# Claude Co-Work Now on Mobile and AI Updates

An episode of [AI Chat: AI News on Claude, ChatGPT, Gemini, Grok, Perplexity](https://invisiblenetwork.ai/podcast/1669990482), 7 Jul 2026, 12 min.

Audio: https://content.rss.com/episodes/365073/2970006/ai-chat-podcast/2026_07_07_20_48_18_5a949fd5-98b1-416f-8330-d3b5b5b7e16b.mp3

Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode! In this episode, we discuss the exciting launch of Claude Co-Work on mobile devices, enhancing accessibility for users. Additionally, we explore AI advancements from companies like DeepSeek, SK Hynix's massive IPO, and innovative tech like Solos’ camera-less smart glasses and Vercel’s strategy in AI model deployment. Chapters 00:00 Introduction to Claude Co-Work 00:10 DeepSeek's Inference Chips 00:21 SK Hynix's $28 Billion IPO 00:35 Launch of Solos Smart Glasses 00:45 Vercel's AI Model Strategy 00:59 Conclusion and Personal Update 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

[0:02] Welcome to the AI Chat Podcast. I'm your host,
[0:04] Jaden Schaefer. Every day I cover cutting -edge
[0:07] AI news and talk with the leaders behind it,
[0:09] breaking down what it means for your life and
[0:11] business. Anthropic is finally letting Claude
[0:17] Cowork go to your mobile device, your phone on
[0:20] the Claude app. You've probably noticed you can
[0:22] get Claude code over there, but Claude Cowork
[0:24] was missing. So they're going to be releasing
[0:26] that soon. DeepSeek is planning their own inference
[0:29] chips to basically cut NVIDIA and Huawei out
[0:32] of the deal for them. They're trying to get around
[0:34] them and they rely on them heavily. SK Hynix
[0:36] is planning a $28 billion US IPO or listing because
[0:41] the demand for their memory has made a 260 %
[0:45] stock jump for them. So this company is growing
[0:48] like crazy. Solos is launching AirGo A6. These
[0:52] are smart glasses. They're only 19 grams and
[0:55] essentially they're smart glasses with voice
[0:57] AI, but there is no camera in them. Vercel's
[1:01] Guillaume Rauch is betting on splitting AI models
[1:04] from agents. So he thinks that you're going to
[1:06] get more bang for your buck and more efficiency
[1:08] if you actually take the AI model away from the
[1:11] agent and have these as two separate things.
[1:13] In completely unrelated news, I got a text today
[1:15] from my city in North Carolina that said that
[1:17] we are in a drought and I'm only allowed to water
[1:19] my lawn if my house number is even every other
[1:22] day of the week. And then it has proceeded to
[1:25] rain all day long. I'm not sure if we were supposed
[1:26] to be saving water so that AI data centers don't
[1:29] run out. but I think we have enough rain now.
[1:30] The drought might be officially over. Okay, I'm
[1:32] super excited. Maybe the most excited I've been
[1:34] in a while. Anthropic has just allowed ClaudeCowork
[1:38] to go on your mobile device on the Claude app
[1:41] on your phone. Previously, it was only on your
[1:43] desktop computer. And it's actually pretty annoying
[1:45] because if you're doing something with ClaudeCode,
[1:47] you could remote control it from your phone,
[1:50] although it had some limitations. One of them
[1:51] being I found I had like all the permissions
[1:53] were set the same way from my phone. So I still
[1:56] had to keep approving things over and over again,
[1:58] like the olden days, which is kind of brutal.
[2:00] But I'm really excited. excited because now you
[2:03] can fire something off on your computer, leave
[2:05] and you can keep, you know, get iterating with
[2:08] Claude cowork on your phone. And I think it's
[2:11] actually going to be able to take control of
[2:12] your computer still and do stuff. So this is
[2:14] actually insanely useful for so many things.
[2:17] At the same time, Claude has also published a
[2:19] list of what people are using Claude cowork for.
[2:22] So in case you need any ideas, or you're curious,
[2:24] there's one graph in particular, they said that
[2:27] 33 % of people are using Claude cowork for business
[2:31] processes and operations. 16 % are using it for
[2:34] content creation and copywriting and only 8 .7%.
