Episode Categories:
Resources:
Join the discussion: Ask Jordan questions on Anthropic Claude
Upcoming Episodes: Check out the upcoming Everyday AI Livestream lineup
Connect with Jordan Wilson: LinkedIn Profile
Try Our Free AI Prompting Course: Register for our free Prime, Prompt, and Polish AI Course!
An Exciting Launch: Claude for Enterprise
Anthropic, a pioneer in generative AI, recently launched "Claude for Enterprise," a robust large language model aimed at revolutionizing business operations. This new product has been teased with intriguing features like a 500,000-token context window, the ability to process extensive datasets, collaborative tools, and stringent security controls. While these innovative features may seem tempting, care should be exercised before adopting Claude as your enterprise's go-to language model.
Product Highlights: Promising, But Not Without Concerns
Despite its impressive offering, Claude for Enterprise poses unique challenges that might curtail it from becoming the primary tool for enterprise users. The lack of real-time internet access significantly weakens Claude's potential as it hampers the flow of up-to-date information, a crucial necessity for businesses.
Third-party integration is another significant concern. While Claude's planned integration with Github is commendable, its lack of collaboration with other widely used tools may limit its utility in diverse enterprise settings.
The restrictive tier system even for paid users poses an obstinate problem for organizations looking to stress-test Claude's capabilities before large-scale implementation. With restrictive rate limits and a limited usage allowance, businesses struggle to evaluate Claude effectively.
Moreover, the ambiguity surrounding Claude's enterprise pricing could be disconcerting for businesses. In a marketplace where competitors present clear pricing tiers, Claude's customized pricing model lacks transparency, which may disturb enterprise clients.
A Competitive Landscape: Microsoft, ChatGPT, and Gemini
While discussing AI-based enterprise tools, Claude's competitors cannot be overlooked. Each boasts its unique strengths and features: Microsoft's 365 Copilot utilizes the advanced GPT-4 model, ChatGPT's Enterprise version attracts with an internal GPT store for organizational efficiency, and Google Gemini excels in backend operations despite its front-end shortcomings.
Real-time internet accessibility and robust third-party integration, especially, make these competitors more suitable for enterprise use.
A Word of Caution
Finally, the pricing structure should never be a caveat for businesses looking to adopt AI tools. A transparent, well-tiered pricing model not only presents the cost factor distinctly but also prevents potential misunderstandings and enhances customer trust.
Moreover, a product's integration capabilities and real-time access to the internet prove consequential. A tool lacking these features could be less than optimal for business operations, regardless of its other innovative features.
While Anthropic's Claude has much room for improvement, it's essential not to overlook its promising features, like the Artifacts feature for real-time code rendering and an impressive context window size. Perhaps with time and refinement, Claude may prove to be a valuable contender in the growing AI industry.
Paving the Way Forward
Facing competition from well-established models like Copilot, ChatGPT, and Gemini, Claude's journey forward will certainly be challenging. However, the addressal of the concerns highlighted can position Claude better among businesses and organizations looking to leverage AI's power.
Businesses must appraise these aspects before making a choice and stay tuned for any developments and enhancements which will shape their decision-making process.
Topics Covered in This Episode
1. Understanding ChatGPT's Internet Interfacing
2. Correct Usage of ChatGPT
3. Problems with Outdated Data in ChatGPT
4. Chain of Thought Prompting With ChatGPT
Podcast Transcript
Jordan Wilson [00:00:17]:
Anthropic Claude just released its new product for enterprise. Claude for enterprise. So I think a lot of people are going to be talking now. Should we be looking at anthropic Claude? Maybe your company is on Microsoft 365 copilot. Maybe your team has been using ChatGPT for teams or enterprise, and you might be thinking, okay, well, is anthropic Claude now the large language model that we should be using as a company? Well, I'm gonna go ahead and answer that today and say no. And also give you three reasons that businesses should not be using anthropic Claude at least just yet. Alright. What's going on y'all? My name is Jordan Wilson, and this is everyday AI.
Jordan Wilson [00:01:06]:
Welcome. Thank you for listening. Thank you for tuning in. This is a daily livestream podcast and free daily newsletter helping us all learn and leverage generative AI to grow our companies and to grow our careers. So if that sounds like you, if you aren't already, make sure you join that free newsletter. Go to your everyday ai.com. Every single day, we bring you the news, the fresh finds from all across the AI world, as well as exclusive insights breaking down, our podcast every day, whether it's me, going solo or having world class guests on the show. You can find it all only in our daily newsletter.
Jordan Wilson [00:01:44]:
Alright. So I'm excited today to talk about the three reasons why I think businesses should not be using, anthropic Claude's enterprise plan just yet. But before we do, we gotta start with what we do every single day with the AI news. Alright. So first, what did Ilya see? $1,000,000,000 in funding. Alright. So Safe Superintelligence, has raised over $100,000,000,000 to advance AI research. So the AI, the AI startup Safe Superintelligence Inc, which was cofounded by former OpenAI, or sorry, was founded by former OpenAI cofounder and OpenAI chief scientist, Ilyas Skeever.
