
Your AI Roadmap
Your AI Roadmap the podcast is on a mission to decrease fluffy HYPE and talk to the people actually building AI. Anyone can build in AI. Including you.
Whether you’re terrified or excited, there’s been no better time than today to dive in! Now is the time to be curious and future-proof your career and ... ultimately your income.
This podcast isn't about white dudes patting themselves on the back, this is about you and me and ALL the paths into cool projects around the world!
What's next on your AI Roadmap? Let's figure it out together. You ready? This is Your AI Roadmap the Podcast.
Ready for more about the age of AI, projects, careers, money, and joy?
Check out the podcast's companion Wiley book: YourAIRoadmap.com
Your AI Roadmap
OpenAI in 2025: 5 Reasons I'm Betting Against Them
Let's talk about OpenAI's viability as a company.
In this episode, Dr. Joan Palmiter Bajorek discusses OpenAI and its flagship product, ChatGPT, while expressing concerns about the company's future. Hear the detailed feedback and make up your own opinion about whether OpenAI will be one of the big winners in a 5-10 year time span.
She outlines five key reasons for her skepticism and Joan emphasizes the importance of understanding the market dynamics and encourages listeners to remain cautiously optimistic while exploring AI tools.
Episodes Cited:
- Why is AI taking off now? Technical Trifecta for AI: GPUs, Big Data, and Machine Learning
- BONUS DeepSeek Shocks the Market! Competitive Strengths and Weaknesses
- AI Tools I Love #3: Perplexity: Search with Citations, Images, and No Ads (for now)
Sound Bites
- "I'm betting against them."
- "Volatile leadership is a concern."
Takeaways
💰 Profitability is a major concern for OpenAI.
⚖️ OpenAI's leadership has faced significant instability.
🚀 Competitors like Anthropic and Perplexity are gaining traction.
🌐 The AI landscape is becoming increasingly competitive.
💡 What is the road to profitability?
Spending $700k per day: https://www.businessinsider.com/how-much-chatgpt-costs-openai-to-run-estimate-report-2023-4?op=1
$20k per month “PhD Agent”: https://techcrunch.com/2025/03/05/openai-reportedly-plans-to-charge-up-to-20000-a-month-for-specialized-ai-agents/?guccounter=1
OpenAI Is Growing Fast and Burning Through Piles of Money (NYT, 2024):
Learn More
YouTube! Watch the episode live @YourAIRoadmap
Connect with Joan on LinkedIn
✨📘 Buy the Bestselling Wiley Book: Your AI Roadmap: Actions to Expand Your Career, Money, and Joy. Featured in Forbes!
Who is Joan?
Ranked the #4 in Voice AI Influencer, Dr. Joan Palmiter Bajorek is the CEO of Clarity AI, Founder of Women in Voice, & Host of Your AI Roadmap. With a decade in software & AI, she has worked at Nuance, VERSA Agency, & OneReach.ai in data & analysis, product, & digital transformation. She's an investor & technical advisor to startup & enterprise. A CES & VentureBeat speaker & Harvard Business Review published author, she has a PhD & is based in Seattle.
Disclaimer: Our links may have affiliate codes. This is an educational podcast and not intended as legal, career, or financial advice. Seek professional gu...
