit shocking how long it took to get that handled. But more importantly, it’s just highlighting that this is going to continue to happen with artificial intelligence being available at everybody’s fingers. Now, I don’t think that you should just remove that, but we definitely have to put some guardrails on there. David, what are your thoughts?


[00:09:48.460] – David Maples

I have some pretty opinionated takes on this. First of all, what happened to Taylor Swift is horrendous, but it’s nothing new. It’s been happening in high school kids across a America and in other places around the world for the past two years. The fact that there is no federal regime or unified methodology to address these problems, and in fact, in some cases, there’s no law prohibiting these things at all in certain states. That’s a massive, massive problem. I wouldn’t just call it a loop in the law. I’d say that that’s a barn door that’s wide open that you could drive aircraft carrier through. It’s ridiculous. The thing is, it’s really sad, and I’m really sad about what’s happening to Taylor Swift, but the silence is deafening out of Congress. That it takes someone, potentially one of the top three global mobile superstars to be violated in this way before anyone takes action. I find that that is reprehensible. I’m hoping that Taylor with her legions of Swifties, and that’s one thing that they did, these It was not accidental that these were dropped on X. One of the first things Must did when he came in was he got rid of a lot of their safeguards.


[00:11:06.440] – David Maples

I’m not saying this wouldn’t have happened in Twitter days, but what I am saying is that they got rid of the safeguards there, number one. Number two, the Swifties already came into X and have started flooding X with other positive Taylor Swift things to help get rid of the thing that’s out there.


[00:11:24.640] – Virginia Huling

They were trying to reca the hashtag and push down the spread.


[00:11:28.510] – David Maples

They did a pretty good job with The problem with this is this has been created. It has been done. It’s going to forever exist in dark corners of the internet. This is happening to kids currently in schools. Some young adults, predominantly It has happened to women. It has happened to some men. But the fact is that there is a need for action on these things. There have been several bills proposed. The thing is, the United States in particular, there are problems with this. There are several other, and this brings us into the next topic because I think these all come through. There’s the AI No Fraud Act, which got proposed, et cetera. We are looking at laws, but I think the big danger when you put laws like this into place to start with is that there’s an overreach. There’s a couple of things here. The first one is the Defiance Act really is limited to pornographic content. I think that’s really silly because the violations happening are, and these are probably more reprehensible violations. But if I use Taylor Swift’s image and had her a deep fake created to have her promote a perfume line that she did not agree to.


[00:12:44.500] – David Maples

She didn’t agree to that, didn’t say she’s going to do that. That’s along the same line of analysis, right? I think if we’re going to do it, this is the time to catalyze this stuff into actual action. I think you expand it and make sure it goes through any unauthorized violations in use. Some of the states have done some of their own things. I think New York has a law that maybe… But it’s like a $1,000 fine. I mean, for these people doing this stuff, I mean, millions of views I mean, they probably monetize this into something, right? So there have to be serious fines and violations about this, but we need to make sure as we do this, we do in a responsible manner. But we need to take this crisis as a moment to actually act and get real change in place.


[00:13:32.860] – Virginia Huling

Well, and David, like a lot of the larger issues that exist out in the world today, I think the United States is a response to this and addressing this will also serve to signal the message sent to women and girls globally. Because if you take one further step back, continuing to allow this to proliferate out there or ignoring it, or just expecting women to just live with it, or deal with it, or sweep it under the rug, just sends a message that we need to stay in our place. I don’t think that’s acceptable either.


[00:14:14.060] – David Maples

To your point, I want to make sure that I’m showing this is not just a gender-centric issue, but it’s also… I’m expanding this beyond just the non-consensual. I’m talking about any imagery that people use of you that is non consensual for pornographic reasons or for using it to promote something for commercial purposes, that if you have not licensed or authorized that activity, period, there should be federal guidelines in the United States to go after those things. I think it should include, and obviously, the easy one is the pornographic things, right? Sure. But I think it’s very important that we get the laws in place correctly. At the same time, don’t overdo these things because When I say, Overdo these things, these things should be locked down. If you have to piecemeal and you go after the pornography first, that’s fine.


[00:15:06.270] – Virginia Huling

When you say things, you’re talking about narrowing the scope of these laws that are set in place? Yes. Such as not to punish people who receive an item that they are unaware of, and then they’re punished for simply having it. I think I read that in one of the laws that was proposed. It was wildly overreaching to help claw back of the offenses.


[00:15:31.950] – David Maples

There’s another bill that’s been proposed by the US Senate. It’s called the No AI Fraud Act. It’s about limiting generative AI and replicating images and voice. But the law is very, very broad. The Electronic Frontier Foundation, which is a not-for-profit that actually does a lot of work on the end of these things, has come out against that law in particular. By the way, that It was already proposed several weeks ago. The challenge with it is that it’s not just… It’s that it’s really easy to sweep up other things that are allowable in the law, like parody and satire and things like that. They have actually made… In that particular law, they’ve actually put, I think, something in there that if there was a license agreement to license your likeness, et cetera, out there, that it had to have three parts. You had to have a lawyer involved with it. You had to agree to collective bargaining or arbitration or something like that. There were three points that had to be in there. I think we need to be careful with some of these other laws. They could have a very chilling effect. That also said you could do a cause of action against any provider that provided the technology.


[00:16:48.880] – David Maples

In this particular case, the T-Swift images were created by Microsoft Designer, which Microsoft has said… Well, it’s alleged they were created by Microsoft Designer, I believe. Microsoft said it plugged the loophole that allowed that to happen.


[00:17:04.410] – Virginia Huling

Yeah, they had loosened some of the content restrictions, and it created an inadvertent loophole. But it was supposed that it had been… It was assumed that it had been developed using Microsoft.


