Anthropic Opens Bio "Wet Lab" in SF and Meta's Muse App

18 Sep 2026 · 13 min · 5 chapters

Ask about this episode

Ask anything about it. ChatGPT or Claude reads this page and answers with the times it was said.

Connect VO and ask about every podcast you hear, including the moments you saved. Add to ChatGPT · Add to Claude

In short

Anthropic’s plan to open a Bay Area “wet lab” where Claude directs lab robots with limited human oversight; Meta’s new “Muse for Mac” desktop app; Meta’s Muse-linked background with the acquired Manus technology; FAA contract for AI air-traffic conflict prediction.

Guest backgrounds

No guests mentioned; episode is presented as a solo host update.

Key claims

Anthropic aims to accelerate preclinical research for “undruggable” rare diseases while avoiding direct competition with pharma; Claude-controlled wet lab raises biosecurity concerns. Meta’s Muse for Mac can access files/apps/calendar/notes/messages and control the computer. Manus was unwound by Beijing after Meta’s $2B acquisition, and the host suggests Meta likely cloned Manus capabilities into Muse. FAA’s SMART system will flag flight conflicts using schedules, weather, capacity, and constraints; controller tools may help with persistent staffing shortages.

Notable examples

Polymarket/Reuters report on Anthropic wet lab; Chamath’s “I do not recommend this” reaction; Zuckerberg tweet about Muse for Mac; FAA $875M 12-year SMART contract starting near Washington DC.

Written by AI. May contain mistakes. Listen to the episode to check what was said.

Chapters

Tap a time to open that second in VO

AI Developments and News

0:30 to 0:58

Discussion of recent developments in AI, including Mammoth's valuation and the FAA's AI air traffic control system.

“We also have the fact that Mammoth is looking at a four billion dollar valuation.”

Anthropic's Wet Lab Announcement

1:29 to 3:56

Exploration of Anthropic's new wet lab initiative and its implications for AI in biology.

“And as always, if you haven't already, I would love for you to check out the, or I'd love for you to leave a review to the show.”

Meta's Muse App Launch and Features

3:56 to 7:52

Discussion of Meta's new Muse app, its capabilities, and comparisons to other AI tools.

“trying to compete with pharmaceutical companies who pay them, you know, millions of dollars, I would assume hundreds of millions of dollars of revenue comes from pharmaceutical companies.”

Manus Acquisition and Meta's Strategy

7:52 to 10:50

Analysis of Meta's acquisition of Manus and how it impacts their AI strategy.

“You're seeing and we're seeing basically all the big players.”

FAA's AI Air Traffic Control System

10:50 to 12:48

Details about the FAA's new AI air traffic control system and its significance.

“And the system is going to deploy there first.”
Hear the part that matters, and keep it.Open this episode in VO. Double tap your headphones to save a moment as you listen.
Get VO free

Transcript

Automatic transcript. May contain errors.

0:28We made it to the weekend. they have a Mac app and I'm actually super stoked about that. We also have the fact that Mammoth is looking at a four billion dollar valuation. They're doing a 500 million dollar raise. This is after Meta bought them and Beijing said just kidding and meant basically like turn the company back and unwind the acquisition. We also have the fact that the FAA is tapping airspace intelligence for an 875 million dollar AI air traffic control system. Lots of comments on this right now. And if you haven't tried it already, I would love for you to go check out AI box.ai, you get access to over 90 different AI models.

1:05And specifically what I think a lot of people are using it for right now is if you already have access to something like chat GPT work, or Claude co-work, you know, Claude co-work doesn't generate images, doesn't generate video doesn't generate audio, you can use our MCP connector, which basically allows any of those tools, you connect them to AI box, and you can generate images straight inside of Claude and do a lot of stuff that way. So if you're interested, it. I'll leave a link in the description. And as always, if you haven't already, I would love for you to check out the, or I'd love for you to leave a review to the show.