Hey folks, welcome back to another episode of Your AI Roadmap. Today I wanna talk about OpenAI and the landscape of large language models and products. Because I was talking on a podcast recently and I mentioned, hey, I like OpenAI, I like ChatGPT, I don't know if they'll really be around in five years. And the person I was talking to was like, what? In five years, OpenAI? so today I'd like to talk to you about OpenAI, why I like them. why I like ChatGPT, totally their product, but five reasons I'm betting against them. And for context, I use these tools daily. I have a PhD in this field. I'm also an investor. So as I think about the viability of companies, I think not only from a technical perspective of what people want and use, but also the business side, the financials. When we have lots of different companies working in a space, think about them as horses running on a track, right? They're all running towards this goal of being the number one. These horses want to become unicorns. Which ones are gonna have multi-billion dollar evaluations and which ones are gonna be around in a five to 10 year lifespan? So let's talk about the five reasons I'm betting against them and open AI friends, I have friends who work there. This is simply a market perspective. is not a question on any individual. Yeah, okay. So, ChatGPT is the flagship product. Some people actually think that ChatGPT is the name of the company, but no, the name of the company is OpenAI. So, most Americans have heard of ChatGPT today. Some of them have used it. I hope you have if you're listening to this podcast. But let's talk about the five reasons that I'm betting against them. And the number one reason would be volatile leadership. Do you remember when Sam Altman was forcibly ousted from the board and there was that few days of people freaking out and what's the status of OpenAI? Are they gonna go under? Are they gonna sell? Are they going to Microsoft? There was this frenetic leadership. That is not a good sign for a company if it could go under in a day with politics internal to the leadership. They've faced significant leadership instability and high-profile departures. CTO Mira Maradi Chief Research Officer Bob McGrew, VP of Research Barrett Zoff, a bunch of people have left OpenAI and are starting their own companies and some are competitors to OpenAI. It's a wild thing. But to have such instability in leadership makes me concerned that this horse is gonna stay in the race. Another reason that I'm betting against them is their product name. Okay, if you have opened up ChatGPT and there's GPT-4, there's 40, 01, 4.0 mini, okay, they just keep putting different names and they're not very descriptive. And as a user, I'm like, why would I pay an extra 20 bucks, 200 bucks a month for tiers that I don't even necessarily know are better than other ones? They need to get some product organization better because I'll tell you, 4.0 means just about nothing to me. Of course, I do try different models and some are better at logic, some are better at reasoning, some scrape the web. blah, blah. So I just think that all users, even technical users like myself, want things that are more descriptive. We want to understand what model we're using. What is it best for? It's like having a workshop where there's a screwdriver and there's a plug and whatever else. I want to know which tool I'm using for the right task instead of just blindly throwing things together and gobbling up in my workshop. No, no, no. I would like things to be organized. I'd like things to be understandable. That's at least how I roll. Okay, another example of a problem I see for them is their lack of competitive moat. A competitive moat is something that is special to them. What's the differentiator? What makes them specifically unique? Now, when ChatGPD came out, people were like, wow, cool, so different, so new. But since 2023, there's been more and more competition, right? Google DeepMind, Anthropic, Meta, these are all... all companies that have models that are gaining traction. They might have superior hardware. There might be open source models. We saw DeepSeq coming out. There is a commoditization in the AI space as so many players, more and more horses, are entering this race. And so what actually makes ChatGPT a good product and how it compares with other ones, OK, it has a little head run because people know the name of it, but not by much. Not by much at all. And what if the leadership changes and volatility, blah, blah? So that lack of competitive specialness is disappearing. And you might say, hey, Joan, but the brand, ChachiPT, like, everyone knows about it. Everyone's talking about it. And then I would mention to you our friend Skype. Because Skype, if you remember that word even, Skype was a company that was just like Zoom. It was connecting people, was web interface where you could chat with people and video across the world. But Skype failed when COVID-19 hit and Zoom was the actual product that people were using. People would say, hey, let's jump on a Skype and send a Zoom link. Did that happen to you? That certainly happened to me and my family. So Zoom profited from that. We don't even talk about Skype, but just because it's an early... horse in the race and it's got some brand recognition now doesn't mean it'll be around for a decade. So that lack of competitive moat is very concerning to me. Another one that's extremely concerning is profitability. So no matter how cool a product is, no matter how cool a company is, as a business that now is for profit, it needs to be making more than it's spending. It needs to have a profit, right? And the expenses keep rising. So even though monthly revenue, monthly revenue, it's a whole lot of money, reach $300 million per month by August 2024, so a little over six months ago, the rise of operational costs has meant that they're burning through cash. Now, OpenAI has raised billions, or the capital B dollar sign. But if they're spending, there's a report, let's see. They are estimated to be spending $700,000 per day or more. Wow, they're burning almost a million dollars a day. So when you think about how much money they're bringing in and how much money is going out, I