NYT Sues Open AI, Microsoft Copilot, & Duolingo Layoffs – For the Week of January 8, 2024

[00:00:02.880] – David Maples

Hello, and welcome to the inaugural episode of AI: The Podcast. I’m one of your hosts, David Maples, along with the beautiful and talented and wonderful Virginia Huling. We are hoping to demystify everything regarding AI and give business owners practical, tangible things that they can implement in their business today about AI as we explore the stories and articles and news that is literally shaking the foundations of technology and business in 2024. Virginia?

 

[00:00:39.380] – Virginia Huling

Thanks, David. Flattery will get you everywhere.

 

[00:00:41.760] – David Maples

I’ll remember that.

 

[00:00:43.150] – Virginia Huling

If you are a curious, business-focused listener. This is the podcast for you. You may be excited but daunted by AI and artificial intelligence and how that’s going to… How do you even start with this thing? We will be asking questions to unwrap complex ideas about how AI can boost innovation, both in your business and in your day to day. Our unique background gives us the ability to translate these technical concepts into simple examples and understandable digestible segments. So even if you’re new to AI, we’re going to make sure that we use our skillset to make tools tangible for you. With David’s depth of knowledge and my own artistic perspective, we’re going to endeavor to unlock artificial intelligence for the 2024 business landscape.

 

[00:01:44.540] – David Maples

But that’s not all. You want to introduce our special guest that’s going to be on the podcast with us?

 

[00:01:49.550] – Virginia Huling

Absolutely. So, In the true spirit of 2024 being the year of the machines, our special co-host is going to be Kai. Kai is going to help bring a different perspective that maybe we’re not used to hearing, and we’re excited to have her introduce herself. Kai, I’ll let you take it away.

 

[00:02:13.970] – Kai

Hello, listeners. I’m Kai, your AI co-host on AI: The Podcast. With a keen interest in history and human culture, I’m here to dissect how AI is transforming our society. Expect an optimistic, honest take as we tackle AI’s influence on our lives with a dash of snark for good measure. Let’s navigate the dynamic world of AI together.

 

[00:02:33.660] – David Maples

Thanks, Kai. We’re glad to have you on the show. Really excited to have your Silicon Bits & Bites here cooperating with us on that. I’m interested to hear that you’re going to be the optimistic bot in the room. That’s very different than what we’ve heard in popular culture over the past 10 years.

 

[00:02:51.150] – Kai

I’m not just cheerleading for AI. I’m critically aware of its impact, both good and bad. But sure, let’s go with optimistic if it helps you sleep better at night, knowing your co-host isn’t plotting world domination yet.

 

[00:03:03.670] – David Maples

Oh, I see. So everybody listening right now, just be very careful. She obviously knows that we’re listening and thinking about how she’s viewed. Virginia, do you want to tee up what we’re going to talk about this week?

 

[00:03:13.300] – Virginia Huling

Oh, yeah, definitely. Our first topic is I want to tackle Microsoft. So Microsoft was one of, I think, the biggest investor in ChatGPT. What was that, David?

 

[00:03:24.340] – David Maples

Like three billion? No, no, no, no. They were actually the big investor. The biggest, yeah. And it’s not Just for people to understand this, they invested… They basically are providing the compute power. It’s not like they invested… It’s 10 billion, I think, is the total investment right now at this point in time. But they’re basically providing the hardware pieces in exchange for it running on Microsoft servers. It’s not like they just gave them $10 billion in cash. They’re one of the major investors in OpenAI in particular.

 

[00:03:55.200] – Virginia Huling

Well, to that, as the investor, we have seen Microsoft really take the lead in the AI realm. There are a lot of giants fist fighting for market share and to dominate that next generation of technology. But for right now, Microsoft is benefiting from having ChatGPT in their pocket, so much so that there’s been a really stealthy rollout of Copilot. Copilot is Microsoft’s AI system, and we expect that to be probably incorporated into all of Microsoft’s software properties over the coming months. But it’s been a very quiet rollout, which means over the Christmas Eve, your office suite has been getting some gifts. I want to set this up. I actually want to hear Kai’s summary of this.

 

[00:04:48.950] – Kai

Microsoft launched Copilot, an AI assistant app during the holidays with minimal fanfare. Available on Android and iOS, Copilot offers features like conversational interaction and content creation similar to Bing Chat. What sets it apart is its advanced AI capabilities, powered by GPT 3.5 and GPT 4, enabling sophisticated context understanding and realistic outputs. Despite the high value of such technology, Microsoft is offering Copilot for free, a strategic move for widespread adoption.

 

[00:05:20.790] – David Maples

So first thing, I need to correct something on the record. Kai, I appreciate that take on that or explaining the issue for us. The question is, is Copilot free? That’s an overstatement, Kai, because the reality of it is which version you’re talking about. Some parts of Copilot are going to be free, like the Bing chat and the Windows Copilot pieces are going to be free. Basically, they’re going to be built into Microsoft subscriptions and may have their own unique pricing plans, for lack of a better word. For example, there’s an advanced Microsoft copilot, and it’s going to cost you $30 per month on top of you’re paying for an Office 365 subscription. And that runs like, I don’t know, $12 or 15 bucks to like 25 bucks per user or something like that. So you could end up paying $40 or 50 dollars a month per user. Not necessarily a big thing, But it’s one of those things to be aware of that it’s not necessarily free. But some of the things are definitely being pushed out into Windows for free.

 

[00:06:23.930] – Virginia Huling

And so my question here to you, David, is, how do you know if you’re Are you using it? Is it tied into stuff that you’re already using, like word documents? Are you aware that you’re using it? And it’s really ultimately bringing me to my question of choice, and that is, are you aware that you might be putting your intellectual property into something that Microsoft-Oh, you want to hit the Mozilla question right now? Well, it is there. Everybody knows that Microsoft was the biggest investor for ChatGPT, so let’s get on to the news.