[2:37] So basically like 9 % are using it for software
[2:39] development. I actually use it for software development
[2:41] a lot. Um, but I kind of toggle between Claude
[2:44] cowork and Claude code. There's like pros and
[2:47] cons to both of them. I find if I'm doing something
[2:50] where I need to be like logging in on the browser
[2:53] to stuff and clicking around and grabbing files
[2:55] off my computer, I'm going to use Claude cowork.
[2:57] Um, if I'm doing something that's more in like,
[2:59] you know, just making pushes to my GitHub and
[3:01] stuff, I'm going to use Claude code. So anyways,
[3:04] there's pros and cons. And I will say the biggest
[3:06] pro with Claude code is that you just get way
[3:08] more mileage out of it. I burned through all
[3:09] my Claude co -work credits. I actually made a
[3:11] new Claude max account just because I didn't
[3:13] want to run out of all my fable credits, which
[3:15] shout out also Claude. Claude said that they're
[3:19] extending Fable 5, the use inside of the Claude
[3:23] app. If you have like a Claude Mac subscription,
[3:24] they're extending it to the 12th. They were supposed
[3:27] to end today on the 7th, but they're extending
[3:29] it to the 12th. So if you have Claude, you're
[3:31] going to get way more Fable 5. Now, the problem
[3:33] is if you're anything like me, you burnt all
[3:35] of your Fable 5 credits already. I made a new
[3:38] account yesterday and I have officially burnt
[3:40] through all of my weekly credits already. So
[3:43] that's definitely a problem. I'm trying to work
[3:46] on that and I have to wait. till tomorrow, Wednesday,
[3:49] before it reset. The next things that people
[3:51] are using Claude for is DevOps and infrastructure,
[3:54] 7%, research and intelligence, 6 .4%, data analysis
[3:58] and business intelligence, 5 .8%, document processing
[4:01] and extraction, 4%, sales and revenue operations,
[4:04] 4%, and then everything else is about 14%. So
[4:06] the biggest thing is business processes and operations,
[4:08] content creation, and copywriting. That is basically
[4:12] 50 % of what everybody is using it for. I know
[4:15] you'll be able to do all of that from your phone,
[4:17] which is a huge unlock. It's also on 24 -7. So
[4:20] you can actually give it a task on your phone
[4:21] and then go to bed and it's done in the morning,
[4:24] which is really nice. I think I've had issues
[4:25] in the past when working with a similar type
[4:27] of software from my phone where it like if you
[4:29] turn your phone screen off, then it stops running,
[4:31] which is bad. So anyways, it's always on. It's
[4:34] from your phone. This will be fantastic. It is
[4:36] launching today. It is in beta and it is for
[4:39] Max Plan subscribers. If you're paying $100 a
[4:41] month, it's also going to roll out to the pro
[4:44] users at $20 a month later on. CoWorks demo that
[4:47] they showed with this. This basically showed
[4:49] it pulling data from email, Slack, and a bunch
[4:52] of different meeting transcripts, and it was
[4:54] writing a document, drafting emails, and it was
[4:56] all doing that from your phone, which is really
[4:58] cool, but it was using your computer to, like
[5:01] your desktop, to do all those things, which is
[5:03] pretty sweet. DeepSeek, the popular Chinese LLM,
[5:06] is going to be designing their own chips to run
[5:08] inference. This is something that essentially
[5:11] they're doing to get away from relying on NVIDIA
[5:13] and, of course, their big competitor, Huawei,
[5:16] over in China, which currently dominates. It's
[5:18] about half of China's data center chip market.
[5:21] What I will say is that inference chips are cheaper
[5:24] and faster to iterate on than training chips.
[5:26] So that's kind of DeepSeq's something that's
[5:29] going to help their own profit margin on every
[5:31] response that their model is actually generating.
[5:34] Alibaba and Baidu are both kind of pushing for
[5:36] similar custom chips. OpeningEye and Broadcom
[5:39] have released their own inference chip, which
[5:41] is called Jalapeno. As far as the timeline goes,
[5:44] custom accelerators typically take about three
[5:46] to five years from design to production shipment.