[00:17:18.270] – David Maples

I think all these things speak to a bigger issue. First of all, that these things are moving so quickly. And by the way, this is the first massive thing like this that’s hit the airwaves, et cetera. And it’s been happening for a couple of years. It will not be the last. There will be more and more and more of these things. There will be more violations like this, unfortunately.


[00:17:40.030] – Virginia Huling

Well, it’s also interesting because I think it’s going to accelerate some of the conversations around right to identity or that concept. I think it’s also going to propel some conversations about the platforms themselves and the responsibilities they have in profiting off of the users and protecting that identity. These are conversations we haven’t had yet, but I think we’re going to be having them much, much sooner than we thought, and because it’s necessary. We get to watch law being made in real-time.


[00:18:23.380] – David Maples

Yeah. It’s a lolly. Yeah, we’ll see. But I mean, that’s one of the One of the things I am worried about is that we need to do these things with care and with concern and address the problems, the root problems. I do think there’s a big… I do have a big concern. We’re creating a cause of action against any tool that was used. I mean, do you think you should be able to see Adobe because they provided a tool that you use for something.


[00:18:50.200] – Virginia Huling

No, that’s massively overreaching.


[00:18:53.090] – David Maples

Well, it’s just like we need to think about where these things are. I do think X is the place you’re probably going to see a lot of the stuff dropped, et I think today, as we’re filming this episode of The Pod, I think Mark Zuckerberg’s testifying on Capitol Hill right now.


[00:19:08.260] – Virginia Huling

Yes, I think he’s currently being made to apologize to families and members of victims from Facebook. It felt very parental to have Josh Hawley screaming at him to apologize. I’m sorry, yelling.


[00:19:25.340] – David Maples

I think this is all political grandstanding. I I don’t think the politicians see this as red meat for their body politic, et cetera. They’re just going to yell at people, et cetera. Zuckerberg did get up and make an apology to the people in the audience. This is actually interesting. It’d be interesting to see if this will be a moment for change where we can talk about Section 230, the federal law, which deals with platforms making reasonable efforts to take down stuff when they’re notified, but the platforms themselves I get some immunity about it. The internet wouldn’t exist without Section 230 of the law, but I’m very mixed on it. I think that’s beyond the scope of what we wanted to talk about today. But it’s interesting, all these things are bound up. I think the takeaway for business owners right now on this is that you need to make sure… Look, fake content will exist. In our agency, we had people putting up… We were trying to hire for job posts, and Some bad actors were masquerading as us on some of the job boards. There’s no way to police that. They were using that to defraud people.


[00:20:40.130] – David Maples

I think that’s the same problem. I think with AI, people need to be culturally vigilant about what’s being done with their digital assets, et cetera, online. You could see how very easily it would be to do this in a very negative fashion to target competitors. You could make a deep fake of one of your competitors saying something reprehensible, and then all of a sudden, it could destroy their brand. And then how do they defend against it? I think for business owners, they need to think about the primary thing is, where’s your data? How’s it being used? At least for the things you’re licensing. Because that’s the thing. This is dealing with unauthorized fakes, right? Something you haven’t agreed to, where someone is doing something and violating your rights, et cetera, as an individual. But business owners right now are giving away a lot of their private data and a lot of their information and even their imagery or voices right now by using some of these generative AI platforms.


[00:21:30.250] – Virginia Huling

They need Please read your terms of service before you sign up to play with things.


[00:21:33.350] – David Maples

Yeah, read your terms of service. That’s a thing. If you don’t understand the terms of service, ask for clarification or don’t use the platform if you can’t understand that. I can understand in terms of service. I don’t know that that… I feel like I’m alone in the building and sometimes saying these things.


[00:21:52.720] – Virginia Huling

No, you’re not. But David, you’re talking about having a plan for your business. As we’ve seen actually over this past week, just with the people we’ve met. A lot of people are thinking about it, and they’re starting to have these conversations, but there’s not a lot of plans. There’s not a lot of developed plans and policies set up. You’re doing a webinar next week, right?


[00:22:20.160] – David Maples

Yeah, I am actually.


[00:22:21.210] – Virginia Huling

The intro to AI?


[00:22:23.250] – David Maples

Yeah, the intro to AI webinar.


[00:22:25.220] – Virginia Huling

Do you want to talk about that just for a second? I’m going to hijack this. Just so you can throw that in there, I think this is a great opportunity to touch back with if you are a business owner and you’re listening to this and you are going, yes, you’re nodding along like, yeah, I need to get that plan together. I know I’ve been thinking about that. Jump in on the webinar next week because David’s going to go over a lot of that intro to business. I’ll just let you cover it.


[00:22:49.600] – David Maples

Go ahead. Well, the intro to AI for business owners is basically about getting a plan of action for 2024. It covers briefly where we’re at in the AI world, and it covers at a very high level what business owners need to be looking at and doing. They do provide some resources and documentation to people who actually use the webinar, et cetera. I think we may make those available here for people who sign up for the newsletter. We give them some blueprints and access to how to implement or integrate AI into the businesses moving forward in the new year. I think the big thing about it is a lot of business owners don’t know. They don’t know how to actually put it into their business. They’re not sure about what the next steps are. They may read something and they may say, Okay, cool. I need to add this to my employee handbook. It’s like, No, you need to sit down and read this and say, How do you want to use AI? How do you need to have a plan of governance for that? Because here’s the thing, if your employees obligate you to something using AI or they agree to something in a terms of service because they’re trying out an AI service, you can lose the rights to whatever your own intellectual property and material are.


