Modern Creator
Why Modern Creator?
Riley Brown · YouTube

Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode

Riley Brown runs Anthropic's freshly-released Opus 5 through a real workload, then puts the new real-time Claude Voice and OpenAI's Codex voice mode side by side to show what talking to an agent can actually do.

Posted
yesterday
Duration
Format
Demo
hype
Views
3.2K
188 likes
Part of the collectionThe Claude Opus 5 PlaybookEvery Opus 5 breakdown, synthesized into one page.
Read the playbook
Big Idea

The argument in one line.

Anthropic and OpenAI shipped competing voice-controlled agent modes within minutes of each other, and the deeper story is that both labs are racing to make cloud-run AI agents as easy to use as a phone call.

Who This Is For

Read if. Skip if.

READ IF YOU ARE…
  • You run a small team or solo business and want a concrete read on whether Opus 5 or the new voice modes are worth changing your workflow for.
  • You're trying to decide how much time to spend testing every new Claude or Codex release versus focusing on your own workflows.
  • You want to see a real voice-controlled agent editing documents, drafting email, and deploying a live web page end to end.
SKIP IF…
  • You're looking for a hands-on technical tutorial with exact prompts and setup steps — this is a reaction/demo video, not a how-to.
  • You need independent benchmarks of Opus 5 — the model claims here come from Anthropic's own comparison card and one early tester's opinion.
TL;DR

The full version, fast.

Anthropic released Claude Opus 5 alongside a real-time voice mode for its iOS app, arriving almost simultaneously with OpenAI's own Codex voice feature. Opus 5 beats Anthropic's prior flagship on most of its own benchmark categories at roughly half the price, though an early tester found it broke compatibility with existing prompt skills until he rebuilt them from scratch. The real demo is voice control: talking to Claude edits a live Notion doc and drafts email, while talking to Codex builds and deploys a landing page and pulls up Gmail drafts from a phone. Claude and Codex have also copied each other's in-app browsers, tab layouts, iOS simulators, and record-and-replay skill creation within weeks. The takeaway: stop debating which model is best and invest in repeatable workflows instead.

Free for members

Chat with this breakdown — free.

Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.

Create a free account →
Chapters

Where the time goes.

00:00 – 01:47

01 · Opus 5 drops, voice mode teased

Cold open: Anthropic just released Opus 5, positioned as cheaper than its own top model, with a bigger tease — a new real-time voice agent that beat OpenAI's own Codex voice feature to market by minutes.

01:47 – 03:04

02 · An early tester's verdict

Riley reads Dan Shipper's early-access post: Opus 5 argued with instructions, stopped early, and broke existing skills and plugins until Shipper rebuilt his workflows from scratch — calling it 'a poor man's Fable.'

03:04 – 04:43

03 · Testing Opus 5 in Claude Cowork

A single Cowork prompt (extract a company's site, fabricate numbers, build an investor deck) takes over 40 minutes but returns a slide deck Riley calls one of the best he's seen, closely matching the source brand.

04:43 – 10:28

04 · Claude Voice: editing docs and drafting email by talking

Using the new voice button in the Claude iOS app, Riley talks to Opus 5 to search and edit a live Notion doc, get an anonymized email summary, draft a reply, and build and deploy a live landing page — learning that voice mode can't read links aloud.

10:28 – 14:23

05 · OpenAI's Codex voice mode

On the Codex desktop app, voice control pulls up a Gmail draft, edits a collaborative Paper design board, and delegates a task to two parallel higher-reasoning model sessions to generate more design variations.

14:23 – 17:23

06 · Codex Remote: controlling a desktop from a phone

A phone app lets Riley fire one-off tasks at his Mac from anywhere — a day recap, a sponsorship-only email summary, and a list of every connected Mac skill (Gmail, Calendar, Drive, Notion, Slack, browser control, and more).

17:23 – 22:05

07 · Four ways Claude and Codex are copying each other

Riley lays out the pattern: both apps added in-app browsers, converged on the same chat/work/code three-tab structure, added iOS simulators for building Swift apps, and both now offer record-and-replay skill creation from a screen recording.

22:05 – 24:18

08 · The labs' real endgame

Citing a Greg Brockman tweet about cloud-run agents working even when a laptop is closed, Riley argues both labs are trying to make the power of coding agents like Codex accessible to millions of non-technical users through simple chat interfaces.

24:18 – 26:23

09 · Takeaway: stop picking models, build workflows

Riley's closing advice: don't obsess over which model or platform is best — more competitors (Cursor, Gemini, China's Z Code and Kimi Code) are coming regardless, so the real advantage is in your own repeatable workflows.

Atomic Insights

Lines worth screenshotting.

