Anthropic launches Claude Science and a new model

0:00Anthropic is launching Claude Science. This is going to be a workbench for researchers. Base44 is launching their very own AI model. They're trying to protect their 100 million dollar in annual recurring revenue. for their Vibe Coding business. And X has just launched a hosted MCP server. So it's going to open its API to Claude Cursor Grok Build, which if you know anything about the backstory of X and their API, this is quite a big deal. Anthropic is shipping Claude Sonnet 5. It's going to be$2 per million input tokens. And Google is shipping Nano Banana 2 Lite and Gemini Omni Flash 2 developers.

0:38So tons coming out from basically every top AI model today. We're going to get into all of that. Hopefully it's not going to be too noisy from my end. It is a hot sunny day in North Carolina where I'm at right now and I am currently bouncing a baby on my lap. My wife did not sleep a lot last night, so trying to let her get some rest holding the baby. You probably heard me say this before on the podcast. But anyways, if you can hear a little gurgling and gooing in the background, that would be what it's from. If there's any repetitive AI tasks that you do, like writing newsletters or emails or generating graphics or anything you do on a repeated basis, I would love for you to go check out my startup, AIbox.ai.

1:18We have a no-code AI app builder where you basically just explain what you want it to do, whatever your workflow is, and it automatically links together different AI models and creates a workflow for you that you can use on a repeated basis. this. This saves me a lot of time because AI definitely is a huge unlock for being more productive. But at the same time, if you're just using it to do the same thing over and over every single day, you should probably just automate that. So if you want to give it a try, go to AI box dot AI slash builder, I'll leave a link in the description to go check out the builder platform that we have.

1:50We have over 80 different AI models that you can link together and build some really incredible automations to hopefully save you a ton of time. And it's only$8.99 a month to get started. I'll leave a link in the description. Kicking it off with Anthropic, they've just launched Claude Science. So this is going to be a workbench. It's connecting 60 different scientific databases and a bunch of prebuilt research tools for genomics, protein structures, chemistry. And all of this is going to run on Claude Opus 4.8. So sorry, Fable 5 not there yet. I'm sure when that comes out, we may get that rolled in as well.

2:21And it's going to come with$30 ,000 in free credits. They're going to be giving this way to 50 different academic projects. and they're making this bet that basically workflow is going to beat raw capacity for getting these researchers to be using anthropic. So there's no new model or gating but Claude Science uses the same Claude Opus 4.8. It's basically available to everybody and there is no biology fine-tuning or specialized weights. They've been testing this with a few different people and have had some impressive wins. They had UCSF Brain Tumor Center that was using it to compress geloma germaline analysis workflows we have the allen institute's jerome lecoq who is building a computational review pipeline and we have novo nordisk who's one of their launch partners what it's actually doing is that it's able to spin up these multi-agent architectures and they're going to let all of these scientific researchers create parallel sub-assistance and they're going to be doing sequence analysis they're going to be doing structural prediction they're going to be doing fact checking and the reason why this is so powerful is because all of that can be happening while they're also keeping their full reproducibility.

3:26So the code, the environments, the message history, all of that they're going to be able to reproduce. So there's a bunch of basically custom tools that scientists need, and Anthropic has built that in. They're giving away a bunch of grants, hoping to show people that Anthropic is the best place if you're doing science. Okay, Base44 is launching their very own AI model, and the reason why is, by the way, I didn't realize, but Base44 is owned by Wix. It's an incredible vibe coding platform. I've been really impressed with it. It's not the one that I primarily use because I'm just using Claude code for mostly everything I do.

3:58But if you're getting started, I think anywhere between Lovable and Base44, those are my two favorite. Anywhere is also a great platform. But in any case, Base44 is$100 million in annual recurring revenue, and they just launched their brand new model called Base1. So this is their very own LLM. This is actually very similar to what Cursor is doing, by the way. But this LLM is trained on tens of millions of real user interactions. So what's interesting is because they're able to, you know, have tens of millions of users using them to create their tools, they can take all of those conversations.