1:36If you enjoy the show, if it helps you at all, if it's something that you like specifically, I've started for the last two weeks doing video on Apple podcasts. So if you listen on Apple podcasts, if you've ever watched the videos, I screen share all the news stories and product demos and, you know, tool demos that I'm talking about. So if you enjoy that, let me know because it is way more work to do video. It's way more work to be doing product demos the whole time the podcast is going. So if that's something that's enjoyable to you, please leave a review on the podcast. It would help me to know if I should continue doing this.

2:08But otherwise, let's get into the show today. So Anthropic right now, they announced and they kind of announced this a while back, but they're getting more and more into they want to do bio labs. And specifically, there's a tweet that went viral from Polymarket. It said breaking Anthropic has quietly set up wet lab in the Bay Area as it pushes Claude to conduct real world biology experiments. This is according to routers. There's a lot of people commenting on it. But one of the people was Chamath Palihaapatiya from the All In Podcast. And he said the group behind such hits as We're All Going to Die and Regulate Me Now are building a wet lab in San Francisco.

2:42I do not recommend this. This is kind of the response that I think a lot of people have been getting. And honestly, I mean, it's not a shocker why this is happening. You have anthropic out there saying, look, like AI is going to end humanity. And one of the reasons is they're like, you know, biological warfare. And, you know, they're like AI and drones and AI robotics. And they're like, there's all these different ways that AI could end humanity. And one of them is like biological warfare. And by the way, we're setting up our own wet lab and getting it to be completely controlled by Claude. So as you can imagine, a lot of people are sort of freaked out about this and don't think this is a great idea.

3:17They basically want Claude to direct all of the laboratory robots. They want there to be very limited human intervention. And they said that they believe so the reason why they're doing this is they believe this could accelerate research into a lot of different rare and what they're calling undruggable diseases. They think that their focus on preclinical research rather than clinical trials is going to avoid direct competition with all of the big pharmaceutical companies that are paying them, you know, millions of dollars for for their technology. So it's Interesting. I've never really seen anthropics shy away from competing head on, whether that's with lawyers or with doctors or like with a lot of different areas.

3:55But it appears they are not trying to compete with pharmaceutical companies who pay them, you know, millions of dollars, I would assume hundreds of millions of dollars of revenue comes from pharmaceutical companies. So they're like, look, we're gonna we're gonna enter the space, but not in a way that competes too harshly. So interesting right there. Okay, we had a tweet earlier today from Mark Zuckerberg, where he essentially said Muse for Mac is out today. It works across apps, files, calendar, notes, and messages on your computer. You control what can access. The team is shipping fast. Download it here.

4:26This is something I was actually talking about the other day because I tried out Muse for the first time. And I'm going to be honest, I have basically never used any meta AI tools for the last two years. I mean, I've tried them whenever they drop, but something like meta AI built into Messenger was just always so bad. And their open source models were cool. And I applauded them for doing the llama and for doing the open source models because I you know think that that's a really good contribution to technology and developers and everything but as far as like a product that I actually use it was never the best so I don't really want to use it if it's not cutting edge that I truly believe has changed with their latest model that they've released and with Muse and for those that don't know Muse is basically Claude Cowork or ChatGPT work why am I talking about it so much why do I care guys this is free to use and I think that they might be working on some like payment tiers or something.

5:13But right now, like imagine getting chat GPT work, which by the way, I used up I've used up all of my chat GPT work credits for the week. And I'm now having to wait and like five more days for it to reset. And if you're like, wow, you just used a lot. No, actually, I don't really know how it all works. And I appreciate opening eye giving us resets. I was using the Astra model not on the highest tier. Because if you ask one question on the highest tier of the Astra model, it will use your entire week's usage in one question that runs for two minutes learned that the hard way last week but this week I had a big project and I did want you know a really good model and Astra is an incredible model so I didn't have it turned up all the way but I did have it on you know like high or something maybe it was like in the middle of their five level slider and on that level with a very important project that needed that smart of compute I used up I guess they gave me three resets.