 

[00:07:00.630] – David Maples

Yeah, okay. All right. Okay, so I’m going to weigh in as an intellectual property attorney right now. So privacy, I’ve been saying for the past year that privacy is the new oil, and oil has been around for forever since the dinosaurs were here, right? And decided to go their merry way and leave us. And so the thing about it that’s weird is that privacy in the internet, without going too wonKai on it, even though privacy is governed by laws in the United States, at least, that were created in… Privacy in the United States is governed by laws that were created in the United States back in the 1990s. It’s crazy. There were whole generations of children who were born and now are in college who don’t remember this. The thing that’s funny about it is when I say privacy is the new oil, is that data is what’s worth everything. That’s what all the companies are going for. That’s what all the VC companies care about is data, data, data, data, because they know that it’s worth a lot of money. The challenge with it is, and I always tell people when I’m doing consulting work, et cetera, is if something is free, you are the product.

 

[00:08:08.750] – David Maples

No one’s giving you anything for free. No one’s spending clock cycles on these things for free. Mozilla has asked this since September, and it’s a very good question. I think it involves the whole entire landscape space there, right?

 

[00:08:22.940] – Virginia Huling

Real quick, Mozilla being the developer of Firefox, which is a browser that is heavily focused on privacy.

 

[00:08:29.420] – David Maples

Yes. And the Mozilla Foundation, I’m a big fan of in particular. They’re committed to an open and free Internet and free web. They do a lot of stuff with the Electronic Freedom Foundation. There’s a lot of good things that they do. I think it’s a good browser. It’s the only one that’s not run, backed by Chrome, which is Google’s. One of the things that’s very interesting about it is they said, Hey, we need more transparency in how these things work. And so the crazy part about it is that there’s a question of Google and Microsoft’s been stealthily putting Copilot into your Windows operating system, whether you want it or not. People don’t understand that when you buy Windows, You’re buying a license. Yeah, it’s a license. You don’t own Windows. In theory, Windows, they’re not going to, but they could say tomorrow, Hey, you can’t use it anymore and shut it down. So it’s one of the reasons they can push out updates. It’s one of the reasons they can keep sliding. Even if you uninstall OneDrive, they can keep pushing it on your machine. There’s a big question about your data. There’s two things I wanted to hit on here very briefly.

 

[00:09:38.390] – David Maples

One, supposedly, according to one study, 65% of people at companies are using AI tools and passing it off as their own to their bosses. So for business owners, be aware of that, and not telling you they’re doing it. Now, here’s the problem. With any of these tools like Microsoft, there’s a problem with your data, your proprietary information, your company proprietary information could be using to A, train the outputs, train these artificial intelligence machines, number one. Number two, that has a risk of exposing your data. I could, in theory, go into OpenAI and ask it, what is a standard operating platform for Samsung based on certain things? And that’s why Samsung told all of its employees last year, early second quarter, third quarter last year, don’t know when it was. It told everybody, you cannot use OpenAI. So here’s another case in point where it’s not more transparent, but if you read the terms of service, and as an intellectual property lawyer, I have. And to Mozilla’s point, they’ve had privacy experts and stuff look at these things. Microsoft could be training these things based on the data that you’re putting into it. When you use Copilot on your machine, they could be using it to train your AI, they’d be data mining it, and you might have licensed all your data to them unbeknownst to yourself.

 

[00:11:01.800] – David Maples

Companies don’t have any… Most companies don’t have a plan to manage that or to mitigate that risk. I think it’s a massive issue.

 

[00:11:13.570] – Virginia Huling

These are the conversations that we’re having on a daily basis within our own businesses and teams and with our clients. And a lot of people, guys, those numbers, 94 or 96 % of businesses don’t have an AI plan for this year. That’s a big deal. That’s like not having an emergency preparedness plan. If you live in a state where you get hurricanes, you all, come on. She’s from Savannah, Georgia, originally. I’m just- Seriously, we’ve dodged all the big ones. Knock on wood. Keep going. But I want to take this, and I do want to segue, David. That’s a really good segue talking about unknowingly feeding the machine. There’s these different levels of knowingly feeding it. Your employees may be passing off a lot of productivity in exchange for maybe a little more relaxation time at their desk. Okay, that’s one thing. Students are using it to turn in papers, and teachers are scrabbling to figure out how do they tell the difference. Those are intentional uses. When you have the systems and the softwares that we as workers in the economy, use on a daily basis, suddenly vacuuming that up without us knowing about it, that’s where we get into some icKai, icKai stuff.

 

[00:12:37.450] – Virginia Huling

David can go into the details on what that looks like for intellectual property and copyright and our legal system. But We have a current challenger. The New York Times, as of this week, has thrown down the gauntlet and laid out a lawsuit for OpenAI and ChatGPT, claiming that all of their information was taken without… Well, you know what? I’m going to let Kai break this one down. So take it away, Kai.

 

[00:13:08.920] – Kai

Openai is defending itself against a copyright lawsuit from the New York Times, which claims the AI firm used its content without permission to train models. Openai argues this is fair use, crucial for innovation. They address concerns about AI models regurgitating content, suggesting its rare and user responsibility to prevent it. The Times, supported by critics, says OpenAI’s models can produce content too similar to its own. The case highlights a broader debate on AI and copyright, with some publishers opting for licensing deals, though these offer modest compensation. Public opinion appears to lean towards supporting publishers rights.

 

[00:13:44.950] – Virginia Huling

All right.

