Elevating Conversations: Anthropic Unveils Premium Paid Chatbot Features

0:00The first thing I want to say here is Anthropic of course this is the AI company established by former OpenAI staff and it has unveiled Claude pro so this is its first consumer centricing subscription subscription subscription plan is a free. analyst chatbot. And this is priced very interestingly at$20 a month, US or 18 pounds in the UK. And the plan offers quintuple the usage of its free version and the availability to send significantly more messages and priority access during high traffic periods, along with early exposure to new features. So the thing that's interesting here, right, is$20 a month.

0:37this is the same price point as chat gpt for when you get gpt4 um pro and so it's interesting right they're they're not discounting themselves they're not saying like hey we're a knockoff they're saying look we're just as good as them um so i think that pricing point is really good they can't go higher right because chat gpt and gpt4 obviously is like the premier most popular tool so you can go more expensive than that unless you're doing enterprise but for consumers i wouldn't go more expensive but if you go less expensive then you seem like you know you're a knockoff you can't quite do as much.

1:07So I think smart move on their part, putting it at the same price point. So the thing I will say about it is you can do four times the usage of its free version, like based off of what you can use. This is a great move. And for those that don't know, Claude 2, one of its main selling points is the fact that you can upload massive, like super long, it has a really big context window. So you can upload super long prompts, aka, you can give it an entire document and say, hey, based off of this document, You can upload a PDF based off of this PDF. And then you ask a questions about it. I've seen people upload like little like small books and ask questions about the book in the PDF.

1:46And it will give them an answer. I, for one, use it when ChatGPT's context window is not big enough. Right. So typically it's going to be like I will post like three or four news articles in there and be like, hey, like tell me like, you know, five interesting things where you're like consolidating. like give me five bullet points based off of these, you know, four different news articles about X, Y, and Z topic. And chat GPT is like, Hey, you know, your, your questions too long. So I post it into Claude and I'll usually get a great response out of it. So that's typically when I use it, but I, and that's one of the biggest like features from it.

2:18Otherwise the quality of it, I've heard different people like in our Facebook group, I've had different people say like, Oh, Claude is better at X, Y, and Z. Um, but purse. And then like, I've actually had some people that say they just switched to exclusively using that. Personally, I feel like, I mean, I still personally like GPT-4 a little bit more on most things. I guess maybe it depends on what you're specifically using it for, but when I need something long, ChaiGPT doesn't do it, so that's where I'm going to go. I think, interestingly, the price point is, you know, coming, well, obviously it's the same as ChaiGPT, but since its inception in July, iCloud has gained attention for a lot of its advanced features.

2:57Anthropic noted in a blog post saying, quote, since launch in July, users tell us they've chosen Cloud as their day-to-day AI assistant for its longer context windows, faster outputs, complex reasoning capabilities, and more. The company also kind of mentioned that Cloud Pro would enable subscribers to make use of the chatbot's advanced model five times more than what is available on the free tier, like I mentioned, right? And addressing why there are limits on messages, Anthropic clarified that the substantial computational requirements to run Claude2 pretty much mandated these restrictions. So users can send up to 100 messages every eight hours, which is a boost from the 50 messages per three hour limit that you're going to find with, you know, ChatGPT plus subscribers, which I am a ChatGPT plus subscriber.

3:45So, you know, they also specify, quote, a model as capable as Claude2 takes a lot of powerful computers to run, especially when responding to large attachments and long conversations. We set these limits to ensure Claude can be made available to many people to try for free while allowing power users to integrate Claude into their daily workflow. So running AI chatbots obviously is not a trivial financial endeavor. This is incredibly expensive. OpenAI once reportedly shelled out, I think, around$700 ,000 daily to maintain chat GPT. And this, you know, amounts to around$21 million a month. So obviously, message limits in CloudPro can be reached kind of quickly, especially if the conversation includes substantial attachments.

4:33So for instance, if a user uploads the complete text of the Great Gatsby, they'd find their ability to send messages capped at roughly 20 for the subsequent eight hour period. So this is because Cloud2 revisits the entire conversation, including attachments with each new message, right? So it's actually looking at all the previous conversations, everything previously on there. And if you're putting something in as long as like the text of an entire book in there, it's going to take it's definitely going to use more. So Anthropic definitely has, I think, some broader aspirations targeting the development of a next generation algorithm aimed at AI self-teaching.

5:09And this was actually revealed in their recent investor pitch deck. So the ultimate goal is to kind of craft visual virtual assistants capable of a range of tasks, including email management, research, and creative output along among a bunch of other functionalities. Very, very interesting, right? And since it's launched in 2021, it was, you know, spearheaded originally by OpenAI VP of Research Dario Armedi, but Anthropic has gained over$1.4 billion in funding at a valuation in the single digit billions. So obviously, a very impressive figure, but it's still far from their estimated 5 billion that they need over the next two years to realize their vision for AI.

5:54So that's, you know, what Anthropik saying, they're saying, if we really want to, you know, hit our vision for AI, do what we would like to do, we're going to need$5 billion over the next two years, which is a fairly big ask. But I mean, it's not that like out of the question, considering Microsoft gave OpenAI 10 bill over$10 billion. So, you know, I mean, that's still half of what OpenAI has already done. And they have a couple years to kind of get this done. So assuming they get mass adoption, and they do some really impressive things, I don't see any reason why they wouldn't be able to get to that number.

6:27So revenue from offerings like CloudPro is slated to kind of fund computational capacity. The company's roadmap indicates a reliance on quote, 10s of 1000s of GPUs to train its models, which is essentially earmarking, I think, nearly a billion dollars for infrastructure spending within the next 18 months. So while Anthropic definitely enjoys the patronage of thousands of customers and partners, including a subscription-based generative AI app from Quora that offers Claude 2, and its less, I guess, powerful kind of counterpart Claude Instant, which is just faster. That's kind of why you'd use that.

7:05It still has some pretty stiff competition. So for those that don't know, you can go to poe.com and that is a platform that Quora owns and you can access, you can try out Claude today for free. That's typically how I use it when I need it for one-off things. If it ever comes to the point where I'm needing to do, you know, very frequently analyze large documents, then I probably just get a subscription to it. It's kind of a no-brainer. It's really good. I would assume if ChatGPT or GPT-4 improves the quality a bit and it comes up with larger context windows, then maybe that kills Claude. I don't know.

7:40I do really like having multiple options in the arena, right? I always hate having like one player in the space that like completely dominates. Then you get to sort of a monopoly thing. And a lot of times the quality of the products suffer. So I think there are a bunch of other competitors in the field. You have Cohere, You have AI21 Labs. Of course, OpenAI is huge. And I think that what's really interesting is OpenAI actually anticipates to generate a billion dollars in revenue next year. So obviously, big players in the space making big moves. But I think Anthropic has a really bright future and I'm excited to follow along on their journey.