4:27And I'm not sure if it's every single user or if they had people opt in. It's probably every single user and it's just part of their terms of service. But they're taking everything that those people have been saying and they're training their model to basically fine tune what people need when they're building websites, apps, and software. Now, what's interesting to me about this is they know exactly like when someone's building software and, you know, theoretically anthropic and codex like for opening, I should have this as well. But they know every time that someone asks it to build something and it does it and then they give a follow up.

4:56They're like, oh, shoot. Every time someone asks for this, you know, for X, Y, Z, they typically ask these next five things. And that could either mean that when they first ask, they don't know exactly what they want or there's traditionally a bug that follows or there's a gotcha or something like that. but essentially they can fine-tune this model that they have to know exactly what they want to what people are basically the direction they're going before they even go there right they're like hey I want like a fitness app and like oh these are probably the five features they want so either they'll ask them right away or they'll just start building stuff in or they'll set up the architecture in a way that is ready to go now there's a couple cool things that they can do with these fine-tuned models one of them being that they can make them much more efficient and so they're able to cut the costs for their users, which of course is fantastic.

5:38But also, if someone's talking to their model, and they're like, Oh, you know what, our, you know, our base one model would be a lot better than sending it to Opus 4.8. All of a sudden, they're able to switch people over to the more efficient model, save the user money. And they're also keeping exclusive data, right? Because if they're not sending data over to OpenAI or Gemini or Anthropic, all of a sudden, they're not able to train with all of that data. So we're also seeing, we're seeing the same thing coming out of cursor. So a bunch of different players that are interacting with code are basically intercepting the code and the conversations people are having and knowing what people want.

6:16And they're being able to train their own fine tuned models with it. So base one, basically they're saying this is designed to replace Anthropics Opus for app generation workloads. And it's going to try to give them a really direct control over latency, cost and inference speed. And they're not going to have to rely on any third party APIs. And also, by the way, you know, this is a big cost for them because when you get a subscription to base 44, you pay them for tokens, but a huge chunk of your token bill just goes straight to Anthropic, right? It's like they're, they're giving a huge chunk, but if they can do it directly on their servers and if they have, um, you know, a cost efficient way to do that, they're actually able to make more money.

6:50So lovable that I've mentioned many times is of course their rival in Sweden, and they still use external LLMs. They have about$500 billion in AR, so they're about five times bigger, but they do have some really solid backing because Wix acquired them for about$80 million when they were only six months old with eight employees. Their parent company is now cutting about 20 % of their workforce. So Wix is cutting down, but evidently Base44 is accelerating. And my prediction here is that if Base44 really pulls this off, Base44 will actually end up being much bigger than Wix, which is pretty wild considering they bought it for$80 million.

7:25It's already doing more than$100 million in annual recurring revenue. So they were just, you know, blasted past even the acquisition price. This was an amazing deal for Wix. All right. The next thing we ought to talk about is that X has just launched a hosted MCP server. And if you know X, like they, when they, when Elon originally bought Twitter, he immediately shut down the API because I think OpenAI and a bunch of other people It just had like a full firehose blast of all of the API data and they were using it for training. Elon shut that down, realizing how valuable it was. We saw similar moves from Reddit and other players afterwards.

7:59But now that X is kind of launching this new MCP, they're going to let tools like Claude, Cursor, Grok, all tap into XAI's API and specifically your own account. So if you have your own account and your own permissions, you can let the MCP tap into that. I'm sure there's a ton of like market sentiment analysis, you know, great things that you can do with this. There is an API. So big firms, you know, like hedge funds that are trying to, you know, measure people's sentiment analysis on the FIFA World Cup to determine how many hot dogs are going to be sold and blah, blah, blah, blah, blah, and do all their like crazy, you know, hedge fund calculations.