  • Opus 5 beats Claude's prior flagship in most benchmark categories except multidisciplinary reasoning, at roughly half the input and output token price.
  • An early Opus 5 tester found it broke backward compatibility with existing skills and plugins, arguing with instructions and stopping before work was finished.
  • Starting fresh without legacy prompt skills produced dramatically better Opus 5 results than reusing workflows built for older models.
  • Anthropic is positioning Opus 5 as an everyday model for enterprises and knowledge workers, not primarily as a coding model.
  • A voice agent can search a live Notion document, edit its sections, and add new content purely through spoken instructions.
  • Voice agents can't read links aloud — a spoken link has to be exited into text mode before it can be displayed.
  • OpenAI's Codex voice mode and Anthropic's Claude Voice both shipped within about ten minutes of each other.
  • A phone-based Codex Remote session can spin up new one-time tasks on a desktop agent without opening a full chat thread.
  • One weaker default model can delegate a design task to two additional higher-reasoning model sessions running in parallel on the same shared board.
  • Both Claude and Codex have independently added in-app browsers, a chat/work/code three-tab layout, iOS simulators, and record-and-replay skill creation within weeks of each other.
  • Record-and-replay lets a user demonstrate a workflow once by recording their screen, then have the agent turn that recording into a reusable named skill.
  • The stated end goal of both labs is to make cloud-run, always-on agents as easy to use as a normal chat app for millions of non-technical users.
  • Selecting the 'right' AI platform matters less than committing to one and building repeatable workflows, since new competitors (Cursor, Gemini, China's Z Code and Kimi Code) keep launching.
  • A voice-agent-built landing page went from a single spoken prompt to a live deployed URL in about a minute of build time.
Takeaway

Stop shopping for the perfect AI model or platform.

WHAT TO LEARN

Two frontier labs releasing near-identical voice-agent features within minutes of each other proves platform choice matters less than committing to workflows you actually repeat.

01Opus 5 drops, voice mode teased
  • Anthropic released Opus 5 and, within roughly ten minutes, OpenAI shipped a competing voice feature inside Codex — read close release timing as two labs actively racing each other, not coincidence.
  • Anthropic's own benchmark card shows Opus 5 ahead of its prior flagship in most categories except multidisciplinary reasoning, at about half the token cost — treat vendor benchmark cards as a starting signal, not proof.
02An early tester's verdict
  • An early Opus 5 reviewer found it broke backward compatibility with his existing prompt skills and plugins — assume any major model upgrade may require rebuilding custom workflows rather than reusing them unchanged.
  • The same reviewer got dramatically better results only after deleting his old skills and starting from scratch — sunk-cost workflows can actively hurt a new model's performance.
03Testing Opus 5 in Claude Cowork
  • A single long-running Cowork prompt (extract a brand, build an investor deck) took over 40 minutes to finish — factor multi-ten-minute wait times into agentic 'cowork' style tasks, not quick-chat expectations.
  • The deck's branding closely matched the source website's actual look and copy — feeding a real site into a generation prompt anchors the output far more than a generic style description would.
04Claude Voice: editing docs and drafting email by talking
  • A voice agent can search and directly edit a live document purely from spoken instructions — treat voice mode as a hands-free editing tool, not just a question-answering assistant.
  • Voice agents can't speak URLs aloud — any workflow that ends in a link needs a manual hand-off step into text mode.
  • A landing page went from spoken prompt to a live deployed URL in about a minute — voice-driven agents can complete an entire build-and-ship loop, not just draft content.
05OpenAI's Codex voice mode
  • Voice control on Codex can pull up a specific saved draft and hand a visual design task to a connected board tool — treat voice mode as a dispatcher to your other connected tools, not a closed system.
  • A single voice session delegated a design revision to two additional higher-reasoning model sessions running in parallel — use weaker default models for orchestration and reserve stronger models for sub-tasks that need real reasoning.
06Codex Remote: controlling a desktop from a phone
  • A phone can fire one-off tasks at a desktop agent without opening a persistent chat thread — this turns checking on your agent into a stateless dispatch action instead of a running conversation to track.
  • Asking an agent to list every tool or skill it can currently access is a fast way to audit its actual surface area before trusting it with a task.
07Four ways Claude and Codex are copying each other
  • When two competing AI platforms converge on the same feature set within weeks of each other, treat that convergence as the industry's current default UX, not a differentiator to chase.
  • Record-and-replay turns a single demonstrated screen action into a reusable named skill — the fastest way to teach an agent a repeatable task is to show it once, not write a spec.
08The labs' real endgame
  • Both major labs are explicitly optimizing for agents that keep running in the cloud after your laptop closes — plan workflows assuming your agent's work continues independent of your device being open.
  • The stated end goal is making the power of professional coding agents usable by non-technical people through ordinary chat — expect the skill gap between power users and casual users to keep shrinking.
09Takeaway: stop picking models, build workflows
  • Don't over-invest time in picking the best model or platform — more entrants (Cursor's knowledge-work tool, Gemini, China's Z Code and Kimi Code) are coming regardless, so today's pick won't stay the obvious best for long.
  • Aim to use AI to do more of what you already do, and do it better — if a workflow isn't measurably faster or higher-quality because of AI, it isn't paying for its token cost.
Glossary

Terms worth knowing.