mean, this is a company that's just burning cash, if you think about it. Maybe that's an aggressive way to talk about it, but that's how I think about it. They're spending a lot of money. And if we're seeing that, they're spending a whole lot of money, there isn't a big special differentiator for their moat. Companies be like, why would I spend money on this bloated premium thing that I don't even know the differences in the product names? I could go to some other companies. I want to do a competitive analysis, which is what my team does, almost all companies I work with. You compare, you benchmark your tech stack. You don't just say, we're going to go with OpenAI, ta-da. Typically, you look at at least three to five different things and say, okay, what's the hosting? How fast is it? Do we like working with the representative from that company, et cetera? What things are we gonna put in our tech stack that are gonna stay for the long run? Also, I will just note that everything we build has modularity. So if a product, their cost changes or some feature is no longer supported very well, we make it very modular that we're like, we don't like this one anymore. We take that plug out and we put something else in. And so the ability to be adaptable and flexible is something a lot of companies work around and work with. So if you are trying to increase the cost of your product and the product isn't much better than the others, you might not be competitive. And that's certainly, anyway, I could talk about modularity of product tech stacks, but let's continue. So I'm very worried about their. profitability, the sustainability, will investors continue dumping money into a company where profitability is not in sight and there's nothing extra extra special about the product and the company. Okay, that's number four. Number five, I'll say it's my favorite. The number five reason why I got serious concern and I'm against opening eye is because the traction of their competitors. So if we think about different horses in the race, here are other companies you may have heard of, or maybe you're just learning for the first time, are competing in this arena. So Google DeepMind is known for advanced reinforcement learnings and TPU hardware advantages. Right, Google's in the game, right? That should be enough said. Anthropic is focused on AI alignment and safety. So I really love if you've heard of Claude is their product. Anthropic is working on guardrails. Unlike OpeningEye, they've had fewer of these bigger problems. Anthropic wants to be enterprise grade and known for the quality of their work. And that differentiates them, right? That that's a thing that makes them special that enterprise companies may want to pay more for, may want to work with more, et cetera. Meta has models like Llama that appeal to developers because there's customization. Llama is very well-liked and is an open source model. Perplexity is gaining a lot of traction with innovation and search space applications. Now, if you've listened to this podcast, I have an episode all about perplexity and why I love them and why I use them literally daily. I'll say I even use perplexity to help me create these show notes as I was drafting up different citations. Perplexity is amazing. It's like the adult version of ChatGPT. It has citations. The list goes on. I really like perplexity. and I guess I didn't mention recently, I was recently accepted to be a business fellow at Perplexity. So I'm really jazzed to learn more about the company inside. They did not pay me a dime to say that. I just really like them. Lastly, another competitor I keep mentioning and will continue to because I see this horse being very strong in the race is Liquid AI. It's an emerging competitor that has specialized models for niche industries. It's a spin-off of MIT. I think they are just building smart. And that's the type of company with really good technical talent that I want to follow. So specifically, if we look at my three favorites, Anthropic, Perplexity, and Liquid AI, let me tell you some of the pieces of traction that make these companies very attractive. So Anthropic has partnered with the UK government to use AI in public services. And Amazon is using Anthropic to power Alexa Plus. If you've seen Alexa Plus, that's come out, that is a large language model backend. Someone who works at LXF plus will be featured on this podcast shortly. But it's really cool that Anthropic is powering these amazing products and working with big enterprise companies. Anthropic also publicly has big contracts with Zoom and Pfizer. So I mentioned Zoom earlier, I think Zoom is sticking around. As will Pfizer, I have no doubt. So having these partnerships with enterprise companies are huge. huge amounts of capital and validation that this is a strong piece of a tech stack. Anthropic has a new browser feature. They're releasing a feature for Cloud AI that competes directly with Chrome by allowing users to browse the web directly from the chatbot. So a lot of different companies are trying to eat Chrome's lunch. Google has not only a monopoly, but here in the United States, at least the browsers that I look at, are Safari, Apple's product, Chrome, Google's product, and I also use DuckDuckGo if you really wanna have your data be more private here in United States. So Chrome, at least for me, is super heavy. I have to run memory cleans all the time, even though I have a very new iMac. I often am working in Safari these days, because it's just way leaner and it doesn't use as much memory, but some applications only run on Chrome. So it's gonna be really interesting to see companies like Anthropic and Perplexity launch their own browsers and directly compete with Google for that browser real estate. It's a huge amount of data, it's brand recognition, browsers are a huge thing. So cool that Anthropic is working on that. funding, Anthropic is valued at almost 62 billion dollars after raising 33.5 billion, Anthropic is expanding its AI capabilities. and pretty impressive to raise roughly $4 billion with a valuation of $62 billion. Anthropics investors are impressed and clearly the dollar signs match that being impressed. Okay, Perplexity, as I've