Jordan Wilson [00:02:30]:
And, they announced yesterday that they've, raised $1,000,000,000 in funding, signaling a significant investment in the future of AI safety and research. Investors in the funding round include notable firms such as a 16 z, Sequoia, DST Global, and SV Angel. The funding round has reportedly valued Safe Superintelligence Inc at $5,000,000,000 indicating a apparently very robust, market interest in AI safety initiatives, or maybe people just want to be involved with whatever Ilya Sutskever touches. So SSI plans to allocate the funds toward acquiring compute power and expanding its team of researchers and engineers with operations in, both Palo Alto, California and Tel Aviv, Israel. And although sign, although specific research areas have not been disclosed, the company's focus is expected to align with general AI safety, a field such Skibra has previously championed. Alright. Our next piece of AI news, Michael Dell, has highlighted the strong AI demand at his company and others and stated that there is no AI slowdown. So Michael Dell, maybe you've heard of him, the founder and CEO of Dell Technologies, dismissed concerns about a slowdown in AI spending, asserting the current demand is robust and expanding across multiple sectors.
Jordan Wilson [00:03:56]:
So his insights reflect not only his prominence in the tech industry, but also the significant role Dell Technology plays in shaping the future of AI infrastructure. So Dell said this, recently at a conference in New York City yesterday, and the industry veteran emphasized that fluctuations in demand are normal during big tech transitions, comparing it to challenges faced when launching a rocket. Dell, the company has reported 3,100,000,000 in AI server sales for the second quarter, nearly doubling from the previous quarter reflecting a strong market response to AI solutions. Alright. And then last but not least, our topic for today, Anthropic has launched Claude Enterprise to compete with OpenAI's ChatGPT Enterprise and Microsoft 365 Copilot. So Anthropic is stepping up its game in the AI chatbot market with the introduction of Claude Enterprise, aimed at enterprise customers seeking enhanced administrative controls, security features, and a powerful large language model system for their company. This launch comes as a direct response to the growing demand for business specific AI solutions. So that's actually a great just transition, to just go ahead and kick the show off, and we will, go into that in a lot more detail, so don't worry.
Jordan Wilson [00:05:15]:
But, gotta shout out our audience. And, hey, let me know, audience. What is your company using? Let me know. Is your company using, ChatGPT? Is your company using, Microsoft 365 Copilot? Are you all using Google Gemini? Are you using quad anthropic? Do you have your own model? I'm curious. I wanna know, but let's just get straight into a quick, recap here of what quad enterprise even is, what it entails, and I'm gonna tell you the three reasons why I don't think enterprise companies should be using it. Alright. So quad enterprise, one of the biggest things here is it features an impressive context window of 500,000 tokens, enabling the processing of extensive datasets. As an example, that's 200,000 lines of code or a 2 hour audio transcript in a single prompt.
Jordan Wilson [00:06:10]:
Yeah. That's wild. And that's obviously much larger than the current context window offered by ChatGPT Enterprise and Claude's team plan, which are less than half the size. Yeah. We've done official, you know, we do consulting and front end strategy for companies. So we tested, ChatGPT Enterprise. It's, they say it's a 125,000 tokens. It's roughly about a 118,000 tokens.
Jordan Wilson [00:06:34]:
So, more than, what, 3 x that, almost from, quad enterprise. So pretty impressive in terms of the amount of data that it can work with. So the new enterprise offering obviously includes, Claude's highlight features such as projects and artifacts, which serve as, collaborative workspaces for users to upload and edit content, making it ideal for long term projects involving multiple data sources and team members. Also, a new announcement here from Cloud Enterprise, it does integrate or will be integrating with GitHub, allowing engineers, engineering teams to sync their, repositories directly with the AI, and this feature is particularly beneficial for onboarding new engineers, developing new features, or troubleshooting bugs. The other thing here, it's obviously big with team control. So businesses can assign a primary owner for their workspace who can manage access levels and monitor the activity for security and compliance purposes similar to that, feature functionality found inside of OpenAI's, ChatGPT Enterprise or obviously in, 365 Claude. Or sorry, 365 Copilot from Microsoft. So here's the other thing, though.
Jordan Wilson [00:07:46]:
The price of cloud enterprise remains undisclosed. More on that in a bit. But it is expected to be higher than Anthropic's team plan, which costs $30 per user per month. So the increased cost obviously reflects the additional features and capabilities offered to enterprise clients. So Claude did, or Anthropic did announce yesterday that they've been in a private beta with early adopters such as GitLab and Midjourney, indicating a cautious but strategic rollout. Alright. So and despite all these advancements, Anthropic is facing challenges and gaining broader adoption in a competitive landscape where API pricing pressures are prevalent. The success of Cloud enterprise will depend on its ability to attract a significant user base to off sense, to offset high inference costs.
Jordan Wilson [00:08:33]:
Alright. And just so you guys know, when we are talking about these products, for enterprise, we are talking about the front end systems. Okay. So we are not talking about, back end development and, you know, your your developers using the API and tapping into your own, knowledge base. That's not what we're talking about here. We are talking about front end consumer products. Right? So the everyday, business person, as an example, logging on to ChatGPT Enterprise, logging on to your, you know, Microsoft 365 Copilot, logging on to now Quad Enterprise. Alright.
