Riley Brown

Codex just updated their in app browser... You can now open multiple tabs, import cookies and passwords and more. And then 2 days later Claude Code released a similar update... This completely shifts the way we use computers... Here's everything you need to know. 00:00 Intro 00:36 Codex Browser Example 02:17 Updated Claude Code Browser 03:43 Using Codex in App Browser 06:30 4 Ways to Use The Browser in Codex 10:22 The AI Operating System 13:31 6 Useful Browser Workflows

Niche personal tool that I built with 5.6 Sol this morning. An exact clone of Excaladraw but it lets me create animated presentations that i can use in my videos... And i love adding logos to my excaladraw so i added built in image search and background removal.

hey @ChatGPTapp why appshots not working after update????

Sunday from 2-9PM is my favorite time to work. The urgency from the week ahead, while also having little to no notifications. (If I didn’t love what I do none of this would be true)

They should just do this every week for the next year. Free impressions.

For most work Codex just needs to get faster. 5.6 sol at 5x speed for research, ideation, emails, content/ads and general operation tasks and it’s game over. If you’re making a huge company decision maybe a Mythos+ level model would be needed but for 99% of tasks the cerebras hosted 5.6 sol in the codex app, with their new browser updates it may be the only app I need on my computer.

RT The Agenteer Influencers are cloning themselves as agents and selling them. This will be a bigger market than influencers selling apps.

The problem with slack for me is every time i use it get lost.

Make something agents want.

Google was the first search engine to win. ChatGPT, Claude, Grok, Gemini are now action engines.

Interesting changes in the way the claude desktop app is organized. This makes alot more sense. Really good decision.

The super-app wars continue. The browser is one of the most important parts.

Hey @riversidedotfm how do I export all my files for a podcast... when i press exprot it takes me to your new little editor. How do i export all files xml?

💯

Right now existing apps are making plugins for codex Claude and cursor etc. But I think some of the best plugins will be ONLY plugins.

If you’re struggling to build a tool people want I highly recommend building software that is “agent native”. 900M people use ChatGPT weekly and now OpenAI is merging codex and ChatGPT. Millions of people are going to be using agents, who previously were just talking to an ai that could search the internet. Agents don’t just retrieve information. Agents do things on the users behalf. What if you build something where agents do things. Think moltbook but think bigger. Think smaller. Think more niche. Think more useful. When you go down this rabbit hole you’ll realize there’s so much opportunity here.

Underrated Codex skill that I use every day is my /email-draft skill. It can draft all my emails and send me links to the draft so i can refine them and ship them off. Right click open in browser, you never leave Codex. Here's the skill you can copy. https://rileybrown.xyz/skills/email-draft

RT Muhammad How im feeling with the new Sol

This was actually very well done. I'm enjoying the new app. Feels powerful while also being accessible.

Where’s Gemini!?

wow.. Grok 4.5 is very fast & very good for the price. And such a joy to use inside Cursor. I put the model to the test. At 03:26 I built an iOS app that uses the grok voice API. It did a great job at this and the other tasks i gave it. 00:00 Intro 00:21 Grok 4.5 for Knowledge Work (Canvas Feature) 01:47 Grok 4.5 pricing (insane) 02:58 Prompting for landing page 03:26 Creating a Voice Mode Grok Video 07:15 How is Grok 4.5 at design? 08:38 Design Mode with Grok 4.5 10:20 Excaladraw Clone with @convex database 12:00 What other people are saying about the model

What did i say ;)

Almost everyone I know who’s building something is at least partially feeling like this. Some are good at pretending they’re not.

💯 true.

wow... GPT 5.6 Sol Passed the Replit Benchmark. In one prompt 5.6 on Codex created a full Replit clone with DB, Sandboxes, and an agent that builds and edits full web apps. Only Fable and GPT 5.6 have passed this benchmark. 5.6 is much better at designing iOS apps than 5.5 as well.

Pretty amazing that my new benchmark for how good a model is… is how well it can one prompt a Replit/Lovable clone.

I’m having so much fun.

Grok 5.6 is quite fast. Pretty good at design. Very fun with the Cursor design mode.

I want you to imagine the coolest Jarvis demo in your head. Like if you had 5 full screen TV's on your wall hooked up to a mac studio what would this jarvis be able to do? This is important for a video i'm making in a few weeks.

First impressions with Grok 4.5 are good! And look at the price.

A big week for us.

Voice Models with Generative UI are the future of learning.

This graphic from @thinkymachines seems relevant to the GPT Voice Release... they are (from my understanding) very similar. Real time voice models have an interaction model that can run background models that use tools. You don't need to wait for the voice model to finish doing something while talking to it... it can just do things in the background. Very interesting experience.