Claude Cowork
A workspace mode inside the Claude app for longer-running knowledge-work tasks — documents, decks, research — rather than quick back-and-forth chat replies.
Codex Work
OpenAI's equivalent tab inside the Codex/ChatGPT app for multi-step tasks like building a site, reviewing code, or automating routine work.
Codex Remote
A phone-based interface that lets a user fire tasks at, and monitor, a Codex agent that's running on a separate desktop machine.
Record and Replay
A skill-creation feature where a user records their own on-screen actions once, and the agent converts that recording into a reusable named skill.
Paper
A collaborative visual design board tool the host connects an agent to, used to co-create and iterate on landing-page mockups in real time.
Firecrawl
A tool that extracts a website's content and branding so it can be reused elsewhere, such as feeding a company's site into a slide-deck generation prompt.
Resources

Things they pointed at.

01:47channelDan Shipper (Every)
03:23channelKieran Klaassen (@KieranKlaassen)
03:36toolFirecrawl
11:16toolPaper (design board tool)
24:53toolZ Code / GLM 5.2
25:02toolKimi K3 / Kimi Code
25:20toolDeepSeek
Quotables

Lines you could clip.

02:30
“it is a poor man's fable. It has many of Fable's personality quirks without its, like, true genius.”
sums up the entire early-access verdict in one line→ TikTok hook↗ Tweet quote
14:15
“This is like talking to Jarvis.”
names the emotional payoff of the whole voice-agent demo→ IG reel cold open↗ Tweet quote
23:56
“You're not gonna pick the perfect tool. It just does not exist.”
crystallizes the video's closing thesis in one sentence→ newsletter pull-quote↗ Tweet quote
The Script

Word for word.

Read-along

Don't just watch it. Burn it in.

See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.