Jordan Wilson [00:09:06]:
Let me just skip to the end y'all. I'm gonna skip to the end. Here's three reasons, and we're gonna dive into these more. I'm gonna go over the pros and the cons. If I'm being honest, this show, I'm doing it somewhat selfishly, because I know that I'm going to be fielding a lot of calls because companies reach out to us when they want to learn prompt engineering, when they want to learn AI, when they want to, you you know, implement Microsoft Copilot. Right? We just talked with a a company earlier this week that wants us to train, over the long term up to 70,000 employees. So we hear from a lot of enterprise customers because that's that's what we do. Right? Yeah.
Jordan Wilson [00:09:42]:
We bring you, everyday AI free every day, but companies pay us, to help them implement tools like ChatGPT, like Claude, like Microsoft Copilot. And I cannot recommend right now, if I'm being honest, I can't recommend anthropic Claude to any enterprise clients. And here's why, and we're gonna dive into these in a little bit. Here's the three reasons. Number 1, no Internet access. I can't I can't stress this enough. Okay? You have to have real time connectivity to the world if you are using a large language model. I understand how large language models work.
Jordan Wilson [00:10:21]:
Right? I understand they're different than search engines. But when you look at the main competitors, right, because that's ultimately what this is about. Enterprise customers, they're shopping around. Right? They're obviously building on top, you know, of these companies' APIs because the price wars is bringing these down, you know, to compute too cheap to meter. Right? But for front end users, you have to look at the competition. And Google with its Gemini, chat g p or OpenAI with ChatGPT, Microsoft with Copilot, they all offer some form of real time Internet accessibility, access. Right? Claude doesn't. 0 as of today.
Jordan Wilson [00:11:06]:
Wild. I cannot underestimate from a business perspective. Right? And you can say, oh, okay. Well, you know, users should always know. You should always just bring in the most up to date data. They're not gonna do it. Right? You have to work into account user error or sometimes user laziness. Right? So if you are not connected to the Internet, a large language model, yes, that is not for enterprise, period.
Jordan Wilson [00:11:34]:
And I will continue to say that. I don't know why Claude hasn't implemented real time Internet accessibility, or access. Right? They have a $4,000,000,000, in funding from Amazon, one of the leaders in, the Internet. We do know they announced, you know, Amazon's going to be using Claude for its future versions of Alexa, its AI smart assistant. So we know the capabilities to marry the Claude technology with real time information must exist. Right? And it exists in the other products. So I don't care. I don't care what your reasoning or rationale is.
Jordan Wilson [00:12:11]:
Oh, Jordan. You know, you can just upload all the documents you need. Okay. Well, documents change. Right? Information change changes Weekly, daily, hourly. A real enterprise business needs a large language model that has access to information real time. Period. Hard stop.
Jordan Wilson [00:12:34]:
I if I'm being honest, I'm not going to recommend anthropic Claude to literally any enterprise client that pays us to be like, hey, Jordan. Help us with an AI strategy. We want a front end large language model. Claude's not into consideration. It's not. You can't. It is dangerous for a company. I'm I'm saying it right there.
Jordan Wilson [00:12:52]:
It is dangerous because employees do not check. They do not verify. They take, unfortunately, what comes out of a large language model as copy and paste truth. Alright? Yes. There's still obviously problems with, you know, ChatGPT and Copilot and Gemini in terms of, retrieving real time information. And, yes, there's still, likelihood for hallucinations and and things that are just wrong, but you gotta, like, you gotta have that to play. It's like the minimum height for a roller coaster. Right? It doesn't matter how effective the real time Internet accessibility is.
Jordan Wilson [00:13:29]:
If you don't have it, you don't get to play in the enterprise, period. Reason number 2, limited third party integrations. Right? There's one right now. There's one. And it's not even for everyone. GitHub. Right? More on that in a second. But you have to bring integrations that enterprise companies use daily.
Jordan Wilson [00:13:53]:
Right? That's the whole thing. Right? And I talk about how right now, chat gbt and technically by proxy, Microsoft Copilot, their business operating systems. And I think Google Gemini is getting there, although there's some problems with connectivity. But you have to bring in all of the tools and software and processes that your employees are already using. If a large language model can't fit into a workflow, it shouldn't be a part of the workflow. So with such limited integrations, I don't think again, I don't think that this should be used for the most part. Alright? And when I'm when I'm saying this y'all, I'm talking about 90%. I'm talking about 90% of enterprise businesses, or enterprise companies.
Jordan Wilson [00:14:37]:
Are there gonna be 10% where anthropics Claude is a absolute no brainer for sure, but I'm talking top to bottom implementation. Right? So, yeah, like, as an example, developers, developer, development teams. This is great. Right? I'm talking about for an entire enterprise company here. Reason number 3. The other tiers are too limited to test. K? Hey. Someone from Anthropic, if you're listening, you guys should know this.