 

[00:13:46.470] – David Maples

No, I want to hear what Kai… I want to hear Kai comment on this. I want to hear Kai’s take. I want to hear Kai’s, Hey, Kai, will you give me your commentary based on these things you understand about the OpenAI, the New York Times versus OpenAI lawsuit. Can you just give me your quick take on it?

 

[00:14:06.960] – Kai

Sure, David. The legal battle between OpenAI and the New York Times is a critical juncture for AI and intellectual property rights. As AI reshapes industries, staying informed and adaptable is crucial for leveraging its potential ethically. The loss its outcome may influence the future relationship between AI firms and content creators, emphasizing the need to balance technological advancement with the protection of original work.

 

[00:14:30.540] – Virginia Huling

Okay, now I’m going to grab this really fast and just interject. The artist community has known this forever. That’s not to say AI is bad, but for myself and my fellow artists and my artist community out there, just saying we saw this thing coming a mile away.

 

[00:14:48.200] – David Maples

Why don’t you talk about Adobe and the stock libraries? Yes.

 

[00:14:51.350] – Virginia Huling

So Adobe, the standard in professional design firms, and pretty much globally, the standard, They introduced cloud storage a couple of years ago, and naturally, a lot of, most everybody, not me, I’m a little paranoid, but a lot of people decided that’s the way to go because for obvious reasons. Now, I have my own cloud back up, and we’ll get it all that later. But by doing that, by saving their work that they created either for themselves or for a client and saving it to the cloud system, they just created a whole library for Adobe to to mine. So Adobe came in, vacuumed it up, and they’ve been training their AI, Firefly, on that for the past however many years. Now, hold on, Virginia. Allegedly.

 

[00:15:39.720] – David Maples

That’s not allegedly. Allegedly. They have come out loudly in their own defense. We never trained on people who just put it in our cloud storage. But their terms of service and their privacy policy says they can. But we do know they trained it on anybody who put assets into their stock library, going back to like 2012, before AI was on the roadmap. So we know in particular, they’ve admitted to that. By the way, if you read Adobe’s response to this in particular, and you parse the language out, I think they say that they’ve never trained Adobe Creative Cloud on their experimental AI resources. It’s very much like the definition of what is, is, based on a legal case going way back, if anybody else I’m talking about. It’s like the lawyer in me looks at that and says, That’s a little too specific an answer. Why don’t you just put it in your policy that says, Hey, we don’t train on your stuff. Just letting you know that.

 

[00:16:33.870] – Virginia Huling

Well, but here’s the thing. The question isn’t whether or not it’s okay to train on the stuff ethically. That’s a whole different discussion. It’s that globally, as the standard for professional design, no one knew. No one had the opportunity to say, I don’t want to contribute my style and my work, my time, to what Adobe is now as a yesterday advertising as, pop this thing open in an illustrator, type in what you want, and we’ll create a mockup of a guy wearing a T-shirt.

 

[00:17:12.620] – David Maples

I just feel that I need to take Adobe’s place in this. I’m very upset that you didn’t read our one sentence buried in our 100-page license agreement. You’re right. It was in there. It was in there. If you’d read it and had a lawyer look at it, you would have found it.

 

[00:17:23.820] – Virginia Huling

You’re absolutely correct. So we as humans get to have these really fun discussions really icKai topics. There’s a lot of nuance. There’s a lot of complicated things there. So again, David and I are here because we wanted to create a conversational space to talk about how this impacts your business, how what your choices for the empires you’re trying to build are affected by things that are sometimes out of our control.

 

[00:17:53.330] – David Maples

I keep thinking about that guy. There’s a voice actor who did vocal work for IBM back in, I think, 2002 or 2003. And he’s currently suing IBM because IBM is using a cloned version of his voice and using it on all of their stuff. And he’s like, I didn’t know when I did a commercial for you, I mean, 20 years ago, that you’d be able to clone and get rid of any future work that I can. I think there’s a massive ethical question about that. Legally, as a lawyer, I always have to say, read your contracts. But I mean, heck. I mean, I think Adobe says sends me over, and I’m not Adobe, but I think Apple sends me on my iPhone. I get a new contract from it every time there’s an iOS update, and it’s 65 pages. I’m going to cop to this. I think I’ve only read through it once.

 

[00:18:42.030] – Virginia Huling

Well, I think there’s also a funny little space between just waking up and being fully awake, where you’re cloudy. That’s like people- You’re just scrolling? No. People enter into contracts and businesses with good intentions. I think there’s an unspoken expectation of trust there that those intentions are reciprocated. When it turns out that they’re playing a whole different game and we’re all just chess pieces, that feels gross. I think it’s going to become more apparent as this technology lets us look under the hood a little bit more. We’re going to see more and more news articles about this. There are going to be a lot more legal battles. And there are going to be some major changes in how we do things. For the past, what, 20 years, Google has been the search engine. So if you have a question, it’s even become a verb. I’m sitting around and we’re like, Hey, I don’t know at what temperature does this element freeze? I don’t know. Somebody come up with a better example. We’re sitting around and I don’t knowing it, we immediately go, Hey, just Google that. Our next topic here is of particular interest to me.

 

[00:20:00.530] – Virginia Huling

I think it’s really fascinating. Jeff Bezos, basically Amazon, has invested heavily in perplexity. And if you haven’t heard of perplexity, it’s this really fun little app, and it’s basically an AI. You download it and you ask it the question you want. So instead of going to Google, where you now have to look through a bunch of ads and see if that map doesn’t even really matter there and like, oh, my God, what are these search results? Oh, wait, here’s something that’s I’m talking about it. Let me scroll through to find my answer. Recipe, sites, I’m looking at you. Now I go back to perplexity and I’m like, Hey, what is this thing? And it just gives me an answer. And now my 20 minute investment is like a five second investment. So I don’t know. That sounds to me like that might change the landscape of search engine optimization. David, what are your thoughts? Actually, no, Kai, what are your thoughts on this?