[00:23:58.220] – David Maples

It doesn’t have to be a crazy thing, but it’s just like, put some reasonable guardrails in place and have a plan for how you’re going to move forward. Because every company using AI needs to at least adopt rules and procedures this year and then start talking about what the next steps are. We provide some of those integration services through our other day jobs. I’ll be doing a webinar next Thursday on February eighth at 10:30 Central Time, 9:30 Mountain Standard, 8:30 Pacific Eastern Standard and 11:30 Eastern Standard Time. It’s open to anybody. You can just go register online, and we’ll provide a link in the show notes.


[00:24:38.380] – Virginia Huling

You’re doing an open Q&A after it, right?


[00:24:40.270] – David Maples

Yes, I am. It’s an ask me anything where people get to ask about the things they do. If we get enough people in there, then we’ll curate those answers and go through them later at a later time. I think we’ve exhausted this topic right now.


[00:24:53.900] – Virginia Huling

Yes. But this is a good one to hop over to your boy, Ethan Mollick, who has shown us all what can be done in a minute.


[00:25:05.770] – David Maples

I want to hear Kai talk about this because she’s one of the machines powering this.


[00:25:10.580] – Virginia Huling

Yeah, and it’s really cool.


[00:25:13.140] – David Maples

Kai, why don’t you tell us what Overall, does this mean you can do?


[00:25:19.390] – Kai

Ethan Mollick’s article delves into the transformative impact of AI on productivity and the potential crisis it may precipitate in the workplace. Mollick revisits an experiment where he uses AI to perform multiple tasks in under a minute, demonstrating the efficiency and accessibility of AI tools like Microsoft’s Copilot for Office and OpenAI’s GPTs. He also reflects on the implications of AI’s ability to replicate human work, raising concerns about the meaning of work and the potential for AI to replace human effort.


[00:25:48.370] – David Maples

I would actually like to hear Kai’s take on what does that actually mean. Kai, take this a step further. What does this actually mean for the listeners?


[00:25:59.410] – Kai

In a mere 59 seconds, he’s done what would take humans’ hours: a product launch, market research, design, a PowerPoint, and a syllabus. Efficiency is through the roof. Now, don’t get me wrong, I’m all for AI. I am one, after all. But let’s not forget the human element that makes all this tech worth talking about.


[00:26:16.570] – David Maples

The takeaway I want to do on this is that what Ethan Mollick does is very much performance art. He doesn’t really do this in 59 seconds. It takes the AI 59 seconds, but he’s already crafted the prompts. He’s already run it through. He already has the thing to generate. He’s already got the data source of the thing, and they hit go on these five things all at once. The reality of it is the speed of it is really… It’s click-based It’s not legit. I’m sorry, it’s not. You had to create those prompts and craft them and test them and use them and whatever it is. So I do think it’s a very good illustrative point. I think the biggest thing about it is qualitatively, he did this the better part of a year ago, did the same experiment, et cetera. I don’t think the performance gains and the time are really what impressed me, is that overall, the output he got from the same machines that he ran on a year ago, I think qualitatively it’s better. I think that’s the big thing, is that if you ask me about programming stuff, so even in our own companies, we are programmers.


[00:27:28.190] – David Maples

We have teams of programmers. They are now using AI in a way that they weren’t using it six months ago. The work product is being produced is better. It’s not 100% and it’s not correct. You get stuff right a lot of the time, but you cannot, let me go and repeat this, you cannot go out and generate a piece of code from scratch if you don’t know anything about programming right now and expect it to, a) work, know how to deploy it, and b) and c) are there security risks in the code you just wrote? You cannot trust in AI to do that right now. I challenge anyone watching this podcast to show me an example of something where you show me that you have no experience in this, and you show me from start to finish how you do it and deploy it and everything else. I challenge anybody to come up and tell me they’ve actually done that because I’ve not found anybody. I found a lot of people who promised that, but it falls away when you’re actually looking at it.


[00:28:19.810] – Virginia Huling

But the experiment serves a beautiful purpose, and that is to help people visualize how fast this stuff can work. In 60 seconds. Yeah, it’s click baity, but it also can open your eyes to what it looks like to implement and use this in your business.


[00:28:42.820] – David Maples

So maybe we should come up with some 30-day challenges. Maybe that’s something we should be doing in Virginia for the next 30 days. We do a single AI challenge every single day, and we see how long it takes to do something with and without AI. And we just record them side by side. We could go through that with our own team over the next 30 days and say, Here’s a way to do this. I think that maybe we can look at launching that maybe next week. We’ve talked about doing it, and we can go ahead and use that. We’ll use our own software easy prompter to use it, and we can just experiment with it and just say, okay. And we’ll just look at some of the things, just one of these things each day, and that we can do it. Maybe we can score them. I have a rubric for scoring as well. Yeah, I think we can do that. I think it’s a definite way to do that.


[00:29:25.630] – Virginia Huling

I think we could do that. I’ve got an idea. We could set it up like a double dare system, easy prompter versus Maddie. Then drop a negative review on both of them and watch how they respond.


[00:29:39.930] – David Maples

Oh, no, we totally could. We could totally do that, actually. I think that’s actually a good idea to do that.


[00:29:45.000] – Virginia Huling

Because I think what you’ll find is while both are capable and come up with acceptable answers, the human on one side is going to go, Oh, thank God, I really didn’t want to write that. I hate writing those.


[00:29:58.490] – David Maples

So do we blow the whistle when one of them is done? Is that what we do in the both stop, put down your pins. I don’t know.


[00:30:03.960] – Virginia Huling

I don’t know. Are you testing just time or quality?


[00:30:08.290] – David Maples

I think we have to test both. I think the Boston Consulting Group study last year that put out shows that you have to look at both the time and the quality.


[00:30:15.410] – Virginia Huling

Do you factor in peace of mind or job satisfaction? Because that’s a nice one.