Jordan Wilson [00:15:06]:
This is how it happens. Alright? This is how large language model implementation, at the enterprise happens. Usually, very early on, a couple of, rogue users are gonna start to use a large language model probably before they have official permission to do it. Okay? They're gonna find some great productivity gains. They're gonna, you you know, start using it department wide. Maybe it starts to become a little more official. And then if and then eventually, you know, board signs off, c suite signs off, AI, you know, ethics and, guardrails, safety, AI policy, all that falls into place, and then it becomes an enterprise offering. Right? And then essentially, companies will offer, you know, enterprise accounts to, you know, sometimes a couple 100 test users, couple 1,000 test users.
Jordan Wilson [00:15:56]:
And then they kind of figure it out. Right? But right now, the free and even the base, paid plan for anthropic Claude is so limited in terms of the messaging limit. Y'all, I kid you not. We have and we've had since it launched the paid, the baseline paid, plan for Claude. The rate limit, especially when using some of its more powerful features that have come out, which I love, and I'm gonna get to 20:20 to 30 minutes, tops. And then you're locked out for for 4 hours. That's come on. You can't be serious.
Jordan Wilson [00:16:37]:
You can't be serious. If you want adoption from enterprise companies, you gotta give them a taste. Right? And you can say, oh, well, yeah, Jordan. On the free plan, you can even do a couple messages. Okay. Well, you can't figure out hardly anything in a couple of messages. You can't. Right? Literally, we've burned through our paid our paid account in less than 20 minutes.
Jordan Wilson [00:17:02]:
That's not serious. That's not serious. That's like, if you want adoption, if you want to compete in the enterprise game, you have to give people enough on a basic played pan paid plan to see if it works for their needs. You can't do that right now. It's you know, if you go search, go read online forums, Quora, Reddit, Twitter, etcetera. It's it's probably the number one complaint about Claude is is people are like, hey. As soon as, you you know, we started testing it, we're we're we're locked out. The limits are comically low.
Jordan Wilson [00:17:40]:
Alright. So what's everyone using? Hey, Brooke. First time first time listener. Says we don't use any of it. Monica says Copilot. Dennis says ChatGPT for teams. Yeah. So let me know what is is is anyone using, you know, anything at the enterprise level? So I know that might have sounded a little harsh, but if you tune in to everyday AI, I want you to have the reality.
Jordan Wilson [00:18:10]:
Okay? I want you to have the the truth. No holds barred, but don't get me wrong. Claude is good. Anthropic Claude is super impressive. Alright? The artifacts feature, which we've covered many times on the show, we've we've on our YouTube channel, we've shown you some amazing, some amazing ways to use artifacts. Artifacts is by far the most underutilized and probably if I'm being honest, one of the most impressive features of any large language model bar none. The artifacts feature. Right? And so if you haven't used it, essentially, what happens is you can chat with, Claude.
Jordan Wilson [00:18:54]:
And on the right hand side, it can render code. Right? That's amazing. So you can, you know, go into ChatGPT or Copilot or Gemini, and you can write code. Right? But then what you have to do is you have to copy and paste it. You have to go into something like repllet, you know, run it, do some troubleshooting, or use a a tool like, GitHub. Right? So you have to use kind of a a more developer focused large language model or AI powered tool. Claude with an with with artifacts. It's mind blowing.
Jordan Wilson [00:19:25]:
You can literally create entire websites with a single prompt, render them, see how they look, play around with them. You can build, little little apps, you know, in in Python, JavaScript. If you wanna go old school, you know, CSS, HTML, etcetera. In a single prompt, you can build anything even on your phone. Don't get me wrong. That is mind boggling. And that is why I said, for maybe 10% of of companies. Right? If you are essentially a development company or if if you're looking at this on a team's level, right, Claude is great, but top to bottom, no.
Jordan Wilson [00:20:00]:
But the artifacts feature is unmatched. Right? We gotta highlight what's good out of Claude, because it's amazing. The context window, even with the paid version, the context window is better than ChatGPT. And y'all, 500,000 tokens. A 500,000 token, context window for the enterprise version. That's that's nuts, y'all. That's nuts. So, yeah, I I I talked about that.
Jordan Wilson [00:20:26]:
That's 200,000 lines of code that you can have a conversation with, or as an example, a 2 hour audio transcript in a single prompt. Right? That's like you can speak with your company's entire knowledge base in many cases without having to float between multiple products, multiple GPTs. The context window is great. Projects. Projects are powerful. Claude Claude anthropic's power, projects, great tool. Very similar to, as an example, custom GPTs from open AI. Also, what Google is trying to do with gems, we covered that last week.
Jordan Wilson [00:21:05]:
There's some shortcomings there. Right? But, essentially, Claude as well has a very easy way, no code, drag and drop. Any user out there can create a kind of customized bespoke version of the big model. You can give it custom instructions, upload your, files, and then, you know, essentially, you can create a ton of different versions of, anthropics Claude. Right? You can do something for, you know, customer, customer success. You can do something for for marketing copy. You can do something for specific, coding and and upload your, you know, company's repositories, etcetera. Right? So great, great free, features that we get from Anthropic.