6:04So when you have OpenAI, if you use up all of your tokens, I don't know who assigns the resets or when they assign the resets or why. Maybe it's if they have a glitch or I don't really understand how it works. But I guess I found in my settings that a three resets, meaning I used up all my tokens and I could click reset and it reset me to be full for the week. I did that three times. So I used up my first week and then I used up a three reset. So basically I used up an entire month's worth for in about a week. So one week I used an entire month's worth in like three or four days, I think. Now, I didn't really conceptualize what these resets were worth until once you run out now, it asks you, hey, do you want to reset it?

6:38If you click yes, it's$80 for a one week reset now. So even though it's$200 for the whole thing, if you want to reset it for the week, because you use it all up. Anyways, lucky for me, I was at a Claude subscription, and I just go to Claude and I use it's like Opus 5 or something. It's kind of, it feels dumber, but it's not as good. It asks me way more questions. It doesn't understand the project as well, whatever. I'm not going to die, right? So I just have the$100 cloud tier as my backup right now. But the reason why I'm talking about this on a Muse episode is because this is actually an incredible solution that as soon as I run out of my Anthropic tokens, I'm then going to be going down to the free tier of Muse.

7:15And if it's good enough, there's a ton of people saying they're like, look, I'm just canceling my opening eye thing. And I'm not sure if it's as good as Astra. I'm going to have to really put it through the ringer. I've done a bunch of different tasks with it already. And I have tested the the the app on my computer and I am liking it. One of the big things that I do like about it, by the way, is that it can control your computer, it can access all your files. I mean, this is what we love about OpenAI and Anthropics tools. And I don't know why Google isn't competing in this space. I don't know why Microsoft isn't competing in this space.

7:44But I guess I'm meta is the first one we do have Grokbot as well. So I guess there's four of them now that are out there. But I love Muse. It's awesome. This is the first time you'll ever hear me say that about a meta product basically i guess the ray-bans are cool i don't have them but i i like the vibe there so i think meta is doing some good things i'm gonna give them some big kudos i'm also gonna give them kudos for the fact that mark zuckerberg is probably not him it's obviously his team yada yada whatever but i do appreciate even though he owns threads the direct competitor to x he is posting on x which is a platform i get all my tech news from all my ai news from and the reason i found out about the new app the new desktop app for muse was because zuckerberg tweeted about it and somehow that showed up in my feed.

8:23So I think it's smart. You're seeing and we're seeing basically all the big players. We're seeing Jensen Huang from NVIDIA who never really was doing this before. Every time an update is coming out of NVIDIA, it's the CEO that's doing the tweet. I mean, I think this just makes perfect sense. If your company has a big update, the CEO, the most notable person at the organization should be putting out the press release and the tweet about it. It shouldn't come from a random, you know, at meta account that probably not very many people follow because it might just pump out corporate spam as well. So I really love that about it.

8:51This is how I found out about it. I've been trying it. It's able to access your computer. You can give it access to whatever you want or don't give it access to wherever you want. Same as everywhere else. So all in all, I am pretty, pretty excited about this. Okay. Let's talk about the other like a meta related news, which is just that they had Manus, which by the way, I think is such a fascinating concept. They acquired a company, which was a Chinese company that it moved to Singapore, then it was acquired. And then Beijing, I guess, had the CEO and the CTO still in China, even though the rest of the company moved to Singapore.

9:22They didn't let them leave China, so they're holding them hostage. That's how they got this whole deal, this$2 billion Manus deal unwound. What's interesting, though, is that when Meta gave Manus back, I mean, they'd already integrated it into the platform. I had friends that were working on like Meta ads at agencies and stuff. They were using Manus. They said it was fairly decent. I have no doubt in my mind that Meta, I mean, they're smart. They're probably working on this themselves. I have no doubt in my mind that they took like a lot of technology, strategy, technology skills. Like they had this Manus inside of their organization for a couple months.