8:35Those types of things, they already have an API, they're going to be using this. This is more for your average user. An MCP means that you can get all of XAI's data into your cloud that you're actively using. So for regular people, I personally am excited and would love to connect the MCP for when I talk about AI news here. I would love to get maybe like some sort of brief on every story I'm going to cover with the top tweets and the top comments on those tweets. So if my MCP could go and pull that, that'd be really useful. Traditionally trying to get like Claude to go and scrape X for good tweets is notoriously hard.

9:06I've had a hard time doing that and I basically just have to go manually scroll X myself. It would be fantastic if I could save some time and just, you know, surface the best things. Basically, I mean, what's kind of cool is if you have an MCP into something like X, and I know I'm getting off the rails here, but you could imagine the same thing with Reddit or any other social platform. You can fine tune your own algorithm, which I think threads and Mark Zuckerberg's trying to let people kind of do that with threads. But it's like in threads anyways, it's like show more posts about this specific topic for the next like four weeks.

9:34And then anyways, it's not like a permanent thing. So this would be cool. I would honestly almost remake my own X feed because I probably get I'm sure I interact with it, but I feel like I get like way too much politics in my thread that I really want to interact with in any given day. So, and it's probably my own fault, but it would be really cool to just intentionally be able to use this MCP on X and say, look, I just want AI news. I just want, you know, news from these, you know, 500 companies that I'm trying to follow in AI and get more of a, more of a streamlined stream of content around that.

10:04That'd be cool for me. Okay. XAI obviously isn't the, you know, the first big company to do an MCP. It feels like basically everyone has an MCP, even my own company, AI box.ai, we have an MCP. So, you know, traditionally people would go to our site, there's 80 different AI models, audio, text, video, and people would go chat with them in our playground. And you know, you pay like$9 a month, and you get access to every model. But a lot of people are like, look, all my workload is how all of my work is happening inside of Claude or inside of ChatGPT. So we created an MCP. And now you can get access to all of our different tools you can get, you know, like Claude can have image generation and audio generation, which it didn't have before.

10:39So MCPs are very valuable. And XAI is not the first one, GitHub, Slack, Stripe, Salesforce, all of them have official MCP endpoints. The API pricing, if you're going to do that with X is still going to be point zero one five per published post and about 20 cents per post with links that was set earlier this year, they're basically said they're trying to curb spam at scale. But the hosted server is showing some new capabilities, there's search, there's post retrieval, there's user lookup, there's trend analysis. All of that is cutting down a lot of integration time for people trying to work with X.

11:13So I'm going to be excited to test that out. Anthropic is shipping Claude Sonnet 5. It's going to be$2 per million input tokens, which is undercutting Opus 4.8. And it is about 63.2 % on agentic coding. So it's within about six points of Opus 4.8. You can only imagine, right, the next model, which is Fable 5. I mean, and they already dropped it, but it's going to be significantly better. It's interesting because the, you know, their headline model got pulled back. So now they just have all of their like, all of their their worst models, or, you know, they're more efficient models. They're less powerful models, they're starting to catch up with their main model, because they can't even, you know, update their main model to be its best, which is really interesting to me.

11:53So this model is going to basically become the default for all free and pro users starting on Tuesday. I think that is just showing that agent capable AI is now really important across basically every tier. This isn't going to be just a premium feature. And I do think that it's promotional pricing, the$2 for the inputs and the$10 for the outputs, that is going to be running through August 31st. And we see this from Anthropica a lot. I think if I go on my account right now, they're like, till the beginning of July, it's 2x credit usage. And I mean, basically, at this point, I'm just trying to max out my usage every single week, because I feel like you're getting a really great deal.

12:30And I don't know how long that will last. But that's going to increase from$2 to$3, which is still cheaper than Opus 4.8 and GPT 5.5 and Gemini 3.1 Pro. So it's still cheaper, but it's, uh, yeah, until August 31st, we got the$2 price. It's going to then go up to$3. Um, and you know, maybe we'll have new models coming out then. So who knows what happens in the future, but on knowledge work benchmarks, Sonnet five is slightly better than Opus 4.8, which is pretty crazy, right? Like Sonnet is the, you know, the, the lower quality model than Opus, but because they can't upgrade Opus, now the worse version is actually better than Opus 4.8.