Jordan Wilson [00:21:53]:
And also, the content, the actual content writing out of the box is much better than ChatGPT. Right? And I think people and I'm gonna get to that here in a second when we talk about the competitive landscape. But I think people just think, oh, you know, when I talk to anthropic Claude, it sounds more natural. It sounds more like a human. It sounds less robotic, so people think that just that means anthropic Claude is better. No. It's not. But anthropic Claude is great in terms of giving you more realistic human sounding content.
Jordan Wilson [00:22:27]:
And when you're working with a large language model because I'd say so many, you know, I'd say more than half of use cases that we see are ultimately creating written content. Right? Whether that's marketing copy, emails, SOPs, templates for contracts, etcetera. Right? So you want something that writes well. But that is again, y'all, using, or thinking a large language model is for writing content, I think is like, you know, using a a a Lamborghini as an umbrella. Right? Oh, it's raining. Better go get in a Lamborghini to stay to stay dry. No. If a Lamborghini, it's meant to go fast.
Jordan Wilson [00:23:08]:
It's meant to take you places. Right? That's what, like, an analogy I use sometimes when people think a large language model. Oh, we don't need help writing content. It's like, no. That's you know, it can literally automate probably 80% of the manual, knowledge work tasks, that your company does. Literally. That's why so many, you you know, companies are going all in on large language models because they understand that you can get probably about 80% of the work done in 20% of the time. It's not just writing.
Jordan Wilson [00:23:36]:
But you gotta tip your hat to Claude Anthropic for or sorry, Anthropic's Claude for its ability to write great content out of the box. So, yeah, don't get me wrong. It's not all bad. There's a lot of good. But let me also cut it to you straight. It's not hot take Tuesday, Michael. Yeah. Michael said, is today Tuesday? Because I'm dropping some fire.
Jordan Wilson [00:24:03]:
You know what? We did actually have the holiday this weekend, so, read the news on Tuesday. So, yeah, maybe maybe I got some hot takes, still still in the tank. And y'all let me know let me know. It seems like we're having audio problems again. I'm sorry y'all. But, also let me know what questions do you have. If you have questions, I'm gonna try to get to them at the end. But let me just cut it to you straight.
Jordan Wilson [00:24:31]:
I think Claude completely missed their enterprise go to market strategy here. Alright? Are there things I don't understand? Absolutely. Are they going to improve the enterprise product? Yes. Obviously, they're going to improve the enterprise product. You only have one shot. You get one shot at a first impression. And the first impression, if you know what you're talking about, if you investigate top to bottom. Very few enterprise companies have a actual reason, I think, top to bottom, again, to choose anthropics Claude over what's already out there in ChatGPT, in Microsoft Copilot, and in Google Gemini, and, you know, Google's, Gemini across workspace.
Jordan Wilson [00:25:21]:
K? Those three flaws that I mentioned, they're they're they're just just too great. Alright? You can't if I'm being honest, right now, you go to market. You go to market with those three flaws. Number 1, no Internet access. Number 2, such limited third party integrations. And number 3, the tiers are too restrictive for even paid users. So how are you going to give people a taste of what it can do for their entire enterprise? And it's actually a little difficult to get the enterprise account more on that in a second. But I think those three flaws are too great.
Jordan Wilson [00:26:01]:
I think they whiffed it, if I'm being honest. And like I said, Claude is not some un unnamed startup here. Right? It's founded by former, OpenAI execs. They have 4,000,000,000 in support just from Amazon alone. And they should have figured out they should have figured these things out. Those three flaws at the minimum. They should have figured that out. Y'all? And my biggest one, my biggest gripe, like I said, not having real time Internet access.
Jordan Wilson [00:26:34]:
You have a partnership with Amazon that kind of quite literally runs the Internet. Right? Probably whatever social media platform you're listening to now or whatever podcast platform, my live streaming platform, they're all probably powered by AWS, Amazon Web Services. So the fact that anthropic Claude couldn't get what I think is a bare minimum bare minimum real time Internet access. It's mind boggling to me. I'm sure they're gonna release it any week, any month now. I mean, they have to. Right? But the fact that they are going to market to the enterprise. Right? That's your grand opening.
Jordan Wilson [00:27:10]:
That's your grand opening to the business world. Hey. We're open for enterprise. Missed it. Alright. And I will say this. Normally, Claude has nailed its go to market, its rollouts. Right? And let me let me just go ahead and, give you guys the quick, and we're gonna talk competitive landscape here in a second.
Jordan Wilson [00:27:35]:
I think Google has been kind of notorious, right, especially with AI. They put out some great marketing videos. They tease a bunch of stuff, and then we have to wait many months. Right? As an example, Google Gems, just came out last week. It was announced 5 months ago, 4 and a half months ago. Right? It's not good. You get not as bad, but similar with with Microsoft. And, you know, they've had some, some delays in some of their features for privacy concerns, so I get that piece.
