You are the one responsible for me getting all up in ai. It was the womens’ AI challenge. You lit a fire under this not so young woman. ANYWAY, I’m all in with Claude, build a couple of very useful agents, skills… now you’re saying go Codex. I can’t keep up. Tell me I’m ok with Claude for now. 🙃
Holy crap what I do with the four months of stuff I built in Claude, specifically using cowork, now that you come out with this?
I’ve built, so many automations, projects, and even a few full-blown apps (including a replica of the Iron Man Jarvis, which is browser-based, fully voice interactive and leverages Claude skills, schedules, and a complete hands-free weekly content automation system with Blotato that publishes over 25 AI Avatar videos and 50+ articles and text posts per week).🏆
I cannot imagine the cost in terms of time and stress that it would take to switch, let alone adding another subscription bill to my already bloated collection!!! 🙁😬
I totally enjoy and really value all your content and so I will trust that this is going to work and I’m gonna try it to see how I like Codex thank you so much for the reply! 🏆💖
One nuance worth having for your specific setup: the article's own migration section splits this by tool. The Codex import handles Claude Code automatically, but for Cowork it's a manual step, copy the files over, reinstall plugins, and rebuild your scheduled tasks by hand. Given four months of Cowork automations, that's the part worth checking before you commit, since Cowork migrates differently than the rest of your setup.
What if we’re new to Claude code and Codex and the main reason we want to use one or the other is to automate things like video editing, and social media creation and scheduling with Blotato. Which one would you recommend?
Worth splitting these into two different jobs before picking one tool for both. Video editing is local and file-heavy, so it depends on how well the tool handles rendering and iteration on your machine.
Social scheduling through Blotato is different: Blotato already is the scheduler, so what actually matters is whether Claude or Codex can drive Blotato's interface reliably, not which one has the better native scheduler. Test both against that specific workflow, since that's the piece doing the real work here.
Sabrina, I’d love to hear your thoughts on this new AI watermark that Anthropic is doing. Do you think all AIs will follow? Will it affect our automations?
This tracks with what I've seen building automations for clients: the model quality gap between Claude and GPT has mostly closed, so the deciding factor is now tooling and execution environment, not raw intelligence. We switched a client's ops stack off Claude for almost the exact same reason, the browser layer was the bottleneck, not the reasoning. Feels like 2025 is the year people stop picking a model and start picking an execution stack.
That matches what we've seen too, Mohamed. The split that matters more than model versus stack, in our experience, is deterministic versus judgment.
Swap the execution layer under a workflow that just needs Y to happen every time X happens and it still works fine. Swap it under something that actually needs judgment and the errors don't shrink, they just move somewhere new.
I’m a “normal” user. I’m an owner operator of a business (2 actually). My history is tied up in extensive projects, claude created artifacts & docs. It “knows me” - my voice, brands etc. i have some cowork stuff that acts as my daily project mgr using slack and asana. Product roll Outs, marketing campaigns. Ties to crm, finances. I like claude’s less obsequious, more grounded responses.
Assuming i wanted to start over, forget code, how does all that come over to chatgpt?
I’m a UIX guy and the claude experience is fragmented, though (if recent cowork changes are any indication though) it appears they are moving towards a monolithic UI. But a smoother UI and a few less keystrokes a day is not the reason to uproot a year of history for me with all the infrastructure i’ve built unless you can tell me its damn near a one button export/import - especially memory, instructions projects?
Here is the deal. I paid for a year of claude pro upfront. I have not put in the time to fully understand all the capabilities of claude to cut away at this point. I am also considering taking an exam and being certified as a Claude associate it’s a small thing but I’m thinking about doing it anyways.
And finally they can always update this software and make an update software for any company so I’m really looking at doing things where I am before I start to think about buying a subscription everywhere I don’t have the time or the interest to just bounce everywhere I do see that different companies have different drinks currently but at some point they’ll all become middle of the road or very unique so there’s a lot that I’m doing bouncing around like that just isn’t gonna benefit me right now.