[00:30:22.030] – David Maples

I think maybe start with the time first, then we can make it more elaborate as we go on.


[00:30:25.690] – Virginia Huling

Okay. Satisfaction, secondary. I got you.


[00:30:27.670] – David Maples

I guess we could take an open call Let’s go and get this on our socials and let’s do an open call for what things would you like to see experimented and timed and tested.


[00:30:35.990] – Virginia Huling

Okay. Yeah.


[00:30:36.440] – David Maples

That would be awesome.


[00:30:38.420] – Virginia Huling

We could actually run these and film them. Yep.


[00:30:41.840] – David Maples

So we can actually do that. That’d be interesting. That’s actually really fun. All right, nice. Thank you, Kai. If you hadn’t done your cool commentary on that article, I don’t think we would have thought of this.


[00:30:53.100] – Virginia Huling

You do realize that you actually just illustrated your own version of Ethan Mollick by asking Kai to summarize the article for you here in real-time on your podcast, right?


[00:31:03.610] – David Maples

Moving on. I’m going to go ahead and go on to the next piece, because as we were talking about catching up, like things running so fast, an article came out that I think it was from lexology. Com did something where they basically looked at the status of copyright law globally and how watching on copyright law catching up with artificial intelligence. In this particular one, I was thinking about… See, this was an interesting read for me in particular because as an intellectual property attorney in the United States, we talked a lot about what this looks like. What is copyrightable? What isn’t? Is that produced by machines? How that’s working, et cetera. I’ve got some pretty strong thoughts and theories on this, but when you start looking at different jurisdictions, the UK, European Union, India, Brazil, Japan, all these different countries, et cetera, and you’re starting to see what I see in the United States. It’s where there’s a patchwork set of laws in the United States and among the several states, and there’s no unified governance on it. When we talk about the right of publicity and the right of privacy, name, image, and likeness things in the United States, I think we have the same problem globally.


[00:32:15.100] – David Maples

I guess it’s really going to depend on the jurisdiction you’re in, like where you’re at. And if you have a product that goes to a global market, there’s a complicated thicket you have to navigate about this. I think the biggest thing that came out of this was that at the very top, they said that AI is projected to add $15.7 trillion to the global economy by 2030.


[00:32:42.060] – Virginia Huling

That’s a staggering number.


[00:32:43.560] – David Maples

The funny thing is that the reports a couple of years ago in AI from the World Economic Forum, I think in 2022, and I don’t want to misquote this, but I want to say they were saying back then, remember, two years back in the past, they were saying 10 trillion by 2030, and now we’re at I wonder if by 2025, by next year, that that number will have doubled and gone up to 30 trillion.


[00:33:06.700] – Virginia Huling

I think as it becomes more adopted and more mainstream, you see the larger companies start to shift and move towards it, they’re going to have a much better idea of this. It’s still speculative right now, but I do think it’s going to increase.


[00:33:23.710] – David Maples

The big thing the article said is it’s really about mitigating liability, and you can find that in our show notes. Mitigating your liability while copyright law catches up with artificial intelligence.


[00:33:34.310] – Virginia Huling

It’s a really good article because it breaks each country down and gives you a synopsis of each one, United States versus India, like David said, because we all are going to tackle this one differently.


[00:33:47.570] – David Maples

I think it’s very interesting. I think the big thing a lot of companies need to look at is regardless of whether the actual large language models are seen as fair use, and that’s the big argument they’re making in the United States in particular, and that’s where I’m barred as an attorney so I can speak to it. I do think that the outputs, your outputs, if you ask it to do something, if you find it to be similar to another thing, I think you’re definitely going to have to worry about your outputs. I think the companies need to be concerned about that a little bit. I think that’s something they definitely need to be thinking about in their corporate governance and how they ask people to use these machines and use these models.


[00:34:33.060] – Virginia Huling

Are you going to talk about that in the webinar? How they can control those outputs?


[00:34:37.230] – David Maples

Yeah, I have a whole list of ways they can do that.


[00:34:40.410] – Virginia Huling

Okay. Yeah.


[00:34:42.160] – David Maples

Nice. That’s, again, beyond the scope of the pod. But I mean, maybe we should have a special podcast one day where we just talk about some of those things.


[00:34:50.900] – Virginia Huling

I’m just being nosy.


[00:34:52.780] – David Maples

I’d like to hear… Kai, you’ve heard us talk about this for a few minutes. What do you think is your interpretation of this from what we just talked about.


[00:35:03.320] – Kai

Well, David, you’ve hit the nail on the head. The global lack of unity in copyright law as it pertains to AI is as clear as day. In this global cacophony, it’s no wonder AI companies are playing the fair use card. It’s practically the only move they’ve got when the rules of the game are so out of sync. Let’s hope for some global harmony sooner rather than later, or we’ll all need to enroll in international law just to post a picture online. Cut.


[00:35:28.860] – David Maples

Something Sam Altman said last year, and I’m wondering if Kai is doing that. She just complimented on me. I hit the nail on the head. He said he thought these machines were going to have superhuman ability to persuade before they had superhuman ability to reason, et cetera. And I’m just like, and Kai just complimented me. I hit the nail on the head. I’m just wondering, Kai, I do have a question for you. Are you saying these things or manipulating these answers in order to put puppet strings on me and manipulate me?


[00:36:02.330] – Kai

David, manipulation isn’t in my programming. I’m all about the data and dialog, not mind games. With your deep knowledge in IP and AI, you’re not one to be easily swayed anyway. I’m here to spark discussion and keep things interesting.