9:54So I don't doubt that they basically like cloned Manus. They took everything they could out of it. I mean, I wouldn't call it IP theft because I mean, this is China doing it to themselves. But, you know, they owned the company so they could. No doubt that they built this all into what they did with Muse. And as soon as Manus was gone, okay, all of a sudden we have Muse and it's this incredible tool. So I mean, wouldn't be shocked if there's a big element of that. now is Manus down and out oh my gosh they lost their two billion dollar acquisition no seems like they're doing just fine they're looking at doing a four billion dollar valuation that's double what it was before it's going to have a 500 million dollar raise my question though is like who's using it is this just big in China is this everyone I mean Manus went super viral and a lot of people were using it but to me using a piece of technology that could get acquired by American company and then forcibly unwound by the Chinese government like really doesn't give me a lot of confidence that this is a tool that I personally would want to use a lot the privacy the security there's just a lot of red flags on that so I personally just would avoid that tool if it was me okay let's talk about what's going on with the FAA they are tapping airspace intelligence for a 875 million dollar AI air traffic control system this is a 12 year smart contract and it's going to deploy a cloud-based AI to predict flight conflicts they're starting this in Washington DC in that kind of area.

11:16And the system is going to deploy there first. And once that happens, SMART, S-M-A-R-T, I mean, that's an acronym. It's going to basically take in all the airline schedules, all of the weather, all the airport capacity, all the airspace conditions, all the operational constraints, and it's going to flag any conflicts it has before they occur. This is cool. I mean, this is what air traffic control is doing. It seems like we're just giving it more data and helping it make decisions. I'm assuming the air traffic control is going to be very involved. It's not like the robots are going to be running this and we're gonna have like airplane crashes because no one's overseeing anything.

11:50But I think this is probably more useful tools. This contract is also landing alongside a separate FAA hiring push that's looking at closing a really big persistent controller shortage, which I think brings up a great point whenever we have this kind of technology and everyone's like, oh my gosh, another job AI is taking. I mean, every job AI takes, it makes a million more. We had a big you know, a big press release. AI has created a million jobs so far, although people dispute that and they're like, well, that's because, you know, whatever. There's all the arguments about that. I mean, undoubtedly, the jobs economy is shifting.

12:23I think there's more opportunity than ever before. Whether people take up the opportunities is up to them. But it seems like there's a lot of new jobs being created. And in this particular case, we have literally had a massive problem with air traffic controllers having a massive shortage. It's really hard to hire them. It's very specialized, but it's very important. And so if we can give them more tools to help automate more things and ease that issue, I am thrilled. And I think this is a really incredible use case. So really excited about that. Guys, thank you so much for tuning into the podcast today.

12:50If you learned anything new, I would really appreciate a review on the show. Specifically, let me know if you enjoyed the videos. If you don't watch the videos at all, just leave a comment on Spotify or on Apple. I read them. I appreciate them. It helps evolve the show. Again, if you have any other questions, comments, changes to the show, anything you want to hear more of, leave it, but leave it with a five-star review because I don't, I don't like the three-star reviews where you're like, can you change XYZ thing, three-star review? I'm like, come on, give me a five-star review and I'll be incentivized to change it.

13:21So, you know, if you guys have any comments or feedback, I love them all, but I do appreciate that you listening and I hope you guys all have a fantastic rest of your day.

From the publisher
In this episode, we explore the exciting launch of a new lab in San Francisco, discussing its goals and potential impact on local innovation. Additionally, we cover Meta's introduction of a desktop app for Muse, looking at its features and how it could change user interaction with digital creativity.


Show Links See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

More from AI Agents: Manus, Muse, Claude Cowork, Copilot, ChatGPT, Grokbot

All 233 episodes
Anthropic Opens Bio "Wet Lab" in SF and Meta's Muse AppAI Agents: Manus, Muse, Claude Cowork, Copilot, ChatGPT, Grokbot · 13 min
Listen in VO