metaphoranalogy
00:00Anthropic just released Opus five, and in many ways, it's better than Fable, and it's only half the cost. We're gonna talk about this today first. But what's even cooler than the model release in my opinion, Anthropic just launched a voice mode in their iOS app that can be used as a real time voice agent with Opus five, and you can connect to this voice agent with all of your existing apps and tools.
00:21But ten minutes before this voice launch, OpenAI released their voice feature inside Codecs that allows you to control Codecs with your voice in real time.
00:31And so I wanna talk about these updates today too. We're also gonna be covering all of the previous product releases by Anthropic and OpenAI over the past few weeks, And I'm gonna tell you how I plan to use all of these features inside my startup because that's what we discuss here on Agent Native.
00:47Okay. So five hours ago as of when I'm recording this video, Claude introduced Opus five, and they described it as a thoughtful and proactive model that comes close to frontier intelligence of Fable five at half the price.
01:02So what I find to be interesting is that Anthropic used this language right here. They said that their model, Opus five, is close to the frontier intelligence of Fable five, Yet they posted the model comparison card, and as you can see, it clearly is better in all but four categories.
01:19They said it's best in agentic coding. They also said it's best at knowledge work, which in my opinion is the most important part. They said it's best in novel problem solving, agentic search.
01:29It's not quite better at multidisciplinary reasoning.
01:33It's better at computer use, which is very useful, especially if it's half the cost because when you get into computer use with Fable, it gets extremely expensive. It's better at agentic coding, business workflows, not legal or health, but it's also better at biology.
01:47I can't really tell the difference between this model and Fable. It just feels kind of worse. But someone did actually get early access, and that is Dan Shipper.
01:55And he has an opinion that I trust a lot when it comes to AI models. The first thing that he said is that this model is very hard to love. He said that we spent the last week at every testing it across coding, writing, knowledge work, and our internal agent.
02:09It argues with instructions, stopped before work was finished, and generally didn't play well with our existing skills and plug ins like Compound Engineering. Our first reaction was, what have they done to my boy?
02:20Then we deleted our existing skills and started from scratch. This right here is why I think many business owners and managers and operators and people who do knowledge work are wasting so much time testing out the new models and the newest and latest models. If every time a new model comes out, you have to redo all of your skills, it is a complete waste of time.
02:42And he all ultimately went on to say that it is a poor man's fable. It has many of Fable's personality quirks without its, like, true genius. So now when you go to Claude desktop app, whether you use chat, whether you use co work, or if you use Claude code, you now have access to a new model, which is Opus five.
03:01You And can use Opus model five and test it for yourself. So Anthropic is actually positioning this Opus five model as its everyday model for enterprises, knowledge workers, and developers.
03:13And in their marketing material, they mentioned knowledge work much more than coding tasks. So for that, I think the best way to test it would probably be in the home tab using Claude Cowork. So to test it, I might do something like this.
03:25Make sure that you switch to Opus five. I'm gonna say please extract the chorus.com site.
03:31This is to get the brand. You can use Firecrawl to extract the brand of any website.
03:38Make a professional slide deck for investors to raise an a round. Make up fake numbers. It should be a final deck that will get the job done.
03:45Make it amazing. So now we're gonna be using co work, and we're gonna fire off this prompt, and there we go. Yeah.
03:51I'm not gonna lie. I regret doing this prompt. I entered this prompt forty one minutes ago, and it's still going.
03:58This is insane. Oh my god. Finally.
04:01Okay. So after forty plus minutes, it's done with the presentation. And, okay, I will say this is actually insane.
04:14This is genuinely one of the best PowerPoints I've ever seen.
04:19If you go to chorus.com, look at the landing page. Like, look at the branding here, and then look at the presentation.
04:29Like, this is very similar.
04:34And these slides are, like, pretty damn good, actually. It even pulled all of these different, like, logos, which I really like. It just took forever.
04:43And so this model's really good. I recommend trying it within Claude Cowork or within Claude Code, but I'm not gonna spend too much time on it because I wanna get to what I think is a much bigger update, and that is the new voice mode inside the Claude iOS app. So this was released yesterday.
05:00So as of yesterday, you can press this button right here, this white voice button. And as of right now, you can actually use it with Fable. I did not think you could do this.
05:11I actually don't think you could yesterday. You can see the model right here, and you can actually select. For the sake of this video, we can talk to to Opus.
05:18What's going on, my man?
05:22Not much. Just materializing into your video, apparently.
05:27Sounds like you're demoing something live. What are you building over there? Yeah.
05:31I'm actually building something. Um, there's
05:34a Notion document, um, that is called what is it called? Agent Native Weekly, um, and I created it on the twenty second.
05:42I'm filming it right now, and it talks about Opus five, um, and the Anthropic, uh, voice mode. What else are we talking about in this video?
05:51Can you search Notion and tell me? And you can see here it's searching Notion.
05:59Got it. Found it. So beyond Opus five and the voice modes, you've got a section on codex and Claude copying each other, whether Chinese models get banned, the Kimi k three launch, whether software is dead, and whether we're already in the Singularity.
06:15Yep. Thank you for that, actually. You know what?
06:18Are you able to edit this Notion document? Yeah. I can.
06:23I've got the update tool for Notion pages, so I can add sections, tweak wording, whatever you need.
06:30What did you want changed?