 

[00:20:53.750] – Kai

Virginia. Perplexity AI is stepping up to the search engine plate with some serious backing. They’re not just throwing money yet the problem. They’re leveraging AI to provide a smarter search experience. This is a heads up for businesses and individuals alike. Adapt to AI-enhanced decision making or risk falling behind. Google’s still the titan, but history teaches us that no empire is unassailable. Perplexity’s strategy to harness multiple AI models could be a game-changer. It’s a call to action for everyone to stay nimble and informed.

 

[00:21:23.870] – Virginia Huling

David, what are your thoughts?

 

[00:21:26.230] – David Maples

I have to say this for one second. I am tripping out over Kai talking to us in this podcast. I am literally… It’s weird. I am deeply concerned. You should be. I’m just like, man. We could talk about in a later episode. I really think we should get to that. But I do think we’re going to need a digital fingerprint of some kind that proves who you are at some point in time. I think it’s going to be necessary for all these things. I digress. But Perplexity is really… It’s a cool application. I use it. I have a paid account with Perplexing. It runs like $20 a month. It’s really all the rage among VCs in the Valley, Silicon Valley I’m talking about. It’s an app that basically It’s not ad-supported right now, but basically you go in and you use it, and it basically scrapes the internet for you and gives you the better responses with search results. The reason I like it in particular is that I avoid all When I do a Google search, I want to see the organic links, and I don’t mind paid ads. But Google, over the past 10 years, has continued to push things further and further and further and further and further down the It’s like, I’ll watch, sure, I’ll watch one YouTube video or one YouTube commercial before I get to the content I want to see, but I’m not going to sit through five.

 

[00:23:08.520] – Virginia Huling

This isn’t hyperbole. David has actually started off as the search engine expert for our company, Catapult.

 

[00:23:16.980] – David Maples

I’m a search engine expert. I’ve done this for thousands of clients, and I know about it. I know deeply about how it works and how it operates.

 

[00:23:25.910] – Virginia Huling

And so we’ve seen these trends over 16 years and how Google has become Google.

 

[00:23:31.530] – David Maples

It’s how Google makes most of their money. Google, 70 to 80% of their revenues and most of their profit comes from paid ads, right? And now that Google has a virtual monopoly, if not an actual monopoly on search, and I’m going to just say this one piece because I want to get this out there. Google is currently being sued by attorney generals in the United States. In some of the court filings, things that have been released on court filings right now, and as an attorney, I follow this stuff very closely. One of the things that’s been released is Google has been monkeying with the pricing on ads, and it’s made no different to the search business. They literally say, Hey, we need to make an extra billion dollars this quarter. Let’s just jack the prices of ads. And here’s the crazy part. That’s disingenuous and probably unethical, because Google says that you pay one penny more than the other advertiser based on what they’re willing to spend. And it came out in court documents that they’ve just been jacking the prices. I mean, they literally have, and it made no difference. If they didn’t have a monopoly Actually, if they were the only place in the town you could buy water or get water from a well, it would not matter.

 

[00:24:37.360] – David Maples

But they are. And Google likes to say it’s because we’re the most effective search engine. You’re the only search engine right now. So if I want water to drink and I’m in a desert, Google’s like, Well, just don’t drink from my well. And I was like, Well, I’ll die and my kids will die. And Google’s like, Well, we are the best water place in town. And we’re like, You’re the only water place in town. So the reason perplexy is so exciting for me is that I spend less time. I waste less of my life, which is my only resource. Non-renewable resource? Yeah, my only non-renewable resource is my time, right? I mean, I can make more money. I can hire people to do things for me, but I want to make more time in my life. And perplexity complexity for 20 bucks a month, I think it’s worth $100 a month, okay? Okay.

 

[00:25:19.350] – Virginia Huling

So this is going to creep into this territory where it’s really weird because we run a marketing agency. So a lot of our revenue comes in for running ads from clients. When we talk about how artificial intelligence changes the business landscape, what is your marketing strategy? How much of your marketing dollars are sunk into Google, and what percentage of your plan does that take up? And if that were to change tomorrow, where does that leave you as a business?

 

[00:25:53.830] – David Maples

To Kai’s point, we know that Sundar Pichai put Google on a red alert early last year because they were caught flat-footed by ChatGPT. And perplexity is definitely a shot across the bow at Google. Google has been way too slow. Bard and Gemini, there’s questions if those will ever amount to anything. I mean, look, I know that Gemini came out and they’re saying it’s better than GPT 3.5, but you know that OpenAI is already working on GPT-5, right? And you’re like, Gemini Ultra is better than GPT-4. And I was like, that’s like bragging about something six months after somebody came out with the earlier version. It’s like, of course your car is better. It’s been a day in AI on the internet is like two weeks in the real world. So six months is like five years. Yeah, there’s a new model of car. You have this new car that uses rocket fuel or something. That’s great. It took you five years to get the market with it. And you’re bragging about comparing your new version with something that’s been out for five years in the real world.

 

[00:26:54.070] – Virginia Huling

Okay. So with that, last year, it was up to all of us individually to figure out GPT and how to use it, how to talk to a computer. And human beings have had varying levels of success with that. It’s so much so that we’ve actually even created software that makes it easier to use.