Similar, Robert. I dove in head first to Claude. I’m using Visual Studio for Claude Code with ease and have developed a couple of hard working agents for my association. If it’s working, why switch it? And if I switch today, will I be switching back tomorrow? AI is moving so fast, I suspect we will hear “XYZ just surpassed ABC”…forever! Claude and Codex as we know them will no doubt be dead and buried in a year. Kinda sad and overwhelming, really.
I was wondering if you're ever gonna switch up from Claude. I found that Claude is just too expensive and over time it had begun hallucinating and just using too many tokens always, even with GitHub parameters set inside of it. I had to get away from Claude code. ChatGPT is my new go-to definitely.
Hi Sabrina. Love your intellect. Been vibe coding for 6 months. I’ve learned that neither product is better. I’ve paired one against the other in testing. Both contradict each other. It’s who’s pain you are willing to live with that will determine who’s tool you will use. I started with Lovable the exported my repo. I use Claude Code in terminal, Claude AI chat in browser. It’s amazibg to watch each window tell each other - each is wrong. I feel like I keep going in circles finishing a task. I don’t let X check the work of X but rather have Y check X’s work. Codex negates Claude’s work. I dont know who advice to approve at times. What are your thoughts?
Curious, any content you generate or repurpose, do you keep in Codex? I also have Codex, but use Clade primarily for creative tasks such as writing. With all the resets Openai does, it’s seriously the bigger bang for your buck.
I'm with you, Sabrina. Codex is a workhorse now — there's still a way to go with the overall experience architecture across the GPT apps/web interface, but its capabilities far surpass my experience with Claude. Also, the GPT voice mode is a game changer .
You are the one responsible for me getting all up in ai. It was the womens’ AI challenge. You lit a fire under this not so young woman. ANYWAY, I’m all in with Claude, build a couple of very useful agents, skills… now you’re saying go Codex. I can’t keep up. Tell me I’m ok with Claude for now. 🙃
ya just stick to Claude, it’s honestly fine
u in here already? https://www.skool.com/women-building-ai-2577
Yes!
Hi Sabrina, will you be doing a training on how to setup and run Blotato using Codex?
yes! it’s pretty easy though you just add it to MCP
Ooo this is so good thank you Sabrina
Holy crap what I do with the four months of stuff I built in Claude, specifically using cowork, now that you come out with this?
I’ve built, so many automations, projects, and even a few full-blown apps (including a replica of the Iron Man Jarvis, which is browser-based, fully voice interactive and leverages Claude skills, schedules, and a complete hands-free weekly content automation system with Blotato that publishes over 25 AI Avatar videos and 50+ articles and text posts per week).🏆
I cannot imagine the cost in terms of time and stress that it would take to switch, let alone adding another subscription bill to my already bloated collection!!! 🙁😬
its actually very easy to import, you don’t have to change anything, just ask Codex to copy over the claude skills to codex skills
I totally enjoy and really value all your content and so I will trust that this is going to work and I’m gonna try it to see how I like Codex thank you so much for the reply! 🏆💖
One nuance worth having for your specific setup: the article's own migration section splits this by tool. The Codex import handles Claude Code automatically, but for Cowork it's a manual step, copy the files over, reinstall plugins, and rebuild your scheduled tasks by hand. Given four months of Cowork automations, that's the part worth checking before you commit, since Cowork migrates differently than the rest of your setup.
What if we’re new to Claude code and Codex and the main reason we want to use one or the other is to automate things like video editing, and social media creation and scheduling with Blotato. Which one would you recommend?
Worth splitting these into two different jobs before picking one tool for both. Video editing is local and file-heavy, so it depends on how well the tool handles rendering and iteration on your machine.
Social scheduling through Blotato is different: Blotato already is the scheduler, so what actually matters is whether Claude or Codex can drive Blotato's interface reliably, not which one has the better native scheduler. Test both against that specific workflow, since that's the piece doing the real work here.