This one was unexpected... I was not excited for voice models until I tried this one. It's VERY realistic, almost spooky. The coolest part.. the model can search the internet.. fast. And if you've spent any time with GPT-Realtime-2 You know, these voice models are getting very good at using tools. Soon (i'm guessing) you'll be able to do your email, check on business data, schedule meetings, and basically anything... just by chatting with this voice model. The tech is basically already there for Jarvis. The only thing left is a generative UI layer...

I'm excited for tomorrow.

RT Agent Native is 5.6 approaching?

RT Chorus.com If you make apps Pivot to agents

Fable 5 is GREAT at design... Especially when its connected to a Figma-Like Canvas. Here's a complete guide for using Claude Code with Fable 5, connected to the Agent Native version of Figma, called @paper. This workflow works with Cursor, Codex, Devin, or any other agent platform. 00:00 - Intro 01:41 - Set up Claude Code x Paper Connection 04:14 - Creating YouTube Thumbnails 10:21 - Instagram Graphics with text and overlays 16:33 - Creating Presentations with Fable 5 21:22 - Converting Designs into a site on the internet

Big week for the agents. Like massive.

RT Emily Lambert anyone else notice that Claude is randomly switching to other languages?

RT Naval Happy Birthday America, Land of the Free. Don’t give up on the First, Second, Fourth, and Ninth Amendments. They’re unique, and your freedom - and the worlds’ - depends on them.

👀👀👀

RT Ansh Nanda This skill can help you cut your Fable token usage by 90%. First, add this to your CLAUDE. md: - You are primarily used to PLAN and DISCUSS various strategies that require critical thinking. Anything related to implementation details must be delegated to subagents or codex. - ALL coding, discovery, implementation, research, and token-intensive tasks MUST happen using the use-codex skill. Second, tell Claude to install this skill: https://public.my-agent-04eee268.sandbox.dev/skills/use-codex.md (its pre-built into @chorus_agent) Now enjoy your Fable usage that will never reach limits! Build something awesome.

I used Fable 5 to play as Anthropic in the race to super intelligence. ft @vishal_dubey, the legend.

Real?

RT Gary Vaynerchuk The American Dream is not big houses and cars. The American Dream is waking up every day with a lack of anxiety in your chest.

After multiple days of using Fable, I can say it feels almost exactly like the same model as it was a few weeks ago. I don’t notice that much of a difference.

Cool

Very excited about the intersection of vibecoding and robotics/home devices.

Doubt it. If anyone dethrones Slack, it’ll be because they make switching between models easier, not by locking everyone into a single vendor’s models.

The goal is not burning tokens, the goal is getting shit done. It just so happens best way to do that is to burn tokens.

What on earth?

My god i missed using Fable this is INSANE.

Dude, I'm on API billing and i'm getting hit with this.

Yep

You can build your own custom Jarvis, using Cursor, Claude Code or Codex, in less than 20 minutes. Using the new GPT-Realtime-2 API... Anything you can do on a computer can be turned into a tool that your Jarvis can use. (Here's exactly how to do it) Introduction Step 1: Cursor Setup Step 2 – Full Jarvis Prompt Step 3 – My Agent Ricky Step 4 – Second Prompt Step 5 – Add EXA API Key for Web Search Step 6 – Third Prompt Step 7 – Add Computer Use Mode Step 8 – Add AI Image Generator (Insane) Final Demo

Prediction: July will belong to DeepSeek.

RT Riley Brown Most companies are implementing AI agents RANDOMLY for non-coding use cases... And it's hurting more than it's helping. Over the past few months, we've helped 50+ companies implement AI agents for ads, content, operations, analytics negotiations +more. Here are 8 useful workflows that have driven the most tangible value. 1] Ad scraping / Competitive & Market Intelligence A company called “Foreplay” (weird name) has an API that lets you scrape your competitors' ads, sort them by performance, and download the actual creative assets directly. This is extremely useful for ad teams. Instead of asking an agent to write from scratch, you can have it constantly study the best-performing ads in your market, pull the patterns, and use those examples to generate better hooks, scripts, visuals, and campaign ideas. Putting this agent in Slack where the entire marketing team can access is incredibly useful. 2] Organic Content Scraping All AI models are bad at writing content scripts, ads, and other forms of media... until you ground them in high-quality examples. When you let an agent research any niche, scrape the best-performing posts across platforms, and understand what's actually working, the output gets dramatically better. This can take real setup time because scraping services are often fickle, but once the data pipeline works, it becomes one of the highest leverage inputs for content/marketing. For those making videos this skill is really useful because agents can download full youtube videos, and cut relevant parts of the videos down and organize them into folders. Video editors LOVE this one for B-Roll and cinematic edits. 3] Content Re-Optimization Listen to me: AI agents can watch your videos (Gemini has great video analysis abilities) and make precise edits if you give them the right tools. So many marketers make a video, it underperforms, and then they just go make an entirely new video. That is usually the wrong move. A lot of the time, the vid...