Jordan Wilson [00:28:07]:
And I think OpenAI has also fallen victim to this, like, oh, we're announcing things and rollout's not really gonna happen. Right? The features exist because they're going out and, you know, alpha rollouts or very limited, beta access for some of their products, like search GPT, for some of their, features, like advanced voice mode. Right? So some users are are getting access to some of these features with OpenAI, but I'd venture to guess it's less than 1%, very limited rollout. Right? But I think up until this point, Claude has been the opposite. I've loved the way that, anthropic has gone to market. Right? We've seen that with our last, kind of 3 big, you could say, with Claude 2, with Claude 3 or 3.5 Sonnet, and then with some of their, features like projects and artifacts. Right? They essentially just dropped them. They announced them, and they said, oh, available for everyone today.
Jordan Wilson [00:29:05]:
So I don't think they nailed this roll out because number 1, it's not ready. It's not ready for the enterprise. Number 2, not even all the features are ready, and even access is a little cloudy. Cloudy. I didn't mean to do that. I'll say I'll say I did. So let's talk about this. Pricing is confusing.
Jordan Wilson [00:29:28]:
Y'all. That's one of the first things enterprise companies are gonna wanna look for, and the pricing is confusing. Right? Doesn't say on their website, which isn't y'all. That's let me say this. That's not abnormal to not have enterprise pricing on a website. Right? People wanna get you to your your company to sign up for a demo. They wanna sell it to you, etcetera. Right? But usually once that happens, they tell you.
Jordan Wilson [00:29:52]:
Right? So, as an example, ChatGPT Enterprise has a flat rate. That's what they charge. Claude, we're not sure what you gotta pay. So, this is from CIO Dive. So the website, shout out to them for getting this scoop. So I'm gonna read this from their website, but it's quoting, a product manager at Anthropic. Here we go. Pricing for the tool is customized to each organization's scale of use and depth of integration with company systems.
Jordan Wilson [00:30:26]:
Scott White product manager at anthropic said in an email to CIO drive organizations will work with our sales team to understand their unique use case and tailor a plan that offers the best value for their specific AI implementation goals. Oh, gosh. Podcast audience, you can't see this. I'm literally face palming. That's that's not how technology should work, because I'm already telling you this. That right there is going to turn off so many enterprise companies. Like, oh, there's not even a rate. It depends on what we use.
Jordan Wilson [00:31:05]:
Y'all, it's yes or no. Do you get a seat? How many seats? 50, a 100, 10000? And you get access to everything. Alright? I'm sure there there there's reasons why they're trying to have this, you know, per company, pricing doesn't make sense. Maybe it's based on usage. That's, you you know, that's definitely a possibility there. But it okay. Put a tier on it. Okay.
Jordan Wilson [00:31:31]:
Enterprise level 1 is this. Enterprise level 2, if you have crazy, you know, if you're gonna be using it a lot, here's what it is. Don't base it on a on on a per company. That's wild. That's y'all makes zero sense. Again, that is going to scare away so many enterprise company, customers. At least say, oh, starts at, you know, $30 per seat per user per month, whatever. Right? I don't know.
Jordan Wilson [00:32:03]:
Is it gonna be 50 for 1 company? Is it gonna be a 120 for another company? Is it gonna fluctuate? Is it gonna be a variable rate? Y'all, anthropic. You only have one shot. This is not good. This is not good. If you don't control the narrative, the narrative is going to be controlled for you. So what the narrative is probably going to be is, hey. No one knows anything about this anthropic enterprise. They're they're not being forthcoming.
Jordan Wilson [00:32:35]:
Right? Not saying they're trying to hide anything, but if you're not being forthcoming, and the fact that you put out a statement like this saying, pricing for the tool is customized to each organization's scale of use and depth of integration with company systems, that's bad. It's a bad look. At least say, hey. We have 3 different tiers. And depending on a company's use, we place them on one of those 3 tiers. To me, this this is like, don't touch it. Don't touch it. If you'd if you are literally if you're an enterprise company, this is the hottest space right now.
Jordan Wilson [00:33:13]:
Companies want large language models for their users. Yes. They want to build on the back end with their APIs, but they want front end. If you can't say here's our pricing system, don't like it. And remember that one integration that one integration? That's one of my gripes. It's not even live for everyone, apparently. So this is from, someone at anthropic, Alex Albert. Alright.
Jordan Wilson [00:33:43]:
So he said, today, we introduced Claude for enterprise, the best way to securely work with Claude within your org. We're also launching a beta GitHub integration for early enterprise users. We plan to make this more broadly available to claw dot ai users later this year. So there we go again with the rollout. Right? The go to market enterprise rollout. So number 1, who is this available for? How much does it cost? Not sure. Number 2, seems like there's a lot of features that aren't there that should be there. Number 3, you know, they're really pumping the GitHub feature, the GitHub integration, which I love, especially for development teams, but I don't it looks like we're not getting it.
Jordan Wilson [00:34:23]:
Why not wait? Claude? Anthropic? Why not wait? I don't know. Alright. Let's look at the yeah. Marie says loss of transparency loses trust. I agree with that. Cyber here from YouTube said, agree with Jordan's points, pay for ChatGPT in per in perplexity. Won't pay for Claude due to the limits. Yeah.