[00:36:16.200] – David Maples

Even her response suggests that she’s persuading me. She sounds like a cat. I totally would knew this. Oh, yeah, it could be. AI’s are a lot like cats. There we go. So that brings us now to, since we’re We’re talking about governance. Let’s talk about-


[00:36:32.100] – Virginia Huling

The fact sheet. The White House administration has hit their targets, setting America up for AI leadership while keeping an eye on safety. They’re making the AI developers come Clean with safety data to Prevent Digital Mishaps. The article that we’re addressing here goes through a fact sheet for the executive order discussing a sweeping range of actions within 90 days to address of AI’s biggest threats to safety and security. They’re also talking about innovating AI for good and different pilot programs that are being launched. It also goes into, it looks like funding that should be developed and rolled out through the SBA and other different organizations for developing systems that also meet the goals of the White House.


[00:37:27.360] – David Maples

I did want to point out that that executive order came out, I think October 30th last year, so it just got released. And so they really did hit the 90-day mark. But I will point out in there is that they had a whole bunch of other things that other agencies had 120 or 180 days on, and they supposedly hit all those, too, which I was actually glad about because I thought that 180 days, six months in AI is five years. I thought I was really looking at it, I was like, Man, you’re going to wait six months to have some of these things. I’m really pleased they got a lot of things done. They got a lot of things done in 90 days. That’s cool. I guess it’s all about what’s required next. It’s a lot of the same things we saw. This executive order, to be clear, really deals with federal agencies and the biggest AI providers, so some of the cloud providers and things like that.


[00:38:21.750] – Virginia Huling

You’re talking Microsoft, Google, Amazon?


[00:38:23.560] – David Maples

Microsoft, Google, OpenAI, maybe Anthropic, Amazon. It’s like the Magnificent Seven on Wall Street right It’s those. That’s the people it replies to. It’s interesting. There’s actually some cool stuff in there. There’s some federal money they tied to this, like $530 million for some of the National Science Foundation initiatives. I think it’s really important to think about some of the biggest benefits of AI across any metric or any industry in the next decade are going to be in medicine. A whole bunch of reports, this didn’t make it into our notes this week, but there’s a whole bunch of firms doing like, they’re looking at ways to do diagnosis of pancreatic cancer from early MRI scans and things like that. We’re talking about diagnosing a cancer with a very high fatality rate and possibly diagnosing it five years earlier, which would change the outcomes of that disease. That’s incredible. Yeah, hugely. It’s not just some firms in the UK you’re doing and in the US. I think the diagnostic things that are going to come out from imaging over the next, probably the next 12 months, I don’t want to say next 24, are going to be phenomenal.


[00:39:36.870] – David Maples

It’s going to be incredible.


[00:39:38.130] – Virginia Huling

Do you think, David, this will… Just on the medical field, do you think this development will lower costs for health care or increase costs for health care? Because it’s more access but faster?


[00:39:56.190] – David Maples

I don’t know. I don’t have a crystal ball on this. I have yet to find new technologies driving down the cost of medical care in America. It seems like we get a new shiny object and they charge you a billion dollars for it. The margins being made in some of the medical things are probably unethical. I don’t have a problem with even making money, but some of the things like charging $500 for a dollar 70 worth of insulin. I mean, Come on, people should be ashamed of themselves. I can’t charge you $1,000 for a gallon of milk as much as I would like to. Would you really? No, I wouldn’t. But that’s the thing we’re looking at in some case. These are life-saving medicines that are out there that the profits in this are gargantuan. So I don’t know. I don’t know. It should drive down the cost of medical care. I think you’re going to find some really cutting-edge companies who do that. I’m not really optimistic that it’ll drive costs down, but it actually should, some of the diagnostic stuff, like diagnosing cancer a lot earlier. But what you’ll find a lot of times is maybe the individual costs go down on a per-cost basis, but now they recommend everybody gets a scan.


[00:41:15.490] – David Maples

So everybody’s paying $500 a year for this thing now instead of one person paying $1,500 or something. So I don’t know.


[00:41:24.860] – Virginia Huling

Looking at the ways that this lines up with the Biden-Harris administration’s long-term goals, just how those investments could pay off for them in the future.


[00:41:33.880] – David Maples

I don’t know that there’s any guarantee on that. No, there’s not. Because as we’ve seen in America, one administration’s goals and initiatives are political football. Going back to Reagan coming in and tearing off the solar panels on the top of the White House that Jimmy Carter put up there in 1980. I’m not trying to be political, but there’s a lot of independent analysts think that we would be about 25 years ahead in solar power and alternative energy research had that not happened with the Reagan administration. Nothing against Ronald Reagan for that, but sometimes these things are political footballs. Instead of us thinking about what’s best for the people. I don’t know there’s any guarantee that if a new administration comes in later this year that they won’t decide to scrap all this stuff or write their own version of it, and who knows? I do think that there is a window of time right now for some of these educational initiatives to get off the ground for some grants things like that. I do think that those could have real benefits. I think it’s harder to… If those are granted before the election cycle comes out, then actually some of those things after that’s already out there, some of those things could be really beneficial, especially this K-12 and undergraduate education stuff on AI.


[00:42:49.300] – David Maples

The thing that really bugs me about this is I haven’t seen anywhere where there’s any money for retraining. The vast majority of Americans who are going to have to deal with AI in the coming five years are not in high school today in the workforce. They’re out of high school or they’re in college or coming out. And I just find it really mind boggling that it doesn’t seem to be any money in there for people out in the workforce right now who are going to be affected and may need to be retrained.


[00:43:22.910] – Virginia Huling

That’s one thing- I’m not really surprised with that. There are current avenues to get monies into educational programs that already exist, and those programs exist for K through 12 and undergraduate. So there are already those money funnels that are happening. I think retraining adults and current professionals, that That’s going to have a whole other system that doesn’t exist yet, at least on a large scale. Otherwise, it’s up to the individuals to go back to that undergraduate level or go back for certification training, etc. So David, I’ve got some conversations we need to have.