[5:49] And DeepSeq, they haven't. said you know what
[5:51] any of their like foundry partners haven't said
[5:53] anything about their timeline they haven't said
[5:54] anything about the performance targets so while
[5:56] they are starting to develop this i don't think
[5:58] we're going to see it anytime very soon SK Hynix
[6:01] is going to be selling shares on the US stock
[6:03] market this week. They're raising about $28 billion
[6:06] and they're letting anyone in the United States
[6:08] get access to the world's second largest memory
[6:12] chip maker. This is the first time this is happening
[6:14] and if you remember my last episode, Samsung,
[6:16] SK Hynix, and one company out of Idaho are basically
[6:19] the only three people that can mass produce These
[6:22] custom memory chips for AI training at scale.
[6:24] This listing in particular is going to, I mean,
[6:27] they're really just trying to capitalize on the
[6:28] fact that in the last couple months, their stock
[6:32] has surged 260%. All of this is because of AI
[6:35] data centers, which desperately need high -speed
[6:38] memory chips to train and run AI models. SK Hynix
[6:41] is going to sell 17 .8 million shares. And a
[6:44] lot of people are calling this entire situation
[6:46] Ramageddon because when they came out with their
[6:49] first quarter revenue, they're up about 200%.
[6:51] 100 % year over year, just because all of the
[6:54] AI workloads have this massive shortage of RAM.
[6:57] And if you take into account Samsung and everything
[7:00] going on there, SK Hynix and Samsung are both
[7:02] spending over $550 billion to build new memory
[7:06] factories. This is one of the largest coordinated
[7:08] equipment investments in semiconductor history.
[7:10] So they're working together and there is a massive
[7:13] investment that is going in right now. Solos
[7:16] is launching Ergo A6 smart glasses. They're only
[7:20] 19. and they've decided to cut out the camera
[7:23] entirely, but they do have voice AI built into
[7:26] them. This is a really interesting play because
[7:28] we have the competitors, which are Meta. We have
[7:30] Apple working on smart glasses. We have Google
[7:32] that has smart glasses going on right now. But
[7:34] Meta and Google both have cameras inside of theirs.
[7:37] I'm not exactly sure what's going on with Apple
[7:39] because, you know, they're not going to come
[7:40] out for a few years and we'll see that in a couple
[7:43] of years because they're pretty slow with all
[7:44] of this. But it's kind of cool that in the meantime,
[7:47] while we're waiting for a lot of these other
[7:48] big players to come out with their smart. glasses,
[7:50] we're getting some new firms in here making some
[7:53] cool stuff. This is super light, they say that
[7:56] they're privacy first design, and it's going
[7:58] to solve one of the biggest problems that smart
[8:00] glasses have. I think a lot of people I've seen
[8:02] complaining on LinkedIn and other places about
[8:04] the cameras. And you know, if you're walking
[8:05] around and people have these matter Ray Bans,
[8:07] you never know if you're being recorded, it might
[8:09] make people uncomfortable if you're talking to
[8:10] them. And so or you know, if it's if you're wearing
[8:14] them, and you're worried that the cameras could
[8:15] be hacked while you're wearing them all day,
[8:17] right, there's a whole bunch of issues. So cutting
[8:18] out the cameras on these. It feels great to not
[8:20] just be walking around with a giant camera on
[8:23] your head all day long. And you get a lot of
[8:25] the benefit of the AI, which is, you know, you
[8:27] can voice to text, talk to it and get answers
[8:29] back on what you're seeing now or what you're
[8:32] talking about. Now, I think, of course, the biggest
[8:34] drawback here is if I'm looking at my car engine
[8:37] and I'm like, where is the specific cap I'm looking
[8:39] for? If I have the glasses on, it can actually
[8:41] let me know. And in Google's case, they're going
[8:43] to be able to like highlight things so you can
[8:45] actually see, which is really cool. But this
[8:47] won't have that capability. These glasses in
[8:49] particular have voice AI and their assistant
[8:52] can handle anything from real -time translation,
[8:55] which is amazing. I think this is one of the
[8:56] greatest use cases of AI in these glasses. Calendar
[8:59] reminders, calls, they also let you... hear the
[9:03] surroundings through speakers behind your ear
[9:05] so they can amplify different things for you,
[9:07] which is really useful. I just might have to
[9:08] get myself a pair. Vercel, which is the platform
[9:10] where millions of AI agents deploy code daily.