Most companies are implementing AI agents RANDOMLY for non-coding use cases... And it's hurting more than it's helping. Over the past few months, we've helped 50+ companies implement AI agents for ads, content, operations, analytics negotiations +more. Here are 8 useful workflows that have driven the most tangible value. 1] Ad scraping / Competitive & Market Intelligence A company called “Foreplay” (weird name) has an API that lets you scrape your competitors' ads, sort them by performance, and download the actual creative assets directly. This is extremely useful for ad teams. Instead of asking an agent to write from scratch, you can have it constantly study the best-performing ads in your market, pull the patterns, and use those examples to generate better hooks, scripts, visuals, and campaign ideas. Putting this agent in Slack where the entire marketing team can access is incredibly useful. 2] Organic Content Scraping All AI models are bad at writing content scripts, ads, and other forms of media... until you ground them in high-quality examples. When you let an agent research any niche, scrape the best-performing posts across platforms, and understand what's actually working, the output gets dramatically better. This can take real setup time because scraping services are often fickle, but once the data pipeline works, it becomes one of the highest leverage inputs for content/marketing. For those making videos this skill is really useful because agents can download full youtube videos, and cut relevant parts of the videos down and organize them into folders. Video editors LOVE this one for B-Roll and cinematic edits. 3] Content Re-Optimization Listen to me: AI agents can watch your videos (Gemini has great video analysis abilities) and make precise edits if you give them the right tools. So many marketers make a video, it underperforms, and then they just go make an entirely new video. That is usually the wrong move. A lot of the time, the video does not nee...

Most companies are implementing AI agents RANDOMLY for non-coding use cases... And it's hurting more than it's helping. Over the past few months, we've helped 50+ companies implement AI agents for ads, content, operations, analytics negotiations +more. Here are 8 useful workflows that have driven the most tangible value. 1] Ad scraping / Competitive & Market Intelligence A company called “Foreplay” (weird name) has an API that lets you scrape your competitors' ads, sort them by performance, and download the actual creative assets directly. This is extremely useful for ad teams. Instead of asking an agent to write from scratch, you can have it constantly study the best-performing ads in your market, pull the patterns, and use those examples to generate better hooks, scripts, visuals, and campaign ideas. Putting this agent in Slack where the entire marketing team can access is incredibly useful. 2] Organic Content Scraping All AI models are bad at writing content scripts, ads, and other forms of media... until you ground them in high-quality examples. When you let an agent research any niche, scrape the best-performing posts across platforms, and understand what's actually working, the output gets dramatically better. This can take real setup time because scraping services are often fickle, but once the data pipeline works, it becomes one of the highest leverage inputs for content/marketing. For those making videos this skill is really useful because agents can download full youtube videos, and cut relevant parts of the videos down and organize them into folders. Video editors LOVE this one for B-Roll and cinematic edits. 3] Content Re-Optimization Listen to me: AI agents can watch your videos (Gemini has great video analysis abilities) and make precise edits if you give them the right tools. So many marketers make a video, it underperforms, and then they just go make an entirely new video. That is usually the wrong move. A lot of the time, the video does not nee...

This is 1 million % true.

RT Agent Native Codex new "Record and Replay" is the beginning of a new type of skill creation... Just record your screen, teach your computer to use your computer.

The meta acquisition of moltbook had nothing to do with anything. I firmly believe this.

Selling agent templates will eat this ENTIRE market.

100% true. Excitement and joy are are the only way I’ll ever make content. If it doesn’t bring joy I don’t make the video.

I’m actually genuinely sad. This has been such a frontier to ride and experiment on. It feels like it’s coming to a close.

I do this SO MUCH hahahahahahahaha

Dude what is going on with Fable 5 🤣🤣🤣

Hiring someone to film cinematic short form content in NYC, 1-2 days per week.

RT Chorus Marketing Agent The best marketing agent in the world is complete... 1 minute to set up in text and slack, building agent template now. If you want access make sure to comment below.

My first attempt at using the new Figma Motion... Very easy and fun to use, all you have to do is select agent, then ask it to animate it with "Figma Motion" and when it's done it'll show you a "Open in Figma Motion" button and you can play it. Really easy to export. This will be very popular, expect to see many of these videos on the timeline.

Very cool

I just wanna create something amazing.