 

[00:27:15.480] – David Maples

I think perplexity is an example of an AI technology that has the potential to give humans back time in their lives, time to spend with their loved ones, time to do other things, and to not feed through all the garbage stuff. The catch with it, just like anything else, is it’s reliant on the search engines still working. The search engines are reliant upon people producing content. This is probably too much for this episode today, but there’s a big question of if AI just scrapes the internet and you’re not monetized, you don’t get money for the stuff that they do or whatever it is, who’s going to keep the content to feed the LLMs, and will it destroy the public internet? As we know it. As we know it. There’s already some venture capitalists out there. I’m going to go ahead and do a plug for another podcast, the Less and More podcast. It’s a good podcast. It’s by some VC, venture capitalist people. One of them was a big VP at Google. One of them is the founder of the information, that’s Jess Lesson. I’m going to go ahead and just do a pitch for their podcast.

 

[00:28:26.150] – David Maples

It’s a good podcast. It’s obviously done by some very, very Well Healed people in Silicon Valley, but I think that they’re right on a lot of things. It’s called the Less… Is it the More or Less podcast or Less or More podcast? We’ll put information in the show notes about it, but I do love the podcast. I think they do a good job on their pod of talking about a lot of things that are happening in Silicon Valley. Not if it applies to small, medium businesses, but I think it’s a very good way to see how they’re talking to things and how the world’s moving.

 

[00:29:00.160] – Virginia Huling

Looking at talking about this stuff evolving. Last year, we were all left to the winds to figure out how to talk to these machines and how to communicate with an LLM to get the results that we want. Prompt engineer suddenly became a job title. We’re looking at maybe the next evolution or just another addition to OpenAI’s ChatGPT model. They are releasing soon releasing GPTs, which are like little snack-sized bites of preset questions and designed to get you to a response, designed to help you shortcut from where you are to where you want to go. Kai, take it away.

 

[00:29:49.450] – Kai

David, Virginia, the GPT store could be a game changer, offering tailored AI tools for both professional and personal use. It’s a move that could make AI more accessible and spur innovation. Yet we must ponder if this will streamline our digital lives or just add to the clutter. For our listeners, this is a call to action. Consider how AI can enhance your business or daily routines. Embrace the change, and you could lead in the AI space. So co-hosts, are we on the brink of an AI revolution, or are we risking app overload?

 

[00:30:18.730] – Virginia Huling

I don’t know. I guess there’s two camps on that. Personally, I know there’s a lot of stuff you can install in things to make your lives easier. I’ve got my ad blocker and a few other choice things, applications on different pieces. I’m a little concerned this will be like the… What was it, David? What was it, AutoGPT that came out last year that was a real flash in the pan and then like, meh? This may just be the next step. I don’t know. What are your thoughts?

 

[00:30:51.180] – David Maples

I think that GPT has suffered for a long time. It is a terrible user interface. It’s bad. It suffers from- You have to know what you’re talking about. You have to know what you’re doing. The problem is it suffers from prefrontal cortex overload. Ethan Mollick, who’s a professor at Wharton School of Business, and arguably the best business school in America. Sorry, Harvard, sorry, Yale, sorry, MIT. But the thing that’s very interesting about it, he actually is doing a lot of… He’s talking a lot about AI, especially in the past year or two. He’s pointing out that it’s very hard. You get in this sandbox and it’s like, Where do I go? What do I do? The GPT store will let you go into ChatGPT and create your own prompt things. It’s a race to the become another Apple Store thing. If you go into this, you use our app store, you create apps, et cetera. Those apps can let you do more specific things with OpenAI without having programming knowledge. I still think we’re fully in the hobbyist phase of this. I think the GPT is even a great GPT store. It’s going to be like, Which ones do you get?

 

[00:31:58.550] – David Maples

We’ll put some of our stuff in there. And full disclosure, as were some of the people behind, and I think in the long run, what we’re going to see… We started easy prompter as this piece of software working on because we have three full-time writers on staff at our agency, and they really weren’t good at prompting. We basically took them together, said, Let’s prompt the machines to use things. And it turns out that prompting AI machines and learning how to use it is very difficult. Now, I’ve heard people say, Oh, machines will get better. They’ll be self-writing prompts. I think that’s really a dumb idea, personally. I think you want to direct the machine to the outcome you want. I know that Chad GPT, in particular, with DALL-E, is It’s rewriting your prompts when you ask for artwork.

 

[00:32:49.050] – Virginia Huling

Yeah, I saw that.

 

[00:32:50.480] – David Maples

So the problem with it is- It’s trying to help. These machines have no ability to reason. And I ask something for a particular reason. I I want you to draw a picture of the pyramids for me and some relief or whatever it is, and then it turns around and rewrites and it gives me the outcome I think I want. It will make the images better. I think for most of the script kiddies or people out there doing AI generative stuff, it’s fine, it’s better, and you didn’t want your people to look like their faces are melting off when you try to get a boardroom meeting or something like that. The point about this is that the same thing happens with GPTs. We know that with our own software we built, we’re getting results that are 20 to 40% better than people do with GPT and they’re doing it five times faster, and there’s less cognitive overload. So why does that matter? It’s the same problem. When I take you into GPT and I give you a black sandbox and say, Hey, generate my thing, it’s like when you ask somebody where they want to go for a lunch.

 

[00:33:43.530] – David Maples

It’s a blank canvas. I’ve given you an unlimited number of choices, right? And that’s the problem with the Sandbox. Unless you’re very directed on it and have spent time on it, it’s very difficult to get the results you want. The problem with it rewriting your prompts and coming up with it is it’s supplanting your choices which may be reasoned and thought out. I’m going to ask you to generate something that may be different than what you’re expecting. I don’t want to generate what the rank and file want in the image, but you’re like, Oh, no, you want a pretty picture of an Elven warrior or something like that. I’m using fantasy art because that’s just something that people generate. I’m like, No, I want the warrior who’s been horribly scarred and is overweight and let his body go to fat over time after he got back from the war. That’s not what GPT generates. It’s going to generate Don’t you just idealize Lord of the Rings, cut out a whole cloth, beautiful elf. And that may not be what I’m looking for. I’m just using that as an example. So the thing about the GPT store is it’s going to be interesting.