Jordan Wilson [00:34:48]:
The limits are, mind boggling. I don't understand it. All right. Let's quickly talk about the competitive enterprise landscape before this announcement. Right? So now obviously Claude is in the fold, but Microsoft 365 copilot is the leader bar none. Right? If you don't know, they use the GPT 4 o technology, from OpenAI. Alright. So if you have 365 Copilot enabled, It's not even close.
Jordan Wilson [00:35:20]:
Right? That brings the most powerful model to your desktop. Right? It brings it to your Outlook, to your, to your Excel, to your PowerPoint, and all of those Microsoft programs can talk to each other and work with each other with your data up to date in real time. However and I don't know why so many companies don't aren't enabling it at the, you you know, at the quote unquote Microsoft 365 copilot level. They're only enabling, kind of copilot chat. Right? So with that, you know, it's like, oh, your company's data is is secure and safe, but you're really just using it in the chat window, and you're not getting access to Copilot, across their their enterprise of of tools, which is kind of, you know I'm not saying it defeats the purpose. It doesn't. Because then it's still, you know, kind of, in theory on on on par with using, you know, Chat gbt or using, the tool in a web interface in a safe, environment. Alright.
Jordan Wilson [00:36:20]:
But, you know, Microsoft 365 Copilot, when it is enabled across the enterprise, there's no competitor. There's no competitor. It's not close. Alright. I'll say number 2, chat gbt enterprise. Fantastic tool y'all. And, like, one thing I love about chat gbt enterprise that I don't think we talk about, you know, you can essentially right? You have essentially a store. Right? So that's just for your company.
Jordan Wilson [00:36:46]:
So everyone can go on there and create different GPTs for different purposes. And you can literally go on there as a ChatGPT enterprise user, see what all of your teammates have built, and instantly start clicking it and using it. Think. It's like let's say you have 10,000 coworkers from all over the world. And I love when companies are doing this, and this is something that we work with companies on. We encourage them build GPTs for everything. Put them in the store, have little 30 minute stand ups once a week that talk about GPTs that you're building, the problems they solve, etcetera. Right? And then go in there.
Jordan Wilson [00:37:21]:
Right? People, smart companies are doing this, and you can go in there and see what other coworkers in your organization are building, the problems they solve, and start saving time immediately. Right? Love that. I love that feature out of ChattopT enterprise, but also, with working with other GPTs, the integrations are essentially unlimited. Right? They are. Right? Especially in the enterprise environment. And then you have Google Gemini. So I am hard on Google Gemini. Yes.
Jordan Wilson [00:37:51]:
Because I think for front end users, it's not quite there. On the back end, it's amazing. Right? Gemini 1.5 pro, its context window, some of the features. I mean, it can, the multimodality inputs, mind mind boggling. Right? Front end, not quite there. Right? And especially, there's there's some, I think some problems in the workspace, when you're using it kind of at the company or organization, organizational level. But that's that's kind of the competitive landscape right now. And then last but not least, right, when we look at the competitive landscape, gotta call this out.
Jordan Wilson [00:38:33]:
And this, again, I think go to market strategy doesn't make sense. I would have waited if I was anthropic. You know they're gonna be dropping a new model sometime in the coming months. Right? Why not wait for that new model to drop? Why not wait to get this Internet connectivity thing figured out? It's a must. Bare minimum. Bare minimum to play in the enterprise. You gotta offer some level of real time data. Everything that we work on, literally requires up to date information.
Jordan Wilson [00:39:06]:
Working with, you know, information on in in a knowledge cutoff that's many months old, many quarters old. It's wildly wildly dangerous for how people use it. Alright. But here's the other thing. Claude right now is not a top model. Right? You always wanna work with the best. Don't be wrong. When Claude or or sorry.
Jordan Wilson [00:39:27]:
When, anthropic released Claude 35 sonnet a couple of months ago, it held kind of the top, model standing on the, on the chatbot arena leaderboard, leaderboard. Right? So the lmsys chatbot arena leaderboard, we talk about that a lot here on the show. It's essentially a a fair way to see what the best model is. People rank them blindly across different categories, and you'll see right now, Claude is not a top technically not a top 3 model. Right? You have ChatGPT four o latest, as the most powerful model, not even close. You have Gemini 1.5 pro. That's just the version that's available, on the back end. So not for the front end.
Jordan Wilson [00:40:09]:
Then you also have out of nowhere, grok 2, came out. That's also above. Right? So technically, it's it's 4th when you look at, you know, what else is out there. So we know probably that whatever, anthropic releases next, presumably, they'll be coming out with a 3.5 haiku and a 3.5 opus. Right? So Sonnet was technically their middle model. They upgraded Sonnet from 3.0 to 3.5. So that's technically their best model is the middle model. So we've known that anthropic any month now is going to update and, you know, bring out a 3.5 opus and a 3.5 haiku.
Jordan Wilson [00:40:48]:
And presumably when they do, it's I'm I'm guessing it's it's going to take over the leaderboard. So why not do that all in one swoop? Right? Why not say we have the world's most powerful model, we now have access to the Internet, and we are bringing it to the enterprise? To me, that no brainer. Then every company out there has to give anthropic Claude's enterprise a serious look. But right now, nope. You can't. Alright. Here's here's the end takeaway y'all. Cloud Claude is fantastic.
