# Apple & Meta Bring Muse to Mac, Anthropic Opens Up & Deezer Filters AI Music

An episode of [AI Chat: AI News on Claude, ChatGPT, Gemini, Grok, Perplexity](https://invisiblenetwork.ai/podcast/1669990482), 16 Oct 2026, 14 min.

Audio: https://content.rss.com/episodes/365073/3238810/ai-chat-podcast/2026_10_16_00_36_28_f4b20322-d2b7-45e1-90d2-964d7721e2fe.mp3

Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode! In today’s AI Chat, Jaeden Schafer covers three stories shaping how AI is researched, used, and heard. Anthropic is giving METR researchers employee-level access to look under the hood. Apple and Meta are teaming up to bring the Muse AI model to Mac computers. And Deezer plans to let listeners filter out AI-generated music. We break down what these developments could mean for independent AI research, AI on personal computers, and listener choice in music streaming. Show Links • Get the top 80+ AI Models for $8.99 at AI Box: https://aibox.ai • How I Grow and Scale My Business with AI: https://www.skool.com/aihustle • Get the AI Chat Daily Newsletter: https://www.aichatdaily.com/newsletter

## Transcript

[0:00] Welcome to the podcast. According to an article
[0:02] from NBC News, cracks are beginning to show in
[0:05] the stock market as doubts grow about AI investment.
[0:09] So there's a lot of talk right now about the
[0:12] AI build out and the economic repercussions of
[0:15] it. And if this is going to pan out, if the ROI
[0:18] is there, as we have two mega IPOs, OpenAI and
[0:21] Anthropic coming up, we're going to dive into
[0:23] all of that. Also, we have Anthropic that is
[0:26] letting METR, famously Jacob Coxton, who was
[0:30] over at, you know, Anthropic for a couple of
[0:31] weeks. And then he did his big whistleblower
[0:33] thing where he went and said that it wasn't safe.
[0:35] And, you know, over Anthropic, they believe it's
[0:37] going to end humanity. He did his big PR stunt
[0:39] and went to all the big news organizations. And
[0:41] after that, he immediately was hired by METR
[0:44] and Anthropic. immediately handed them a multi
[0:46] -million dollar contract. And now Meta researchers
[0:50] are going to be able to look under the hood with
[0:53] employee access at Anthropic. So there's some
[0:56] interesting things happening there. According
[0:58] to CNBC, Apple and Meta are teaming up to make
[1:01] Muse AI models run on Mac computers. I actually
[1:04] got an email from Meta about this that I'm super
[1:07] excited about as far as they're, you know, they're
[1:09] still committed to open source AI models. Deezer
[1:13] is going to let you filter out AI generated music
[1:15] soon. And if I'm being honest, I actually think
[1:17] this is a phenomenal thing. It's funny for people
[1:19] that don't know, I actually make probably 20
[1:22] million streams a year on Spotify. And a big
[1:26] chunk of that is from AI generated content that
[1:29] I've produced over the years. And then a bunch
[1:31] of it is with my own voice. So it's a little
[1:33] known fact about me. I don't even publish who
[1:35] the artist is, but I make across a handful of
[1:38] artists that I've published. I think about 20
[1:40] million streams every single year. But I still
[1:42] think this is a phenomenal feature, and I'll
[1:44] tell you why. And it's not just Deezer, by the
[1:46] way. YouTube Music is also launching a new label
[1:48] that will show you if something's AI -generated
[1:51] music because there is so much of it on the platform
[1:53] right now. So I will tell you all the pros and
[1:55] cons to all of this. APAT is sending AI agents
[2:00] undercover to get scammed so they can gather
[2:03] fraud intelligence. This is super fascinating
[2:05] to me. And then we have a meta that is launching
[2:08] an open -weight AI model that can run on your
[2:10] Mac, sort of like what I was mentioning earlier.
[2:13] We also have Amazon's device chief, Panos Pene,
[2:17] who says there is a future where AI devices replace
[2:20] the need to use apps. I want to give my take
[2:22] on all of that because I don't think everything
[2:25] he said is accurate. So we're getting into all
[2:28] of that. Before we do, if you want to get access
[2:30] to over 100 different AI models for just one
[2:33] subscription, so Anthropic, OpenAI, Google Gemini,
[2:37] Grok, all of the image, audio, video, 11 labs,
[2:39] all of that in one platform for $8 .99 a month.
[2:42] Go check out AIbox .ai. That's my own startup.
[2:45] We've done some massive upgrades last week to
[2:48] it. As far as one of the big things is we have
[2:50] a groups feature where you can make a group.
[2:53] I make a group of four different AI models. So
[2:55] for me, it's Gemini, Claude, ChatGPT, and Grok.