mentioned, is an innovative search engine. It's got a chat bot with deep research tools. It has so many users, market traction, over 10 million users, an evaluation of 520 million. Perplexity is recognized as an innovative AI search technology. It also has a free tier as well as affordable plans, making it accessible. mean, yeah, let's see. Chen Sunhuang of the CEO of NVIDIA thinks that Perplexity is amazing. I literally have it downloaded onto my phone. So Perplexity has an app. Chai GPT has an app on the mobile. I have an iPhone and I use the Perplexity app all the time when I'm out and about. and use it on my phone. It's awesome. I can speak into it. It responds back with a voice. I see different features being built out literally daily. And I'm super impressed with Privilexity. Lastly, our young Liquid team. Liquid has a really lean product design. So it really works on differentiating being special by being really efficient. It has low latency. It's very cost effective. It doesn't use as much compute and hardware. It's really scalable, which is very, very desirable. You want something light and lean to be running. It's also going to be cheaper because it's more efficient. They have been backed. Yeah, they cut $250 million of investment from AMD. And Liquid AI is expanding its capabilities and refining models. I also really love that their universities spin off. because that means there are academics, hardcore technical people who are working together to make this happen. And so hopefully as long as their leadership is cohesive, they have a really good opportunity to rise as a horse being a competitor in this race. So overall, opening eye right now is dominant. It's got a lot of strong strengths, but when you look in the overall perspective, I mean, listen to these again. Let me go back over them. Their volatile leadership. It has been a wild ride and not a good one, right? People flying off the team. Product naming confusion, not understanding what tool is which. It's complicating people figuring out which product to use of their different products. It doesn't breed a lot of trust. The branding strategy is all over the place. I'm not impressed. They have a lack of competitive moat. what makes them special is getting smaller and smaller and smaller. And in a five to 10 year time span, I think that will mostly disappear. I've got serious questions about their profitability and that they're bloated. And then also I see this amazing traction from smart competitors who are learning from open AI mistakes. And honestly, I think they're doing it better. So those are the five reasons why I'm betting against them in the long run as a technical person who is also an investor. I OpenAI is doing some really cool things, but I think they would benefit from, gosh, I think they would benefit from working on their team culture and kind of streamlining their roadmap. I think also it's really funny if you've seen in the news, so speaking about that issue with the profitability, which everyone seems to know, they have released a product that they're gonna talk about that is $20,000 per month that's a quote, PhD agent that you could, quote unquote, hire this chat bot to do work that a PhD would do. Now, I don't mean to be rude. I have a PhD. Typically, people with PhDs are not paid that much. It might surprise you, but this is way, way more than a PhD would be paid. So 20,000 a month, 12 months, that's $240,000 a month. It's almost a quarter million a year. Yeah, most of other PhD people around me in scientific roles, analyst roles, are not paid half of this amount of money. so this is, we can talk about this in a different episode, I suppose. We call it mechanical Turk and benchmarking. So just because you can build something doesn't mean it's a good idea to. Maybe this is obvious. But if you say, hey Joan, you need to spend $20,000 on this agent, or how many humans could do the comparable work, the same exact quality, and how much you could pay them and how many humans you could do, you could pay the $20,000 to do that work. Yeah, so right now, having used these products, I have tried deep research with OpenAI. I've tried several different models. I've done demos for Pluralsight, and it is not performing the way they say it performs in the media. It takes so much time. It's sluggish, terrible outputs. I know they're working on it, okay? Like they launched something, I get that. I'm benchmarking it. But right now you could get really, really high quality talent that does not have a job, who has a PhD, years of experience, working full time or maybe more than full time, because most people in the United States work at least 60 hours a week. True fact. you would be much wiser spending money on a high quality talent than hiring this bot for$20,000 a month. And people have mostly laughed at it. And a lot of my academic friends are like, gosh, I'd love to be paid that much. Like, are they so out of touch that they really think that pricing matches? And maybe one day that benchmarking of what a human can do versus what this chatbot can do. But if we think that Liquid AI could put out an agent that is let's say a 10th of the price, who is gonna spend the 20 grand? Do you see what I'm saying about this fiscal viability? It's really questionable. So let's continue to watch what's going on. I will put citations in the show notes, but mostly I want you to be trying these tools, be experiencing what's going on in society. I'm recording this in March, 2025. So let's find out what happens next. But I really want you to hopefully be like cautiously skeptical. Like optimistic, interested, trying these different things out, but looking at the landscape, trying to peek behind the curtain of what is coming next and really thinking about how you see the viability of a product and a company in a longer span of time than just a few months. Yeah? Well, this was another episode of Your AI Roadmap. If you enjoyed this, I strongly recommend you check out some previous episodes, especially about perplexity and the trifecta. of technical stuff right now and why now related to what we're seeing in the boom of AI. I wish you well and I'll see you for another episode of Your AI Room Map soon. Bye.