06:31Yeah. So, actually, um, you notice in the Anthropic OpenAI voice mode section, there's a tab in there called Claude Voice. It's empty right now.
06:40Can you please, um, do some research on the new Claude Voice and write out, like, seven or eight concise bullets on what it can do in use cases and have hyperlinks to any sources you may use there?
06:54Or it says Claude Voice. I just removed the bullets there. I want you to just add them in the Claude Voice tab.
07:03There you go. Do you see that? I just added this information here.
07:06Nice job, Claude.
07:11Ah, nice. Glad it landed. And you know what?
07:13At the bottom of this, you know, so we can go all the way down and we see what does this mean for business. Below this, I want you to add a new section in tabs of whatever you think I might wanna include in the video.
07:27Done. I added a section called your move this week with four tabs. Do this Monday, what to watch, my take, and outro.
07:37Wow. That's really cool. Moving on, can you please, like, um, tell me a little bit about my emails?
07:44Like, give me the latest update. Don't give any names of companies or people. Just tell me an update on what's going on in my email.
07:51Sure. Keeping it anonymous, the stuff that actually needs you. Two failed or overdue payments on subscription tools.
07:58So a card on file has probably expired. Beyond that, a warm follow-up from someone you met with today who wants coffee in New York next week. Can you actually draft up an email
08:10to emily at not a number dot com? Just create a draft for me so that when I'm at my computer, I can send it letting her know that this video is complete, that we're recording, and it's the one that matches the Notion doc.
08:24Done. It's sitting in your drafts. Subject line is agent native weekly recording done, and I included the link to the Notion outline so she can match them up.
08:35Hey. I actually want you to create a little landing page for me, Riley Brown, and this should just be our personal landing page. Can you please create it and then deploy it to Vercel for me?
08:44Oh, fun pivot. Love it. Quick question before I build.
08:49Yeah. Um, in terms of yeah. I want it to be like, I have the agent native podcast, the number one place to learn about AI agents.
08:57Please make it very cool.
09:01Love it. Let me pull up the deploy tool and build you something slick. Let me just build it.
09:07Give me a moment to put together a cool
09:10Okay. Good. No need to say anything.
09:15Okay. As you can see, it kind of pauses and freezes. And deploy See, for whatever reason, it does this, like, pausing thing, which is kind of annoying, but I think they're gonna patch that pretty soon.
09:27Animated glowing background. A badge that says Okay. So you it's live.
09:33That's pretty cool. Where's the link? The link is a bit of a mouthful to say aloud.
09:40So Okay. So this is a good learning moment. What you need to do whenever there's a link that you want, it actually can't do it in voice.
09:48You need to get it to type it out, so you can exit voice mode, and then you can say, link here, please. It just can't do it while on voice mode.
09:56So if you ever need a link from the voice agent, you have to exit voice mode and say, where's the link? Okay. Here's the public shareable link.
10:03Let's open this up. And there we go. It has this clearly vibe coded podcast, but the point is is you can vibe code in real time voice on Claude.
10:13Anthropic But wasn't the only one to release a cool new voice feature. OpenAI also released a voice feature directly into the Codex app or the ChatGPT Codex app that they combined, and you can now create a new voice chat. So I'm on the Codex desktop app here.
10:32Hey, buddy. I need you to look at my drafts in my email. I I created one a little bit ago, uh, sent to Emily.
10:38I need you to pull that up in the Codex browser, please. Sure. I'll find the Emily draft and pull it up for you.
10:53Alright. So thank you. You pulled it up for me.
10:58Thanks. I'm gonna go ahead and send that. Another thing that it can do is it can design directly on a Canva like board.
11:05For this, I use in a tool called paper, but check this out. Hey, buddy, old pal. I'm curious.
11:11Do you see my open paper board right now? I'm I'm using paper. You should have access to a paper MCP or tool.
11:19Do you see the board? It says something on there. Please tell me what it says.
11:25Checking that now.
11:30It says hello, monkey. I know. But you you cheated.
11:33You're supposed to use the MCP. Can you use it? Like, can you add please add a, uh, little landing page.
11:40Actually, you should have in a previous chat, you should have access to an agent native site that we created. I think it was yesterday. Can you please add that to the board here?
11:48I wanna redesign it. Okay. I'll pull that together.
11:51I'm using the paper integration and finding that agent native site now. I'm drafting a compact editorial landing page reference with a fall gray background,
12:02deep navy text, and the single electric blue accent. Wow.
12:07And you can see here the agent is editing it. This is pretty cool. I actually want you to create two more variations of this.
12:13Make them in different styles, please. Sure. I'll add two more.
12:17Working in paper again Actually, and because you're you're not that strong of a model, can you actually delegate this to Mhmm. A more powerful model? Like, spin up two more chats to do this.
12:27Keep the ones that you're go doing now. Don't remove these. I want you to spin up two more chats to create them with a better model.
12:33Let me see what I can run-in parallel.
12:37The magazine and terminal directions are on the board, and I'm keeping the original. I'm also spinning up two high reasoning tasks for more variations. Thank you.
12:47I've added a fourth direction as a high reasoning task, and I'm keeping the three existing boards intact. The two new high reasoning tasks are live, and all current corrections remain untouched. I don't see them on paper, though.
13:01Checking that now.
13:12Right. You're right. Those two tasks are still running, so their variations are not on the board yet.
13:18You should see them appear as separate artboards once they're finished. Artboards? It's gonna it's they're not gonna be on the same board?
13:25Checking that now. They're all going into the same gentle moon paperboard Nice. As separate artboards inside it.
13:35There it is. Very yellow. Okay.
13:38The yellow, um, I don't like the one that says your work has new superpowers. Like, that whole direction is very boring to me.