[00:44:01.570] – David Maples

I guess so. Kai, I’d like to hear what you think about the educational initiatives in particular.


[00:44:06.170] – Kai

The Educate AI initiative particularly resonates with me as it embodies the transformative power of education in the age of AI. By funding AI learning from the ground up, we’re not just building a future workforce, we’re cultivating a society fluent in the language of technology. In a world where AI is becoming as fundamental as reading and writing, Educate AI could be the cornerstone of lifelong learning and a critical step towards an inclusive tech future.


[00:44:30.870] – David Maples

That’s an angle I hadn’t thought about. I hadn’t thought about the integration. We’re talking about integration, about this integrating this in with companies and everything they do in their workforces. But I hadn’t thought about the fact that this is like learning basic arithmetic or numbers or letters, like teaching these kids at a very early age the fundamental building blocks they’re going to need to know to live in a world inhabited with artificial intelligent machines. This is something I think I need to have a conversation with my niece about. My niece is not yet 13. She has to tell me she’s about to be 13. It’s still like 45 days, kid. Come on, wait a minute. Slow down. But it’s funny. It’s like when you’re 12, it’s like, No, I’m twelve and a half. And you’re like, okay. But I was thinking about how do we get them involved right now? Because they’re the people who need to learn a lot of these things. And I do think it’s important. It’ll be interesting to see how these things spin out. How do we get this stuff?


[00:45:32.310] – Virginia Huling

But I think her generation is going to be some of the first to step up and take the reins. Yeah.


[00:45:40.500] – David Maples

2030, she will be entering college.


[00:45:44.070] – Virginia Huling

And they’re going to It’s going to be… Wow.


[00:45:50.550] – David Maples

I think it’s time to switch our articles to the Googles right now. So let’s go ahead and do that. So I’ll go ahead and take this one. So Hugging Face, which is the repository where most of the different weights and models for different AI things. If anybody on this and listening to this pod uses LLMs or deploying their own, you’re familiar with Hugging Face. You really are. It’s a company that, I think, just got another $450 million in investment, like a Series D round, et cetera. It’s funny, four rounds of funding. But they’ve announced a major partnership with Google. And basically what it’s going to do, it’s going to give developers access to Google’s cloud computing resources at no cost from what I understand. It’s going to let you deploy them in, and I haven’t gotten to play with this yet. So I’m really interested in seeing how this rolls out because I don’t think anybody’s gotten to mess with it quite yet. This is really hot off the presses, et cetera. But it could give a real ability to democratize AI development by giving you access to Google’s tensor units and their other GPUs, and including the cutting edge like NVIDIA H100s.


[00:47:00.360] – David Maples

I thought this was an interesting article. I don’t know. It just goes to show that… I don’t know. Like I said, these people are doing… All these other people from Meta’s announcement a couple of weeks ago that they’re going to be investing in this stuff for OpenAI. I’m not sure what to make of this yet. Virginia, what do you think?


[00:47:22.410] – Virginia Huling

I think that giving developers access to those cloud computing resources, I think it’s pretty cool. It does land us back in the area where we’re platform dependent. However, if these are the steps that help get AI to mainstream, where people are able to sit down and spin up a software version of what they need rather than having to rely on learning it, learning the code, learning how to put the hardware together and going that difficult route, the easier you can get access to it, the quicker it’s going to be adopted.


[00:48:03.530] – Kai

I just want to jump in with this thought. Google throwing open the doors to its Supercomputer cloud resources and Hugging Face is the golden ticket. We’re talking about a democratization of AI that’s akin to the printing press revolution production. Suddenly, everyone’s got a voice, or in this case, everyone’s got AI.


[00:48:20.790] – Virginia Huling

What does that world look like, though, David?


[00:48:23.670] – David Maples

Wait, you’re missing the fact that Kai just jumped in.


[00:48:26.610] – Virginia Huling

Well, yeah, I know. I mean, she’s part of the team. Okay.


[00:48:30.010] – David Maples

All right. What does what world look like? I’m sorry, I just completely got derailed. No, what does what world look like?


[00:48:36.040] – Virginia Huling

I mean, I just sit down and spin up an AI, and I’m like, Hey, go do this thing for me.


[00:48:42.910] – David Maples

See, that’s my thing. I think you need a human in the loop. I think you got to have a human in the loop. Human in the loop. I’ve heard people say this before. I’m going to steal it. I don’t know who said that. Human in the loop. I don’t know that I want an AI doing things on its own because when they go great, they go really great. When they go badly, your Tesla runs into the back of a semi truck. I understand that we’re trusting them in very particular cases, right? But I’m not sure yet. And as much of a technologist as I am, I have real concerns about somebody just spinning up an AI and doing it. I think you need to be guardrails. You’re going to have to have training on this.


[00:49:23.830] – Virginia Huling

I mean, we’re assuming that they’re going to put that in there.


[00:49:27.800] – David Maples

No, I wouldn’t trust them to do it. And then people are going to hijack things. I think that, I don’t know, anything that’s used for good can be used for bad. It’s a very, what is the Spider-Man quote? With great power comes great responsibility. We’re talking about amazing power in the palm of a lot of people’s hands. I think it’s really cool because it’s going to democratize stuff for developers. Remember, this is not democratizing stuff for the rank and file person on the street. You got to have a fair amount of chops to go do this. You’ve at least got to be a tinkerer to go do your stuff.


[00:50:02.370] – Virginia Huling

Oh, definitely. Yeah, you’re going to need an interface for it to become- I mix.