[9:14] And when I say that, like personally, I'm one
[9:16] of those people. Basically, I think all of my
[9:17] different softwares and apps are using Vercel
[9:21] to host all of the different sites that I have
[9:23] running right now. And the reason why is because
[9:26] it works incredibly well. And I think I think
[9:28] there's like a bunch of other good ones like
[9:30] Netlify probably does this. And I think Amazon
[9:31] has a version as well. But Vercel seems to be
[9:35] one that has some incredible integrations and
[9:37] APIs and MCPs, which connect to your cloud. And
[9:40] so you can have cloud or codecs or whatever you're
[9:43] using to code. integrated really well, so it's
[9:45] easy to deploy code. So Vercel is betting that
[9:48] the future belongs to companies that are going
[9:50] to keep AI models and agents separate. So you're
[9:53] not bundling these together by one lab, which
[9:55] is basically what we've seen everyone doing so
[9:56] far. Their CEO, Guillermo Rauch, says that customers
[9:59] are already mixing OpenAI, Anthropic Gemini,
[10:01] and open models like DeepSeek into infrastructure
[10:04] that treats models as a swappable part. And he
[10:07] says that companies that are facilitating that
[10:09] are the ones that are going to win. Vercel right
[10:11] now, they say that they process about 6 million
[10:13] deployments per day. That's roughly half triggered
[10:17] by coding agents and over 1 trillion tokens flowing
[10:20] through their AI gateway daily. So I mean, this
[10:23] really is a giant in the industry. They said
[10:25] that they shipped EBE, which is a framework for
[10:27] defining agents in plain language. And Vercel
[10:30] Sandbox lets you control what data internal agents
[10:33] can actually access, which I think is kind of
[10:35] the code leaking and training problem that a
[10:37] lot of people are working on. worried about.
[10:38] They have framed this whole new paradigm by saying
[10:41] that this is putting them in direct competition
[10:43] with AI labs like OpenAI, which are building
[10:46] their own hosting and deployment tools directly
[10:48] into their products. Now, I will say I would,
[10:51] you know, I don't think I would ever use a hosting
[10:54] platform exclusively owned by OpenAI or Claude
[10:57] or, you know, Anthropic. The same way I actually
[10:59] don't love Lovable or other people that like
[11:02] are pretty deeply integrated, even though Lovable
[11:04] uses Supabase as the database service. I still
[11:08] don't like how deeply integrated it was for me
[11:10] to get off of Lovable was a huge pain. So I think
[11:12] there's going to be always a place for companies
[11:15] like Vercel that are independent and separate
[11:18] from the actual labs because if you ever want
[11:20] to switch from OpenAI to Cloud or from Cloud
[11:22] to Codex, you want to have that kind of flexibility
[11:25] where it doesn't really matter what coding agent
[11:27] you're using. You've got to stay away from that
[11:29] vendor lock -in. And speaking of which, if you
[11:31] want to get access to over 80 different AI models
[11:34] all on the same account, I'd love for you to
[11:36] try out AIbox .ai. That's my own platform. I
[11:39] let you access everything from audio, image,
[11:41] video, even music generation models. There's
[11:44] over 80 of them. I mean, there's all the basic
[11:46] ones that you know, like Cloud and OpenAI and
[11:47] Gemini, tons of open source ones you haven't
[11:49] heard of or tried before, some really cool video
[11:52] and audio powered AI tools on there. So if you
[11:54] want to check out all of that for only $8 .99
[11:57] a month, go check out AIbox .ai. I'll leave a
[12:00] link in the description. Otherwise, I hope you
[12:02] all have a fantastic rest of your day.

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.