13:45Can you please just, like, adjust it to just make it, like, have more energy? Like, make some graphic on it. That's cool.
13:52Got it. I'll adjust that. I'm reworking the yellow terminal one into something bolder and more kinetic, not just a copy tweak.
14:02Oh, very nice. Very nice. That's a good direction.
14:07And, yeah, look at this. We are just freely creating here with our voice,
14:12and this is just so fun. This is like talking to Jarvis. Done.
14:15The dark version now sits right beside the yellow one with a night navigation feel and a glowing route graphic.
14:23Very cool. Thank you. Okay.
14:25So the video I did before this, I went much more in detail using this feature. I don't wanna spend too much time on it, but I do wanna show the other thing that OpenAI did is they actually allow you to do this from your phone. So I can click on remote here.
14:38So this remote is the way that I can use codecs from my phone, meaning I can control codecs that's running on my desktop from my phone. For example, before, I could select which computer, and this is the computer that I'm using, this Mac right here, and I can change this to be a normal chat, and I can say hi. So when I type in this prompt, it'll show up right here.
15:02You'll see that right now. Where is it?
15:06Oh, it's literally here. Yeah. So I can chat with it, and I can say, are you connected?
15:15And the point is is that these are, like, in perfect sync. But now there's a new way to use codecs from your phone, and that is with real time voice. So I can spin this up, and this isn't like a chat thread.
15:29All this agent can do is spin up new tasks. I'll show you exactly what I mean. I do want to, like, go through and clear these real quick.
15:38Let me show you how this works. So I'm gonna go ahead and use real time voice. It's connecting to my Mac.
15:43Hey, buddy. Are you there? Yeah.
15:45I'm right here. Hey. I I need you to just, like, go through all my chat sessions from the day, and I I need you to spin up a new task and make a summary of all the things that I talked about today, and then give me an include calendar in this.
15:57I just kinda want a summary of my day. So go through all my Codex chats and such. Sure.
16:01I'll take a Very cool. I can see it's working in Codex. Can you spin up a new chat which gives me a little summary in my emails for the day?
16:08Don't include any product stuff, just like sponsorship stuff. Okay. Narrowing it to sponsorship only.
16:15I've created a separate one time task for sponsorship brand deal and paid media emails only. It excludes product mail. Wow.
16:24Thank you. Do you see what tools I have or, like, skills you have access to on Mac right now? Checking that now.
16:30Yes. I can see your available Mac skills and connected tools. And your July 24 recap's ready too.
16:37You made progress across Notion research, audio setup, agent native design, and automations.
16:43Very cool. You know, you're good. You're good.
16:45One thing I wanna discuss yeah. I just tell me the skills that you can see on my Mac.
16:51Checking what's available on your Mac right now. I can help with paper, Gmail, calendar, drive, Notion, Slack, browser control, image and video work, social posting, PDF, Docs, and Sheets. So
17:05as you can see here, I can fully control my Mac. Right? So when I go home for the day, I leave my computer at work, but I can always access it through this remote control.
17:14And before, I would use this remote control, but now I can just speak to it. It is pretty insane that I can do real time voice and talk to my computer even if I'm across the country. So those are some really cool updates to both Anthropix tools and OpenAI's tools.
17:28And I wanna dive a little bit deeper into this because Claude and Codex are seemingly at war. Right? The the two front runners for super apps, whether it's Claude's desktop app or Codex or ChadGPT, their desktop app, they're adding features basically every few days, and it feels like they're just copying each other.
17:46And I actually wanna explain why this is a good thing for us. And by us, I mean a guy who wants to use AI agents effectively in my business. Right?
17:53That is the ultimate goal. And the first or the first of four ways that these two are copying each other are the browser updates.
18:02So I've talked about this a lot. I've made a video on this specifically. If you go to the Claude app, you can actually just spin up a browser.
18:10And this is only in the Claude code tab. Right? If we go into any chat here, we can open up the browser by clicking this button.
18:17And so now this is a full browser, and you can open up multiple browser tabs. Same thing on Codecs. Right?
18:23I've talked about this a lot. You can open up a browser, and we can go to any website, and we can open up many tabs.
18:29And we can ask our agent to open up browser tabs. And so that is what they're copying. Right?
18:34Claude added this feature very quickly after OpenAI added this feature, and that's just one of the features because these two have been copying each other in terms of the three tabs within the app.
18:47And so, basically, these Frontier Labs are trying to figure out how to stuff all of the things that you can do into one app, which is why I call it a super app. And it comes down to three things. Right?
18:57You have chat, Claude Cowork, and code. And on Codex, you have Codex, and then you can switch to ChatGPT mode, which has chat and work.
19:08So they're the same three things. Right? You at the end of the day, there's chat tasks, are like quick one off conversations, work tasks, which are create any type of document, do any type of research, reach out to people, control email, use Notion, etcetera, and then you have code tasks.
19:27And all of these companies, they're seemingly going back and forth trying to figure out the best placement for these three tabs, and they just seem to be copying each other, and it's changing like every single week. Also, the iOS simulator.
19:39So in the Claude app, you can actually open up the iOS simulator directly in the app now. If you ask it to create a Swift app, just ask it to open up the iOS simulator. In fact, this Hello World iOS app used a simulator, and you can actually see it right here, and it will actually open up the simulator.
19:58And when you tell it to run it, it will run it, and it will actually show up here in the simulator. In the Codex app, you can do the same thing by using this build iOS app skill.
20:08This will automatically create the same workflow where the iOS app will open in your browser. So this is an easy way to create iOS apps. And then the fourth and final way that these two are copying each other is the record and replay.