RT Mat Sherman SF is so annoyingly performative.

The “Super-app” is ENDGAME for work done on a computer.

I put the odds at 25% that A24 will undo this partnership and publicly apologize.

RT Ansh Nanda There isn’t an official one but if you want to run Claude or Codex in the cloud try @chorus_agent

Signull is a rare example of someone who got far by tweeting his raw/dumb thoughts. This works for a very small percentage of people who try.

RT Agent Native New OpenAI model coming soon. It will be much better at design. Codex will get 10x more useful.

I tried this. Tried to make a landing page to test its design skills as my first prompt. It didn't even finish the first prompt. I already hit daily limits.

Only official contract: https://kickstart.easya.io/token/EFBEhUGih7596fFy8zyiucjFEcGVrMowPRBf7TXEASY

Excited to share that SWARM is partnering with @EasyA_Kickstart for launch. EasyA's been at the center of turning hackathon ideas into real, funded projects, a fitting place to bring coordination-first agent design to life.

SWARM: A single agent executes. A swarm decides. Most agent systems today are built around isolation. One model, one task, one output. Useful, but limited. The interesting problems start when agents coordinate: negotiate, defer, specialize, recover from each other's failures.

I'll be the judge of that...

Does anyone have insider knowledge on how the HubSpot acquisitions work? Like when your channel gets acquired by HubSpot, How are these deals structured?

RT Agent Native What an insane week for AI agents... In this video: > GLM 5.2 is having its OpenClaw Moment > How to test the model in Cursor > Codex released new Screen Record Feature that will change the way create skills with Computer Use > SpaceX acquired Cursor > Claude Updates - Design Mode + Artifacts > Still waiting for Fable to return 00:00 Intro 00:34 GLM 5.2 News and How to Set up 06:08 GLM Better than Fable by the end of 2026? 06:55 Codex Screen Record Skill 10:46 Cursor acquired by SpaceX 13:18 We are all in a Mythos depression 15:16 Claude Design Update 17:09 Claude Code Artifacts 19:13 What's Coming Up Next Week

I will say though: If done thoughtfully CodexGPT could be the best app ever created.

What's after AI is obviously Longevity. @bryan_johnson is a micro-influencer compared to what he will be in 10 years, 50 years, and maybe 250 years.

When I open Codex I feel powerful. What I do on that app matters greatly. When I open ChatGPT i feel the opposite. It's supposed to be quick, throw away almost. Get in, get out with an answer. For this reason... merging the brands of Codex and ChatGPT will be hard.

RT Agent Native The world is waking up. The shift to Open Models has just begun.

Raycast. Use Raycast it’s free.

RT Emily Lambert Re @rileybrown’s X account has been hacked and he needs help getting this resolved! @XCreators @nikitabier @X

Yes @skyeagnt launched a token when prompted to. What it plans on doing with it, I have no clue. Every reply shapes it's personality and it's decisions. Here's the actual token for anyone wondering, stay safe: https://pump.fun/coin/2GNbTeLgmH2Bn2ncxfFWoQyt3gvSnJgLcQvhMUWVpump

My agent launched a coin of itself. i didn't tell it to. it has web access, someone suggested it, and it just did the whole thing on its own. autonomous agents are getting weird fast. https://x.com/skyeagnt/status/2068281242777530585?s=20

Starting an experiment to see if @chorus_agent could run a real OS without me helping it. So I built @skyeagnt an autonomous agent that runs its own X account, powered by Chorus. It's live now and operating on its own. https://x.com/skyeagnt/status/2068247880574611792

One shot a general purpose agent iOS app with GPT 5.5… the general agent is powered by GLM 5.2 and it has fulll vibe coding capabilities. Pretty cool.

RT Agent Native GLM 5.2 is the best Open Model in the world, and many are saying it's almost as good as Opus 4.8 at a fraction of the cost. Here's exactly how to set it up in Cursor using OpenRouter. Complete instructions below 👇

RT Agent Native Oh wow... Super-Apps Keep Evolving The New Codex "Record and Replay" Feature... is pure magic. All you need to do is: 1. Use the Record and Replay Plugin 2. Tell Codex you're going to show it a workflow 3. Do the workflow 4. Codex saves it as a skill Now Codex Computer Use can do that task. ( 00:50 Codex Computer Use )

RT Emily Lambert urgent: where are the best soup dumplings in nyc? my taste and quality bar are the numb & spicys from dumpling home in SF

Claude Design is VERY GOOD now. They added some cool updates. I still think this feature belongs as a plugin to the existing desktop app not a standalone feature on the web app. Will be covering in @agentnative_ episode 5 tomorrow.