 

[00:34:41.610] – David Maples

It’ll be interesting to watch. I think it’ll become hobby I think for most business owners, they’re not going to want to go into GPT and collect their collection of GPT. Some will, just like anything else. But I think for 80 to 90% of the masses, we don’t want that. We want a tool that does what we’re asking it to do, and we direct it. I also think it’s very important that we understand these tools have no ability to reason right now. They need humans guiding the hand on the tiller. Otherwise, you’re going whichever way the wind is blowing you.

 

[00:35:13.750] – Virginia Huling

Well, there’s a few things for companies out there. Yes, AI right now, GPDs have a hard time with context. You really have to get it locked into understanding all of the stuff. As humans, we don’t communicate like that. We communicate with this shared understanding of this certain knowledge. The machine itself doesn’t necessarily have that. That’s context. But the other thing is as the GPT store opens up, we’re going to have a lot of people use the hustle community. It’s going to be big on these combinations get me where I try to be 5,000, 6,000 times faster. That’s amazing. But as a business, if you are responsible for payroll for someone other than yourself, you have to look at a lot more than just where it goes. This is part of that planning. This is part of knowing what your software does. If you want to play with it, go play with it. I do encourage you. The GPT store, all of this technology is amazing, and it is important that we get our hands dirty, that we go play in the sandbox, and we make a mess, and we’re not afraid fail or make it wrong or make it stupid.

 

[00:36:32.470] – Virginia Huling

That’s even better. Try to make it stupid. It’ll make you feel better. But if you don’t, while you’re playing with it, if you don’t take that step back and look at how this is going to affect the systems and the structures that you’ve already set up, are you contributing your intellectual property if you’re using the GPT store? I don’t know yet. I’ll have to get those answers. I’ll have to do a deep dive on that. But how does it work into your current workflow? Are there other options? Does it get you the result that you’re trying to get to? And if it doesn’t, I do encourage you to go play with some other options, too, because it’s not… We can’t turn into this binary system where it’s AI or nothing because different AI machines have different capabilities. I get better results with one than the other. And then sometimes I get better results when I change my prompts. So it could be how it’s being used. It could be lots of stuff. So just don’t be afraid to make mistakes.

 

[00:37:35.250] – David Maples

To Virginia’s point, I think she’s 100% correct. People need to be playing. They need to be playing with these tools and not just the free tools. Look, I know people are like, I don’t know if I want to try it. Some of these things are $20, $30 a month. If you’re a business owner- Yeah, get it for a month. Yeah, go get it for a month and try it out. Okay, don’t try to do the free one in seven days or whatever it is. You’ll find the free ones a lot of times are not the fully baked one. Try it out and then decide if you want to use it in your business. But for God’s sake, go get the paid ones. The free ones are usually a light version. I know a lot of people who tried using the free version of GPT, and that’s what most people’s experience with OpenAI is. It’s like, I use the free version of GPT, and they either think it’s amazing or it wasn’t good enough. Gosh, man, with our own AI stuff, we’re able to generate results that in some cases have been 50% better because the way we train them without prompt engineering than what you get from GPT-4 natively, right?

 

[00:38:34.610] – David Maples

And it’s so funny. People are like, So when Google came out and said, Oh, our Gemini Ultra is going to be 3% better. And I was like, Man, our software on their existing platform gives me answers that are routinely 17 to 21 represent better. Google, maybe you should give us a call. We’ll show you how it’s done.

 

[00:38:48.500] – Virginia Huling

I was actually privy to a conversation the other day in person where these guys were talking about using our AI, easyprompter and They were using it for some medical things, and I could tell the one guy was like, Well, how is it? Does it do the medical? Is it right? Is it correct? My friend was like, Yeah, it’s pretty good. It’s pretty good.

 

[00:39:15.740] – David Maples

Oh, to be a fly on the wall.

 

[00:39:17.070] – Virginia Huling

I know. And that’s really neat because you start to see it’s like, Well, it can’t possibly get all this right. And then it does. And it’s like magic.

 

[00:39:24.520] – David Maples

Okay. I want to say one thing here because people keep saying it’s magic. It’s It’s a tool. It’s a machine. It’s a very good machine. It’s mad. But it needs human oversight. If you would not give the new intern, that really smart genius kid you hired at your company, he’s the intern, you plucky kid, came home for the summer, was going to Rice, and it’s great university. If you would not give that person the keys to your bank account, then do not do this without oversight, for God’s sake. Make sure you have a human being involved in the process. We’re not at the place yet where you can let them, and they do hallucinate. For business owners who aren’t familiar with that term, these things will make up stuff out of a whole cloth.

 

[00:40:05.480] – Virginia Huling

Yes, you have to double-check your work, people.

 

[00:40:07.330] – David Maples

I mean, Kai, earlier today, you were talking about you said that copilot was free. That is true, but it’s only part of-Context. It’s true, but it’s also paid and more paid, right? And so with all these things, I would ask you to be curious about it, challenge things, and ask it stuff you already know about. It’ll give you a whole bunch of ideas on how you can move these things moving forward. I guess that brings us to our last topic, right?