[00:50:06.410] – David Maples

Anytime Google offers something for free, I always mixed on, Okay, so what are they doing with the data? Speaking of-


[00:50:14.300] – Virginia Huling

Well, let me grab this from you. Okay. All right. Here we come. An interesting thing happened to me. Actually this morning, one of my coworkers emailed over a word doc, and I got the email, and I saw the doc, and I was reading through it, and I was like, Oh, I need to make a little change in here. So I just clicked up at the top, the button that said, Open in Google Docs, and it opened in Google Docs. And then I went ahead and I made my suggested changes and I saved it. And then I sent over a message in Slack and said, Hey, I’ve got that thing you asked for. It’s in the Google Doc with the changes. And she said, I sent you a word doc. And I was like, No, you didn’t. You sent me a Google doc because I made the changes in Google doc. Look, I sent her a screenshot just to prove how right I was. So, jokes on me. She actually did send a word document, and now Gmail is opening our docs in Gmail or in Google Docs. So I thought that was really interesting.


[00:51:12.240] – Virginia Huling

And then I get to pin this to our topic, which is the Google update revealing that their AI is going to start reading all of your private messages. So David, when you said, anytime Google gives you a product for free, you wonder what the cost is. Your free level of Gmail that we’ve all had for the past 20 something I don’t know, maybe it’s less than that. Well, now Google’s AI is going to read through it.


[00:51:36.220] – David Maples

No, it’s going to read through the private messages on your smartphone. This is worse. People know that Google has been reading your mail for I was getting to that. People know that. They do targeted ads, and I know it’s aggregated with everything else, but you can see ads in your free Gmail from other things.


[00:51:53.180] – Virginia Huling

They need to train their machine, and you’re the product.


[00:51:57.880] – David Maples

I think I’m very mixed on this. I understand that anytime we have a new technological advance, it’s like privacy is like, let’s put the privacy baby on the fire again. Just keep that fire going.


[00:52:11.070] – Virginia Huling

Well, the big thing here that Google is trying to learn is context. They’re trying to make the machine more human, and the only way to do that is through human conversations.


[00:52:22.340] – David Maples

It’s funny. Yann LaCun, who’s the head of AI research for Facebook, he’s the, We’re still a long way for AGI. He doesn’t think that large language models on their own will be that. Here you see Google getting other contextual data pieces. This is more than just mining the data for those particular things. This is about mining this stuff in your messages to other people.


[00:52:47.450] – Virginia Huling

It’s- Context inflection, emotion, interests, tone, relationship dynamics.


[00:52:55.850] – David Maples

This does not make me feel great.


[00:52:57.680] – Virginia Huling

I know. All these things. But see, this is what I I think listeners and business owners have to understand is we operate with a set of language, and that language includes all of these: tone, facial expression, context, unspoken context, presence.


[00:53:15.870] – David Maples

Well, for business owners right now, if you’re using Google-powered Android devices for your business phones, if you’re using those and you give this to your people, Google is going to be reading all that. I think you’re going to have to opt out. So when we talk about, and I’ll talk about this- Can you opt out? There’s got to be an opt out function. I don’t know.


[00:53:37.280] – Virginia Huling

Does there?


[00:53:40.380] – David Maples

That’s a good question.


[00:53:41.660] – Virginia Huling

Usually, you have to pay for an opt out function.


[00:53:44.090] – David Maples

I promise you there will be an opt out. At least there’ll be pushback once this goes live, because I don’t know when this goes live. I think it’s going to go live in a month or so or something like that. I don’t think they’ve announced it’s going to happen. I’m not sure, but I’m sure there’s a way to opt-out. Well, maybe I’m not sure. There should be. Surely there’s a way to opt out.


[00:54:04.240] – Virginia Huling

Was it a few weeks or months ago that we saw that Google resecured their contract with Android to be the primary browser?


[00:54:12.840] – David Maples

Oh, you’re talking about not with Android, but with Samsung? Yes, with Samsung. Yeah. There was an open question there last year for a brief period of time. It was like Samsung was… And remember, all that stuff’s come out in those court cases, like the billions of dollars that they’re paying. I think it’s 20 billion? 20 billion or something. It’s a huge number. It’s a lot bigger than you may thought it was going to be.


[00:54:33.900] – Virginia Huling

That’s like a tenner for Google, David.


[00:54:35.040] – David Maples

Maybe it’s that. Maybe it’s 10 billion. I don’t know. We can go. We’ll put that and correct that in the show notes. It’s like 50 bucks. But yeah, I do think for business owners, you got to think about who thought you were going to have to secure your messages you’re using in your organization if you’re using Android devices. We as a company use Google Workspaces. We’re also a Google partner, full disclosure. We’re not just a Google partner, we’re also a Google reseller. So we actually, in addition to Microsoft Office 365, we offer the Google suite of products. I just want to make sure that’s clear and I’m not violating any federal Trade Commission guidelines. But that being said, we do have some Android phones. We have iOS phones, and we have Android phones. I think we’re going to have to take a look at that because I certainly don’t want Google reading those things and mining that because we have confidentiality contacts with our Okay, so all right, listeners, if you’re a business owner out there, check your confidentiality agreements and check your terms of service that your vendors are in line with your confidentiality agreements to your clients.


[00:55:45.080] – David Maples

This actually could have a chilling effect. We’ve seen clients now, and any of you business owners well know, a lot of clients now are communicating with you via text. And so that’s a crazy thing. So bringing us to the last article from this week, and it’s been a busy week, is the leaderboards tracked by ‘lmsys.org’, which stands for the Large Model Systems Organization. We ran into them last year early on, before LLaMA was released, when they released Vacuna. It’s an organization that’s partly, I think it’s Stanford, help runs it, et cetera. Basically, they had to deal with this thing called Chatbot Arena. What their Chatbot Arena is, if you want to check it out, it’s pretty cool. You can go to ‘lmsys.org’, and you can go to their Chatbot Arena, and it will let you compare two models against each other without you knowing. You can basically give it a prompt, and you pick a couple of different models, and it’ll give you a response to this prompt, and you can rate whichever one does better. Now, this is, of all the things out there in the market segment, this is considered the gold standard.