20:21I've mentioned this in multiple videos before, but in the ChatGPT app or Codex app, you can use this skill called record and replay, which allows you to just tell it to record your screen and then turn it into a skill. Well, Claude just added the same thing.
20:35If you go to a new chat, and I believe you have to do it on the home tab, and you can do it in like co work, you can hit this plus sign, and you can say record a skill, and you can start recording. And so now it's recording my screen, and notice here it has capturing and it shows zero steps. So if I were to want it to make it like uploading a YouTube video, I would say go to YouTube.
21:01I don't know where it's getting all these steps from, which is kind of weird. But I could go to create, upload video, and I could show that I can upload a video like this.
21:12And the point is is I can record my screen, and this context will immediately be uploaded into Claude. And the workflow, I think, is slightly better on the Claude desktop app.
21:22I can, like, record this here, and Claude will analyze it. And so I can say, please turn this into a skill called y t video upload.
21:34And this will allow me to take whatever steps that I do on my computer and turn it into a skill. Because remember, Claude and Codex can fully control your computer, and it's incredibly powerful, and it's really nice to be able to turn that into a skill, and it's way better than writing out the steps.
21:49Just show the agent exactly how to do it. And so what does this all mean? And I think it's important to zoom out a little bit.
21:56Obviously, I like to talk about the day to day updates of these super apps, but it's important to, like, take note of, like, what the ultimate goal is of these labs. And I think this Greg Brockman tweet really sums it up. It's like one of the best features of ChatGPT work is that it runs in the cloud, meaning that it works from mobile and when your laptop is closed.
22:16So what these companies are trying to do is they're trying to take all of the power of Claude Code or OpenClaw or these really powerful agents like Codex, and you're they're trying to give it to everyone. They're trying to get millions of people to use AI agents effectively. And the way that they're doing this is they're creating these co work or GPT work environments so that you can use these tools directly from ChatGPT.
22:39And so the one thing that I know for sure is it's going to get incredibly easy to use all of these AI agent tools. And all you'll need to know is how to set up skills or how to set up workflows, and they're gonna be accessible from any device.
22:52Whether you're using the desktop app or if you're using the web app, you can select work, and you can basically do any task that you would need to do on Codex. That is the ultimate goal. You know, they're trying to take all of the things that everyone loved about OpenClaw and make it as easy to use as ChatGPT.
23:10That is the ultimate goal. Because on the phone, you can do the same thing. There's this toggle.
23:14You can toggle it to work. This is literally codex. Like, it's codex light.
23:19Every time you use it, it spins up a little computer. And so if you need it to do a coding task, it will do a coding task. It can code anything.
23:26It can upload it to GitHub. It can deploy it to Vercel. You can have it control your emails.
23:32You can have it create a document. Anything you might wanna do, you can do inside the work tab. And not a lot of people realize this.
23:38And you can do this on Claude as well. This is not just about ChatGPT. My point is is these companies that have millions of users using their app, they want you to be able to use agents because agents use up a lot of tokens and they want that revenue.
23:52But that's not necessarily a bad thing because using agents can be incredibly useful for business. So my main takeaway is as follows.
24:01Both companies are building something similar, and you shouldn't worry about picking the right tool or the right model. Like, Opus came out today, and I do wanna talk about it because I think it's good to know that it came out, but don't don't spend too much time on selecting a tool and selecting a model. If you use Claude or OpenAI's ChatGPT app, just go super deep and focus on your workflows.
24:23You're not gonna pick the perfect tool. It just does not exist. And not to mention, Cursor, they're gonna be releasing their knowledge work version of their tool really quickly.
24:32Gemini, they're gonna make a competitor. Right?
24:35It's hard to tell if it's AI Studio or Anti Gravity, but you know that these are not just gonna be the only players. Not to mention, there's also some people in China doing some cool things. Right?
24:46Z code, the people who created GLM 5.2, an open source model. They have a z code, which is their own super app, and it's very similar to Codex. It's actually frighteningly close to Codex.
24:58Kimi k three also has Kimi code, which is a very similar idea. And so selecting the right platform is far less important than just focusing on your workflows as a business, trying to use these tools effectively. And if there's one thing I wanna leave with you today, it's just figure out how to make your life better with AI.
25:18And that's probably means taking the tasks that you do every day and doing them faster or doing more and doing them better. Use AI to do more and use AI to do better. And if you're not doing that, then there's no real point to using AI other than entertainment, and it's a very expensive hobby.
25:37These tokens are not cheap. And so that's what I'll leave you with is, like, although we talk about these models and we talk about all these updates, I like to put everything in context to make you realize that at the end of the day, you need to learn the skills. And that's why I make all the other videos that I do.
25:52I try to create ideas and show you examples of ways that I actually get value from these tools, because that's what you need to be doing.
25:59You need to figure out how to get value from these tools and only focus on the value, not the tools themselves. That's where all the value will come. And so that's what I wanna leave you with today.
26:10My name is Riley Brown. Thank you so much for watching. If you could like and subscribe, that would mean the world to me.
26:15I would love you forever. It helps me keep doing these videos. And, anyway, this was an update.
26:21I'll see you here for the next one.
The Hook