[2:58] I ask one question and all of them at the same
[3:00] time are computing and they give you the same
[3:02] answer. all back at the same time. For me, this
[3:04] is super useful because some models are better
[3:07] at other things. And it's really cool when you
[3:08] do a text. But it's also really cool when you
[3:10] use something like image generation. Sometimes
[3:12] I like open eyes image generation, but sometimes
[3:14] it has this like brown look to it that you might
[3:17] be familiar with. And anyways, there's a bunch
[3:19] of different image generators, you can put all
[3:20] of them into a group, ask a question, and all
[3:23] of them will generate an image, you can pick
[3:24] your favorite one. So if you want to speed up
[3:26] how fast it and the time and the quality on the
[3:30] image, audio, text, or even video outputs that
[3:33] you get, go check out the new groups feature
[3:34] on AIbox .ai, along with tons of other incredible
[3:38] cool features. You can build tools with it. There's
[3:40] so much more inside of it, but go check it out,
[3:42] AIbox .ai. So the first thing I want to talk
[3:45] about is the... the market, the AI investment
[3:48] market. People are saying there's doubts about
[3:51] the ROI that's coming just one week, according
[3:53] to NBC News, just one week after hitting fresh
[3:56] all time highs. U .S. stocks are showing signs
[3:58] of strain. And so, you know, this is normal.
[4:01] The stock market goes up and down. But today
[4:03] we are pegging this on. AI. They said that investors
[4:06] are starting to worry about returns on trillions
[4:09] of dollars that have put into AI. This is interesting.
[4:12] I think after trading closed on Thursday, the
[4:15] S &P 500 was off 1 % from its week ago record.
[4:18] If you look at the tech heavy NASDAQ, that was
[4:20] down more than 3 % from its peak. Okay. The stock
[4:22] market goes up and down. We know this, but I
[4:24] do think this is an interesting moment. And I
[4:26] actually don't think this is unhealthy. Sometimes
[4:28] I think that we get... We get a lot of hype in
[4:31] AI and in technology where we can be very bullish
[4:34] because, oh my gosh, all these incredible things
[4:36] came out and the whole market rises with them
[4:38] because we look at the implications, right? Like
[4:40] OpenAI and Anthropic are surging if their usage
[4:42] metrics and if the revenue of those companies
[4:44] are surging, that means there's a lot of downstream
[4:46] winners as well. There's, you know, NVIDIA, who's
[4:49] going to be selling more GPUs. There's a lot
[4:51] of compute companies. There's a lot of data center
[4:52] companies. There's companies that build steel
[4:54] that is sent to data centers. There's companies,
[4:57] there's like so many different. people down the
[4:58] supply chain. So anyways, we have these big hype,
[5:00] these big hype moments that do lift the whole
[5:03] stock market. And I think it's okay every once
[5:05] in a while to have these resets where, you know,
[5:07] we're obviously growing if you look at the last
[5:08] year at an insane rate. But I think it's healthy
[5:11] to say like, look, maybe this thing isn't going
[5:13] to. replace everyone in the world next week.
[5:16] And I think we see some of these drawbacks in
[5:18] stock. So I don't think this is a huge cause
[5:21] for concern or alarm right now. But I do think
[5:23] it's healthy when we see some of these drawbacks.
[5:25] Okay, the big story that I was really excited
[5:27] about this week, or interested in, I guess you
[5:29] could say, is that Anthropic is letting METR
[5:32] researchers look under the hood. They're giving
[5:34] them employee level access to all of the latest
[5:38] models that are coming out. This is I think this
[5:40] is actually an interesting and exciting precedent.
[5:43] METR, I mean, I'm not, I don't have anything,
[5:46] any sort of special allegiance to that particular
[5:48] group. But I do think it's kind of cool when
[5:50] we have some of these open source or we have
[5:52] some of these AI model companies allowing independent
[5:55] researchers, independent security checkers to
[5:57] come in and look at this. And if I'm being honest,
[5:59] I don't think this is actually that much different
[6:01] than what a lot of these firms, OpenAI has been
[6:03] doing red teaming with third party vendors for
[6:05] a long time. And so anyways, it just feels like
[6:07] anthropic. did their big, you know, spiel where
[6:10] they were like, look, there's a 10 % chance that
[6:13] Claude is going to end all of humanity and it's
[6:15] super dangerous, yada, yada. And then this is
[6:17] kind of their response. But look, we're letting
[6:19] safety researchers come in and check our models
[6:21] before they go public. I don't think there's
[6:23] anything wrong with that. I don't like the way
[6:24] that they did the PR on all of that. At the end
[6:26] of the day, their model is really great. It is
[6:28] very capable. Letting third party people come
[6:30] and independently verify it. fantastic. And I
[6:33] love this way more than, you know, let's say
[6:35] enforcing some sort of regulatory body in the
[6:37] government where they might not even be competent
[6:39] enough to do some of this research. So I think
[6:42] this this approach is fantastic. We had a recent
[6:45] AI safety accord that a lot of these big firms
[6:47] signed and Anthropic was there. I think Mark
[6:49] Zuckerberg had a phenomenal point where he was
[6:52] like, look, as far as like getting regulated,
[6:54] like as companies, we're beholden to our shareholders,
[6:56] we're beholden to our board for making some sort
[6:59] of software service that is unsafe. We are liable
[7:02] for that. So third party auditors should absolutely
[7:06] come and look at our stuff and then they should
[7:07] report to the board. And if it's something that's
[7:09] unsafe, then the board decides to push the product
[7:12] forward anyways, then they're liable for it.