13:08I think that's just kind of showing though the cost per task efficiency is now really important. It's more than more important than maybe the raw reasoning rank. And if you want to spend a ton of money, you can go use something like Cloud Code and turn on like, you know, their super code mode or whatever. It gives you, you know, waste more reasoning, but it uses a lot more tokens. A lot of early testers say that there are fewer mid-task stalls in some of the multi-step workflows. Zapier's two-part Salesforce automation completed where I think a bunch of previous models, they were trying to do that two-part Salesforce automation, and it was getting about halfway through and it abandoned it.

13:43So anyways, it seems like this is going to be much more efficient. Speaking of new models, they're not the only one. Google is shipping Nano Banana 2 Lite and also Gemini Omni Flash to developers. It is a text image model, It's going to generate in about four seconds and it costs 0.034 cents for a 1K image. Gemini Omni Flash, which is a video generation model, is about 10 cents per second of output, which honestly is not bad. These things can be insanely expensive. We literally had OpenAI that canceled Sora and was discontinuing it because it was just so expensive. So like we had the capability to do AI video.

14:20It's just really expensive. So the fact that Gemini is getting that down to 10 cents per second of output, not bad. You know, if you were doing a one hour long video, it's going to be pretty expensive. But most of the videos that these are generating are like five seconds long. So that's 50 cents a video, which I know actually sounds kind of expensive. If it's 10 second video, it's a dollar. If it's a, you know, 100 second video, it's$10. But, you know, that's that's the cost. OK, Nano Banana 2 Lite is going to be replacing the original Nano Banana. It's going to be the recommend as the recommended tier.

14:50And I think they're saying that it's like if you have a prompt that worked really well for the original Nano Banana, it's going to still work for Nano Banana 2 Lite, which is great. I hate it when the model changes and the prompts are kind of get off. They said there's going to be character consistency. They have really aggressive speed cost optimization that they've applied. So it's kind of interesting, right? It's like two, it's Nano Banana 2 Lite. Like this is supposed to be like the super fast, super light version. And it's they're trying to get this thing to still be comparable to Nano Banana, but way faster and way cheaper.

15:20And I'm really excited when we do these kind of optimizations on the image models. They also have OmniFlash, which is going to match VO3.1 fast pricing, while it's also adding conversational editing, multimodal input mixing, so they could do text, image, and video. And it's going to do text to action synchronization for any sort of like graphics that you have on the screen, which I think is phenomenal. So for videographers, this is actually really cool. Nano Banana 2 Lite is going to ship across nine of the different products that Google already has. So it's going to be inside of AI studio. It's going to be inside of the Gemini API.

15:53It's going to be inside of AI mode in search. So when you're on Google, just searching, it's going to be on the Gemini app, notebook LM, Google photos, stitch, Google flow. And then it's also inside of Google ads, right? They got to, got to make money from the ads off of all of this. I'm excited to see what Google comes out with. I'm excited for all the new models today, guys. Thank you so much for tuning into the podcast. Make sure if you want to check out the MCP for AI box that you go check out AI box. dot AI slash MCP. Basically, you're going to get all 80 of the top AI models right inside of Claude or Gemini or ChatGPT, whatever you work with the most.

16:24You can get all of the other models inside of there. So you can actually call them really, really useful. And as always, if you want to get all of these news stories that I talk about on the podcast straight into your inbox every day, go to AI chat daily.com. That's my website that goes along with this podcast. There's a big subscribe button you can hit in the top corner. And if you hit that, you can put in your email and I'll send you all of these stories and more in-depth articles where it breaks down all of the information. You can see all of that as well. Thank you so much for tuning into the podcast, guys.

16:51I hope you have a fantastic day and I'll catch you in the next episode.