The bait, then the rug-pull.

Anthropic dropped Claude Opus 5 and, within ten minutes of it, OpenAI shipped its own real-time voice feature inside Codex — so the host runs both live: talking to Claude to edit a Notion doc and draft email, then talking to Codex to redesign a landing page and deploy it, before laying out exactly how the two labs keep copying each other's features.

Frameworks

Named ideas worth stealing.

18:16list

Four ways Claude and Codex are copying each other

  1. Browser updates
  2. The chat/work/code three-tab structure
  3. iOS simulator
  4. Record and Replay

A running tally of feature-parity moves between Anthropic's Claude apps and OpenAI's Codex/ChatGPT apps, presented as evidence both labs are converging on the same 'super app' shape.

Steal fordeciding which AI-platform features are now table stakes versus which are still genuine differentiators
CTA Breakdown

How they asked for the click.

VERBAL ASK
26:10subscribe
“If you could like and subscribe, that would mean the world to me. I would love you forever.”

Direct-to-camera sign-off ask at the very end, no on-screen graphic or link.

Storyboard

Visual structure at a glance.

open
hookopen00:00
early tester take
promiseearly tester take01:47
Claude voice demo
valueClaude voice demo04:43
Codex voice demo
valueCodex voice demo10:28
copying pattern breakdown
valuecopying pattern breakdown17:23
closing takeaway
ctaclosing takeaway24:18
Frame Gallery

Visual moments.

open00:00Frame at 00:21 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode00:21Frame at 00:49 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode00:49Frame at 01:02 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode01:02Frame at 01:19 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode01:19early tester take01:47
open00:00Frame at 00:21 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode00:21Frame at 00:49 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode00:49Frame at 01:02 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode01:02Frame at 01:19 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode01:19early tester take01:47Frame at 02:08 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode02:08Frame at 02:28 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode02:28Frame at 02:48 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode02:48Frame at 03:04 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode03:04Frame at 03:27 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode03:27Frame at 03:47 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode03:47Frame at 04:00 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode04:00Frame at 04:18 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode04:18Frame at 04:42 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode04:42Frame at 04:58 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode04:58Claude voice demo05:26Frame at 05:46 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode05:46Frame at 06:06 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode06:06Frame at 06:31 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode06:31Frame at 06:45 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode06:45Frame at 07:05 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode07:05Frame at 07:25 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode07:25Frame at 07:45 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode07:45Frame at 08:04 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode08:04Frame at 08:24 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode08:24Frame at 08:44 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode08:44Frame at 09:04 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode09:04Frame at 09:23 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode09:23Frame at 09:43 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode09:43Frame at 10:03 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode10:03Codex voice demo10:23Frame at 10:51 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode10:51Frame at 11:00 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode11:00Frame at 11:27 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode11:27Frame at 11:42 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode11:42Frame at 12:02 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode12:02Frame at 12:22 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode12:22Frame at 12:41 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode12:41Frame at 13:04 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode13:04Frame at 13:21 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode13:21Frame at 13:36 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode13:36Frame at 13:52 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode13:52Frame at 14:25 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode14:25Frame at 14:40 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode14:40Frame at 15:00 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode15:00Frame at 15:20 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode15:20Frame at 15:39 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode15:39Frame at 15:59 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode15:59Frame at 16:19 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode16:19Frame at 16:39 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode16:39Frame at 16:59 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode16:59Frame at 17:18 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode17:18copying pattern breakdown17:38Frame at 18:06 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode18:06Frame at 18:08 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode18:08Frame at 18:29 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode18:29Frame at 18:57 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode18:57Frame at 19:10 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode19:10Frame at 19:30 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode19:30Frame at 19:57 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode19:57Frame at 20:16 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode20:16Frame at 20:36 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode20:36Frame at 20:56 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode20:56Frame at 21:16 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode21:16Frame at 21:36 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode21:36Frame at 21:55 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode21:55Frame at 22:15 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode22:15Frame at 22:35 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode22:35Frame at 22:55 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode22:55Frame at 23:15 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode23:15Frame at 23:34 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode23:34Frame at 23:54 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode23:54closing takeaway24:14Frame at 24:34 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode24:34Frame at 24:53 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode24:53Frame at 25:13 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode25:13Frame at 25:33 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode25:33Frame at 25:58 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode25:58Frame at 26:13 from Opus 5 and Claude Voice land minutes before OpenAI's own Codex voice mode26:13
Watch next

More from this channel + related breakdowns.

47:51
Riley Brown · Interview

OpenAI Merges ChatGPT and Codex

Riley Brown and Ras Mic dig into GPT-5.6, Codex's background computer-use, and why self-scoring agent loops are turning coding tools into a general operating system.

July 12th
Chat about this