[7:14] And we as a company are liable for it. This is
[7:16] all just kind of like taking care of in regular
[7:18] laws and liability that we as companies take
[7:21] on. So I really love that framing from Zuckerberg
[7:24] on all of this. And I think that it basically
[7:27] will replace the need for heavy handed. government
[7:30] regulation that could slow down the entire field
[7:32] so anthropic letting them in i think this is
[7:35] great in a public statement they said in their
[7:38] blog post they said we want to be publicly accountable
[7:40] for the decisions we're making. We aim to close
[7:42] the gap between what evaluators understand about
[7:45] a frontier AI company systems and what the company
[7:47] itself understands. So under all of this, METR
[7:51] has Anthropic issued company laptops. They have
[7:54] office space. They have access to internal tools
[7:56] that they said are basically exactly what Anthropic's
[7:59] own risk assessment teams use. And then they
[8:02] said Anthropic employees are encouraged to share
[8:05] information with them just as they would with
[8:06] a fellow employee working on alignment risk.
[8:09] This is cool because I think this is the next
[8:11] level of transparency where they're like, look,
[8:13] there's a third party company, you know, evaluating
[8:15] us. And I think they're putting their money where
[8:17] their mouth is. They're like, our employees can
[8:19] go talk to them about concerns that they might
[8:21] have just like anyone else. And I think having
[8:23] this level of openness and knowing that there's
[8:25] this third party company inside your organization
[8:27] that's going to, you know, make a PR statement
[8:30] if there's a big issue. We'll just we'll encourage
[8:32] everyone to be very transparent about the safety
[8:35] risks and the guardrails and what they're doing
[8:37] to work on this. I don't think this is bad. You
[8:39] know, they've opted into doing this. And if this
[8:41] may gain some trust with users and with corporations
[8:44] that use their company, I think that's a phenomenal
[8:46] outcome. And so I think that that's that's the
[8:48] best direction. And again, I prefer this much
[8:50] more than heavy handed government bureaucracy
[8:52] determining a lot of these things to regulation.
[8:55] Apple and Meta are teaming up. This is kind of
[8:57] funny. If you know much about Apple and Meta,
[8:59] they've had a lot of beef. Tim Cook. Geesh, why
[9:02] do I keep wanting to call him Tim Apple, man?
[9:04] X has really warped my brain. Tim Cook and Mark
[9:06] Zuckerberg have had a lot of beef over the years.
[9:09] But and like very famously, you know, a lot of
[9:12] their privacy things, which stopped Facebook
[9:14] being allowed to track you. And then, you know,
[9:16] Facebook lost billions of dollars from their
[9:18] ad network and all sorts of stuff. And there
[9:20] is whatever, all sorts of legal campaigns around
[9:22] this. And probably I would say like. you know,
[9:25] like different campaigns where Meta was like
[9:27] telling small business owners to contact them
[9:29] and whatever, whatever. So, you know, voice their
[9:33] displeasure with them. So I think it's all interesting.
[9:35] What I will say is after all of that happens,
[9:38] it looks like they are coming together and working
[9:41] together again to help Meta be able to run their
[9:45] Muse AI model on Mac computers. So this is a
[9:48] collaboration between the two of them. We have
[9:49] a, you know, we have new CEO, new leadership
[9:52] over at Meta. apple and it seems like we're now
[9:54] starting to work together what's interesting
[9:56] with this particular collaboration is that it's
[9:59] going to be the meta spark which is honestly
[10:02] this is what's been running their muse model
[10:03] or sorry the muse spark and so i think that this
[10:07] is a this is a great model being able to run
[10:09] this on your own computer what they've said about
[10:10] this is that you're going to be actually able
[10:11] to run this on two different, I think, Mac Studio
[10:16] PCs, and they're going to both be able to run
[10:18] it. You can run MuseSpark code. You can analyze
[10:21] video. You can do agentic work. It's all going
[10:23] to run locally. It's not in the cloud. So two
[10:25] big perks is that there's more security, and
[10:28] it's also going to save you money. If you can
[10:29] run one of these models, Meta announced, which
[10:32] I kind of mentioned earlier, but Meta has a new
[10:35] open -weight AI model. that they've announced.