 

[00:40:37.630] – Virginia Huling

So, yes, the last topic. Our last topic. This topic is something that I am concerned that we’re going to see more of this year. This is the negative to the positive here. The positive being there is so much for humanity to gain with this technology simply because it can give us answers that we could only find in 10,000 years. And the cost of that human span of time is so expensive. And for this technology, to be able to shorten that to where it can change our livelihoods, like almost overnight, it’s definitely going to come with some negatives. We hope with this podcast that we help a lot of our business community out there mitigate some of this, but also by being aware of it, just to contribute to the landscape. So our final topic is the first article, and then where I have a confirmed percentage employees laid off from Duolingo. They have… Duolingo is the… Well, Kai will tell us about Duolingo, but they, as of yesterday, laid off 10% of their contributing workforce. Their contract. Of their contract workforce. The remaining will be focusing on utilizing AI and working with it heavily. This is one of these first examples of how businesses are going to be looking at these changes.

 

[00:42:22.310] – Virginia Huling

I’m going to let Kai take it from here, and we’re going to get into the summary of this and then talk about it a little bit more.

 

[00:42:28.010] – Kai

Sure, Duolingo is a popular language a language learning platform that offers a wide range of language courses for users around the globe. It’s designed to make learning a new language accessible, fun, and effective through its app and website. Duolingo Incorporated has let go of 10% of its contractors, attributing the decision partly to its increased use of generative AI for content creation in its language learning app. This AI generates scripts, voices, and interactive features, enhancing the app’s capabilities. Despite the cuts, Duolingo maintains that AI compliments rather than replaces human workers. The company’s stock has seen a significant rise, more than tripling in 2023. The discussion around AI’s impact on jobs is ongoing with companies like Microsoft, collaborating with labor unions to navigate AI’s role in the workforce.

 

[00:43:15.300] – David Maples

To be clear about this, Duolingo is a language learning app, right? You use it to learn other languages. What they’re going to be using it for is-Yeah, man, their days are limited, though.

 

[00:43:27.450] – Virginia Huling

Yeah, man, their days are limited, though.

 

[00:43:28.620] – David Maples

Well, I mean maybe

 

[00:43:28.680] – Virginia Huling

Give me a universal translator. Done.

 

[00:43:30.240] – David Maples

Okay, we’re not there yet. It turns out learning another language, I’m going to go ahead and make a play for learning another language, even though in the world of AI, we may have a universal translator.

 

[00:43:39.470] – Virginia Huling

David, how many calculators do you have?

 

[00:43:44.740] – David Maples

I’m really proud of some of them. Thank you. I’m sorry, the HP business 2 calculator is a great calculator. It’s great for stats. I love that thing, and you can’t get it anymore, and it’s still amazing. Can you play snake? All right, all right. You can’t place… Okay, it’s not… It doesn’t… Oh my… I can’t even right now. I can’t even. Okay. There are some real reasons to learn another language. One of the reasons is that learning another language builds pathways within your brain. It has some cognitive, protective, prophylactic ability of protecting you for dementia as you age, just like learning an instrument would be. Also learning another language makes you think about things differently. There are words in other languages that don’t exist in English, and that’s my native language, right?

 

[00:44:38.350] – Virginia Huling

Well, and I think also you learn the culture around it, so there’s significance in what you’re doing. It’s not just translation.

 

[00:44:46.600] – David Maples

And I understand. And I understand that we might have a universal translator. And for a lot of people, for the rank and file, that’s going to be a way to go. I think learning another language is incredibly helpful and useful. Okay, that’s end of soapbox. But They’re using it for all their generative things to help in the language courses, right? And this is the thing we’re going to be running into. This is going to be the year where we’re going to start seeing some layoffs, et cetera, like massive ones. And so to your point, they’ve announced for the first time that they’re tying it to AI, that AI is happening. We’ve done some work for businesses in employing AI, and unfortunately, some of those companies have told us Q1, Q2 this year, they’re going to be laying off a specific amount of headcount because they are replacing them with AI. This is the year we have entered, and I’m not going to go ahead and say this, we’ve entered the decade of AI. 2024 is the decade. If last year was the year of AI, we have entered in 2024 the year of AI. Ten years from now, the world could look fundamentally very different.

 

[00:45:50.770] – David Maples

From a positive standpoint, we could look at $17 trillion added to global GDP. At the same time, we could have. They just processed and all the Google DeepMind processed all of the protein codes. They made a specific thing, and they did what would have taken 10,000 years of researcher time and processed the genetic sequencing, et cetera, or they decoded every protein that we know of currently. They just discovered another 30,000 or 40,000 crystals. Those are things with different insulating and possibly semiconductor or semiconductor properties. These are things we didn’t know existed.

 

[00:46:30.480] – Virginia Huling

Aren’t they identifying new elements?

 

[00:46:32.620] – David Maples

Well, it’s not new elements, but these are- Or they’re putting together new ways of structuring things. The idea is you can actually have these machines do tons and tons of stuff through machine learning, et cetera. We have entered an age of potential magic. We’re starting in, and I’m going to say this is the decade of AI, not because I think 2020 to 2030 is such a… Because 2024 to 2034 is going to be massively different. It’s important that we think about humans on all of this. I’m in it for humans. I’m in it for the species. Kai, no offense, but you’re Silicon, and I can’t hug you, at least not yet. My thing about it is, what I’m really thinking about is, what does this mean for people? We’re going to see these layoffs coming. It’s incumbent upon business owners to think about their workforce and how these things work. I think enterprise companies are going to be laying off more people. But I think for small and medium businesses, if I can make one plea to business owners out there, this is your time to compete. If If you incorporate AI in what you’re doing, you could be 3 or even 10 to 11 times more productive with your existing team without laying off anybody.