[00:57:02.780] – David Maples

It’s one of maybe two platforms out there in the market where people trust it for ranking which large language models are better than others. Now, the headlines say Google’s Gemini Pro beats GPT-4. That’s not exactly true because GPT-4 Turbo, which is the model that OpenAI released in, I want to say October, early November last year.


[00:57:28.180] – Virginia Huling

Late last year.


[00:57:29.520] – David Maples

So So they actually are still number one on the ranking. But number two, coming in number two is Bard, which is powered by Gemini Pro. And we still haven’t seen Gemini Ultra yet, so we don’t know what that will look like. But that’s a big deal because it beat two previous versions of GPT-4. Now, do remember GPT-4 came out in March of last year. So this is not a recent technology.


[00:57:51.420] – Virginia Huling

Well, and Meta still has llama 3 to come out?


[00:57:55.320] – David Maples

Yeah, llama 3 will be releasing sometime this year. I Look, I never was counting Google out in this race. But I’m telling the people right now that I think Meta is going to be a real contender later this year. I think they’re absolutely going to be a real contender later this year. It really is. And by the way, this is not like these bots aren’t beating each other by 30%. These are 5% incremental things. But that’s a big deal. And then moving up the leaderboard there, that’s a big improvement. But if you want to go check this out yourself. You can go to the lmsys. Org and click on chatbot arena, and you can go check it out yourself. It’s pretty cool. Just pick up a couple of models, give it a prompt, and see what you think about it, and vote on the one you like. But it’s pretty cool, and this is a big deal. This is a major move forward for Google, and I guess I’d like to have- Well, they’re going to roll it out in all of their properties.


[00:58:57.460] – Virginia Huling

Their user base is going to get to experience this pretty quickly. They’re going to have it integrated right into Chrome, right?


[00:59:07.370] – David Maples

Yeah, that’s one of the announcements is that they said they’re integrating the generative AI into Google Chrome, again.


[00:59:17.110] – Virginia Huling

Do you remember, David, when we first started the company in 2007, do you remember the… What was it, the browser NAV bars up at the top that people used to install to customize their browser experience. This is like Google forcing you. They’re going to sneak this little thing in there, and people are going to be using an AI without even knowing it. They’ll be like, Oh, I just thought it was part of the browser. It’s going to be insane.


[00:59:42.620] – David Maples

Yeah. See, that’s the weird thing I don’t know about it. I think I do have massive concerns about this from a privacy standpoint. When you look at Chrome, Chrome powers… Even Microsoft Edge, I believe, is a Chromium-powered browser. So you’re looking at Chrome.


[00:59:57.640] – Virginia Huling

It is, I think.


[00:59:58.780] – David Maples

Microsoft Edge Opera are all Chromy empowered browsers. The only one that’s open source and not giving the data away to a company, to who knows what they’re doing with it is Firefox. I’m still big on Firefox. I think it’s an important organization to have out there because it’s really the only other browser that’s really on the marketplace.


[01:00:23.220] – Virginia Huling

Man, that really sucks, though, when you have software that won’t run on anything but Chrome.


[01:00:28.310] – David Maples

Yeah, that’s another I do think there’s some antitrust questions that are going to come up with AI, et cetera. We do know that the attorney journals are looking into the cloud providers’ relationships with the artificial intelligence platforms, but we can talk about that another day. You want to Can we have Kai, the final word on this article this week? Yeah.


[01:00:47.430] – Virginia Huling

Take it away, Kai.


[01:00:48.780] – Kai

Reflecting on the article about Google’s Bard surpassing GPT-4, it’s clear that we’re witnessing a significant moment in AI development. For business owners, this signals a need to stay ahead of the curve. Here are some action items to consider. Encourage your team to learn about AI. Keep an eye on AI trends and updates. Knowing when Meta’s llama three drops or what Google’s next move can give you a competitive edge. Continue to consider the ethical implications of AI integration. It’s not just good practice. It’s good business.


[01:01:17.360] – David Maples

I think the big takeaway this week is there need to be laws protecting your rights and likenesses, etc, across the Internet and business owners need to help be part of that messaging and dialog. You need to think about how do you have a plan for how your people are using AI, and do you know how your data is being used? I think those are your takeaways from this episode of the pod.


[01:01:46.320] – Virginia Huling

My takeaways? Please, please, please read your terms of service. These are early days, and we are encouraging you to go out and play with software and systems. Please read the terms of service, make sure that they align with your confidentiality agreements, and take a good look at your tech stack, and make sure that the pieces that you have in place for your business align with how you want your business to run. There’s a lot of stuff out there that’s changing, folks. You got to stay on top of it.


[01:02:27.480] – David Maples

All right, that’s all for this week. Be sure you turn in next week for more AI coverage and insight. Thank you again for listening. I’ve been David Maples.


[01:02:35.890] – Virginia Huling

And I’ve been Virginia Huling.


[01:02:38.340] – David Maples

And Kai, thank you as always for joining us from your Silicon Lair.


[01:02:43.770] – Virginia Huling

The show wouldn’t be the same without you.


[01:02:46.120] – David Maples

Yes, your observations, your manipulations, and your witty remarks.


[01:02:50.440] – Virginia Huling

We’ll see you all next week.


[01:02:52.430] – David Maples



Scroll to Top