[10:37] That's going to be able to run directly on your
[10:39] MacBook. You don't have to pay Meta. And this
[10:41] is kind of crazy because it's at the time when
[10:43] Meta is really hitting the frontier. Their new
[10:46] Meta wonderful model is, you know, just as good
[10:49] as everything from Anthropic and from OpenAI,
[10:52] allegedly, according to a lot of different benchmarks.
[10:53] And with that, or, you know, you could say it's
[10:56] 10 % behind, but it's right up there. So this
[10:58] big model is right up there. It'd be like basically
[11:01] if OpenAI was, you know, they have their three
[11:03] different tiers of models or four different tiers
[11:05] of models. It'd be like if they're saying, look,
[11:07] our lowest tier of model, which you can pay us
[11:09] for, is just free to run it on your computer.
[11:11] So that's what Meta is doing. I'm really impressed
[11:13] with them. It might not even be the lowest tier
[11:15] for Anthropic or for OpenAI. It might be slightly
[11:17] higher than that. So I think this is very...
[11:21] I don't know, charitable might be a good way
[11:23] to say this. They're trying to win trust. They're
[11:25] trying to win users in the best way meta knows
[11:27] how, which is to give away free stuff. And honestly,
[11:29] I'm all here for it. I love the Muse tool. I
[11:31] get it to triage my emails every single day.
[11:34] It writes draft emails for me and it's just very
[11:37] useful and it's free. So I think that they've
[11:40] made some great products there. Deezer is going
[11:42] to let you filter out AI generated music. If
[11:45] I'm being honest, I actually really like this.
[11:46] Even as someone that makes thousands of dollars
[11:48] a month from AI generated music, I've made this
[11:50] over the years. And by the way, like AI generated
[11:52] music, I started making like probably like seven
[11:54] years ago. So AI music generators are not anything
[11:57] incredibly new. And it's also, there's a weird
[11:59] line to draw because We have two different things.
[12:02] We have something like Suno or Udio where you
[12:04] give it prompt and it just one shots an entire
[12:06] song with the vocalist and all that. But then
[12:08] we have a lot of tools that have existed for
[12:10] a long time inside of Premiere where you're like
[12:12] adding AI effects on top of your music or you're
[12:15] using quote unquote AI to generate like a piano
[12:18] or to. Or even just like the computers generating
[12:21] a piano or like... Anyway, so it's hard to draw
[12:24] the line on what's fully AI generated versus
[12:26] not. I tend to think just one -shotted songs,
[12:29] right? A song about being in the mountains with
[12:32] rustic sound to it and then Suno spits out a
[12:34] song. I don't tend to find those to be super
[12:36] high quality. And recently I've been listening
[12:38] to some music and the recommended songs that
[12:41] come after a song that I like and I clicked on
[12:43] is like AI generated and it's kind of annoying
[12:45] and I skip past them. So I appreciate the ability
[12:48] to filter them out if you want, to add a label
[12:51] onto them if you want. We've had disclosures
[12:53] on a lot of the streaming distribution platforms
[12:55] like Deezer or DistroKid or Lander. These are
[12:59] the places where you release your music. We've
[13:01] had disclosures there where you say if you use
[13:02] AI and all of that's in the metadata, so now
[13:05] they're going to be able to filter it. I don't
[13:06] think there's anything wrong with that. What
[13:08] I will say is the next step that... YouTube music
[13:10] is doing. And you know, these are YouTube music,
[13:13] I could imagine Apple Music and Spotify will
[13:15] do similar things soon in the future. But YouTube
[13:17] music is launching a new thing to label stuff
[13:19] is AI generated music. Again, there's like tricky
[13:22] line of like, how much of it was AI generated,
[13:25] how much of it wasn't, I've released a ton of
[13:27] content where it's my own voice, but like some
[13:29] of the instruments in the background are like
[13:30] AI generated. I'm not an incredible musician,
[13:34] I don't play a lot of instruments. And that's
[13:36] kind of what I've done. So anyways, I think that
[13:39] there's a lot of options. There's a lot of things
[13:40] you can do there. This is good. It's giving people
[13:42] more options. And I think that that is phenomenal.
[13:45] Guys, thank you so much for tuning into the podcast
[13:48] today. If you enjoyed this episode, make sure
[13:50] to leave a rating and review wherever you get
[13:52] your podcasts and go check out AIbox .ai. And
[13:55] I will see you all in the next episode.

Shows are listed from Apple Podcasts' charts and directory, and Podcast Index where it is used. Every show and episode belongs to its publisher and plays from the publisher's own feed; this site hosts none of them.