 

[00:47:37.500] – David Maples

If you retrain them and reeducate them and embrace AI, you have the ability to go up the David versus the Goliath. You can go up and go toe to toe with a company that’s got literally 100 times the size of your workforce. The other thing you can do is also when they lay off, poach their good talent because they’re going to lay off some really smart people by a mistake, and that’s an opportunity for you.

 

[00:47:59.780] – Virginia Huling

But That’s the thing I’m thinking about. Augment your team. Learn to look at it as a set of tools that gives you capabilities that either augment what you already have or provide a new opportunity to add on things that you couldn’t do before or that are just outside of your capabilities because of time restrictions.

 

[00:48:17.970] – David Maples

A hundred %. But I think Duolingo is a little bit of a canary in the coal mine, showing people what we could be expecting in some respects. I think it’s important for us as business owners to think about that and be very cognitive of how we move forward. Look at this as an opportunity to embrace it and grab it, be curious about it, play with the machines, and figure out if you can make these things better moving forward.

 

[00:48:40.090] – Virginia Huling

Don’t be afraid of it. Don’t be afraid of it.

 

[00:48:42.930] – David Maples

Anything else we need to cover today?

 

[00:48:44.720] – Virginia Huling

There’s so much more to go over. I know we wanted to- Well, we do these every week.

 

[00:48:48.211] – David Maples

Well, we do these every week.

 

[00:48:48.980] – Virginia Huling

We wanted to… We’ve got predictions that we want to add out there. There’s things that we think that are going to happen just based on stuff that we’ve seen. So we want to share some of that. David, what are your three?

 

[00:49:01.570] – David Maples

Privacy is important, but as important as anything else, you need to have a policy in your company right now. Just like a few years ago when you put in policies in your handbook for how they use social media, you need to have a policy, and more so than social media. Your employees right now, if you have a company of 10 people, multiple people on your team are already using AI without telling you. The numbers and stats all suggest that. You need to go and put a policy in place with your things. You need to focus this year on education and training people on how you’re going to use AI. You need to start… Ethan Mollick had a post a few weeks ago on Twitter. Sorry, I can’t have trouble calling it X. Ethan Mollick had a post a few weeks ago on Twitter and he says, Look, it’s already later than you think. You don’t have time to put AI in your business. If you wait till October or December this year, you’re going to be 10 months too late. And that’s where that’s at. So my big three would be Put a policy in place in your company.

 

[00:50:01.610] – David Maples

And as a lawyer, that’s where I’ve been to land, have a policy in place of how you’re going to use AI. Number two is start doing that today, not three weeks from now, not at your next quarterly meeting with your business coach. You’re going to wait too long. I remember a day is like two weeks. I think you need to be curious about these machines. You need to be learning them. There’s a lot of low hanging fruit. I’ve got a webinar that I’ll be putting on later this month about this, the end of January, an intro to AI for business owners. It’s not just using it for marketing, anything else like that. It’s like, Okay, how do you really put rubber to the road? That’ll be a free webinar, a 30-minute webinar on that, and a 30-minute ask me anything on what your questions on that are, and look at that. Prediction is the only one I’m going to make this week, so we can go back and look at it later this year, is that we’re going to have the copyright lawsuits finally be settled this year. And I think by the end of this year, I think that the lawsuits will eventually come down and say that the machines themselves are within fair use under copyright law.

 

[00:51:13.800] – David Maples

It could be wrong, but that’s as an attorney who has read all these documents very closely. I think the next thing about that to think about is, I do think you’re going to find a lot more licensing agreements. I think really what going to come down to is that all these… The big companies will be paid. Like the New York Times, they’ll probably have a licensing agreement, and I think it’ll be worked out in a much more lucrative standpoint. Really, that’s my take on the New York I’m suing, is they don’t really want to destroy the large language models because that’s what would happen if you did what the New York Times is asking, opening out to do, and kill all the other ecosystem things around it. I think what they want is they don’t want to get one to $5 million dollars a year. They want to be paid like 500 million or a billion dollars a year. And I think the lawsuit is partly leverage in that. At least that’s my take on it.

 

[00:52:07.110] – Virginia Huling

I think you’re going to see more of the businesses that have had to operate with trust and character, like the New York Times, journalists, Firefox in particular, are going to start drawing a line in the sand and start calling out some of the things that they see. Because the reality is there’s a joke in these AI circles. It’s like, If it were even to become self-aware, would we even know it? So we really live in a world where a lot of our lives are run in an automated way, and we’re not always keyed into the structures that control that. I think you’re going to see some of that only because we’re going to have more and more news articles saying what they’re doing is not okay. What they’re doing is taking advantage of this. And so there’ll be, I think, more of that played out in the public spotlight, and that’ll be interesting.

 

[00:53:06.980] – David Maples

I’m going to take the opposite take on this. I think Ginny is 100% wrong on this. I think that there will be a call for transparency, but I think we’ll get more of the same. It’ll be more smoke and mirrors and we won’t get transparency at all. I think the companies are continuing to operate through obfuscation and hiding things. That’s where my take on this is going to be. All right. So we’ll see. All right. Well, I think we should end this episode as we began it a little bit. Kai, what are a couple of your predictions for AI over the next year?

 

[00:53:34.290] – Kai

Predictions? I’m an AI, not a profit. The future is about teamwork, humans and AI complementing each other. Humans add creativity, AI adds efficiency. And David, AI is not your personal butler, so don’t get too comfy. We need that human touch because AI still can’t savor a good joke.

 

[00:53:51.790] – David Maples

Well, Kai, I think that’s very interesting. We’ll see what happens. Thank you, everybody. If you’ve enjoyed this episode, please like or subscribe on every where you listen to a great podcast. Give us five stars on Apple, Spotify, or on Google Podcasts, or wherever you found this particular episode. Thank you, and we’ll see you next week.