Where are the learning modules for the basics?
Sabrina, I’d love to hear your thoughts on this new AI watermark that Anthropic is doing. Do you think all AIs will follow? Will it affect our automations?
This tracks with what I've seen building automations for clients: the model quality gap between Claude and GPT has mostly closed, so the deciding factor is now tooling and execution environment, not raw intelligence. We switched a client's ops stack off Claude for almost the exact same reason, the browser layer was the bottleneck, not the reasoning. Feels like 2025 is the year people stop picking a model and start picking an execution stack.
That matches what we've seen too, Mohamed. The split that matters more than model versus stack, in our experience, is deterministic versus judgment.
Swap the execution layer under a workflow that just needs Y to happen every time X happens and it still works fine. Swap it under something that actually needs judgment and the errors don't shrink, they just move somewhere new.
I’m a “normal” user. I’m an owner operator of a business (2 actually). My history is tied up in extensive projects, claude created artifacts & docs. It “knows me” - my voice, brands etc. i have some cowork stuff that acts as my daily project mgr using slack and asana. Product roll Outs, marketing campaigns. Ties to crm, finances. I like claude’s less obsequious, more grounded responses.
Assuming i wanted to start over, forget code, how does all that come over to chatgpt?
I’m a UIX guy and the claude experience is fragmented, though (if recent cowork changes are any indication though) it appears they are moving towards a monolithic UI. But a smoother UI and a few less keystrokes a day is not the reason to uproot a year of history for me with all the infrastructure i’ve built unless you can tell me its damn near a one button export/import - especially memory, instructions projects?
Here is the deal. I paid for a year of claude pro upfront. I have not put in the time to fully understand all the capabilities of claude to cut away at this point. I am also considering taking an exam and being certified as a Claude associate it’s a small thing but I’m thinking about doing it anyways.
And finally they can always update this software and make an update software for any company so I’m really looking at doing things where I am before I start to think about buying a subscription everywhere I don’t have the time or the interest to just bounce everywhere I do see that different companies have different drinks currently but at some point they’ll all become middle of the road or very unique so there’s a lot that I’m doing bouncing around like that just isn’t gonna benefit me right now.
Similar, Robert. I dove in head first to Claude. I’m using Visual Studio for Claude Code with ease and have developed a couple of hard working agents for my association. If it’s working, why switch it? And if I switch today, will I be switching back tomorrow? AI is moving so fast, I suspect we will hear “XYZ just surpassed ABC”…forever! Claude and Codex as we know them will no doubt be dead and buried in a year. Kinda sad and overwhelming, really.
I’d love to hear everyone else’s thoughts.
I was wondering if you're ever gonna switch up from Claude. I found that Claude is just too expensive and over time it had begun hallucinating and just using too many tokens always, even with GitHub parameters set inside of it. I had to get away from Claude code. ChatGPT is my new go-to definitely.
Hi Sabrina. Love your intellect. Been vibe coding for 6 months. I’ve learned that neither product is better. I’ve paired one against the other in testing. Both contradict each other. It’s who’s pain you are willing to live with that will determine who’s tool you will use. I started with Lovable the exported my repo. I use Claude Code in terminal, Claude AI chat in browser. It’s amazibg to watch each window tell each other - each is wrong. I feel like I keep going in circles finishing a task. I don’t let X check the work of X but rather have Y check X’s work. Codex negates Claude’s work. I dont know who advice to approve at times. What are your thoughts?
This is very good
Curious, any content you generate or repurpose, do you keep in Codex? I also have Codex, but use Clade primarily for creative tasks such as writing. With all the resets Openai does, it’s seriously the bigger bang for your buck.
I listen to you, Sabrina.
I appreciate your knowledge and advice.
Thank you.
I'm with you, Sabrina. Codex is a workhorse now — there's still a way to go with the overall experience architecture across the GPT apps/web interface, but its capabilities far surpass my experience with Claude. Also, the GPT voice mode is a game changer .