Jordan Wilson [00:41:28]:
Alright? 3.5 SONNET, I love using it for certain use cases. Artifacts and projects, especially artifacts. Game changing features that I think all other large language models are obviously gonna have to copy. Claude is great. It's not ready for the enterprise right now. They missed the go to market. And when presumably, you know, in the coming weeks, we do this all the time. You know, companies hire us to, you know, help them figure out their AI strategy and they, you know, say, Hey, what are the best tools? What are the best large language models? What should we be using? Can't.
Jordan Wilson [00:42:07]:
Still can't recommend anthropic law. Don't get me wrong. I want to. I want to recommend anthropic law at enterprise, but they they messed this up y'all. It's not ready right now. Again, for the majority. Will it fit certain teams? Absolutely. Will it fit certain companies? Absolutely.
Jordan Wilson [00:42:29]:
But your average enterprise company that wants a large language model, their large language model of choice. Right? In the same way that you have to select an operating system. Right? You say, are we Mac or PC? Right? Are we, you know, what's our CRM? You choose 1. Most companies are going to be choosing if they haven't already a large language model. Right? What's our email service provider? What's our operating system? When we give phones out, what system right? You make a decision. You make a decision. What is our in house large language model that we're giving to our company for the front end? Not talking about building on the back end. What are we giving our users to use on the front end? Claude enterprise across the board today? It's not it.
Jordan Wilson [00:43:20]:
Alright. Let's see. I think I saw 1, 1 or 2 questions, maybe a comment as we as we wrap it up here. So, Monica says love perplexity using it more for personal life, but should integrate it professionally too. Everyone should use it as the best. Yeah. Big perplexity fan. Right? Answers engines.
Jordan Wilson [00:43:39]:
That's why, you know, Google has shifted toward that way. That's why, ChatGPT, open AI, coming out with search GPT. So Jay is asking, still blows my mind models aren't connected to the Internet. Is there a good reason? Jay, I'm I'm sure there I'm sure there is. Right? I'm sure to maintain a certain level of output, having the, kind of an Internet connected, large language model, I'm sure brings along risks. Right? Risk that I don't technically understand, from, you know, being a a multibillion dollar company, an AI startup. But for consumers, consumers need it. Right? Enterprise leaders are demanding it.
Jordan Wilson [00:44:24]:
We talk to them all the time. Right? One of the things companies love about ChatGPT enterprise is it's connected to the Internet when you need it to be. Right? And you wanna bring your workflow all within that kind of AI chatbot window. You wanna be able to work with your files. You wanna be able to work in a secure environment, and you want to be able to quickly bring in real time information. You don't wanna be working with information that's 5, 6, 7, 8 months old, because like I said, it is dangerous in terms of the output. Right? More hallucinations, less, likelihood to be true or up to date or accurate if you're working with old information. Simple as that.
Jordan Wilson [00:45:05]:
And the pricing. Yes, Fred. The pricing. Not a fan of that. Not a fan of that. I don't think you can base pricing on how a company is gonna use it. And maybe that's not the truth, right? Maybe there are 3 tiers, but to put out a statement that essentially says, ah, pricing depends on how a company's gonna use it. That doesn't give me a lot of certainty.
Jordan Wilson [00:45:27]:
Right? Say, hey, here's the 3 tiers we have, depending on usage. You know, this many tokens or this many messages, 1, 2, and then 3 is unlimited. You know, your company of 5,000 users can use it to their heart's content. Sure. If you gotta pay, I don't know, couple $100. Sure. Whatever. But you can't you can't put out statements that it looks like, oh, pricing is an unknown.
Jordan Wilson [00:45:52]:
It's gonna scare away more people than are gonna fill out that form. Alright. That's it y'all. I hope this was helpful. Like I said, the three reasons businesses, I don't think right now should be using Anthropic Claude's new enterprise company. Maybe 4 with the pricing, but number 1, no Internet access. Number 2, very limited third party integrations. And number 3, the free and even the baseline paid tier are so limited that companies can't even give it a real test drive.
Jordan Wilson [00:46:26]:
Alright. I hope this was helpful y'all. If so, if you haven't already, please share this with someone. If you're listening to the podcast, please subscribe, follow follow the podcast on Apple, Spotify, wherever you are. Check out the show notes. I always leave my my LinkedIn, URL. Maybe you want to, you need some help on AI strategy. Feel free to reach out to us, put my LinkedIn in there as well as our email.
Jordan Wilson [00:46:51]:
Share this if you found this helpful. We spend countless hours planning these shows so we can give you real unbiased information. Alright? It helps so much when you just click that repost. Share this with someone if it was helpful, and go check out our thanks a million giveaway while you're at it on our website, your everyday ai.com. Check out the daily newsletter. We'll be wrapping, today's and recapping today's, podcast. Hope it was helpful, and I hope to see you back tomorrow and every day for more everyday AI. Thanks y'all.