Using Parallel AI Agents and Agentic IDEs to Scale Fast (James Dooley Interviews Kasra Dash)
AI SEO & Business Automation Podcast
September 28, 2026
This video explains which digital marketing strategies agencies should focus on in 2026 to improve output speed, website deployment and content quality.
Speakers James Dooley, Kasra Dash
TopicsMarketingBusinessManagement
James Dooley (0:00)
Agentic IDE with Parallel Agents. Today, I'm joined with Kasra Dash, and we're going to be talking about agentic IDEs, and why it's important to be looking at autonomous agents and AI harnesses to be able to scale fast. There's a lot of people know that are wanting to move a lot quicker, they're wanting to test things out and move fast, and everything's coming down to speed, speed, speed. And it's trying to be getting the stuff out there, seeing what works, almost throwing as much mud at the wall, see what sticks, and then they can then scale out and double down in that lane. So with regards to an agentic IDE, like let's say CODIS, for anyone who's watching this, this is something that I use myself. Check out the link in the description to the CODIS review.
Why is it important Kasra to be using an agentic IDE to have parallel agents being able to run asynchronous with each other?
Kasra Dash (0:57)
So the biggest thing is when you're using a harness that isn't, for example, Codex, Desktop or Cloud Code Desktop, is you're not tied to a specific company. So there's so many of these agentic IDEs, like there's Superset, there's CODIS that you mentioned, there's Conductor, there's so many. But what you can do when you're using these IDEs is you can basically bring your own Cloud Code or you can bring your own ChemEK or you can bring your own Codex, ChatGBT. So it technically doesn't matter which one is actually winning the AI race because you can just swap and change whichever AI agent you actually want to pick.
Which for business owners like yourself is very important because it means that one minute you're not needing to download Cloud Code, then the next you're uninstalling Cloud Code to then install Codex. You're not then relying on a singular company, so you've not got a single point of failure.
James Dooley (2:06)
Yeah, for sure. I mean, for me, the Agentic IDEs, it's just all in one dashboard. I don't need to then start deploying, looking for a different agent to be looking at errors. I don't need to have it as a different dashboard looking for something else. It's all in one platform. It's that code editor that's able to run asynchronous and in parallel to each other in trying to make certain that, as let's say new vibe coding websites is being done, quality control measures is being done alongside it, content writers is being done alongside it, image creations are being done alongside it, and everything's running in parallel with each other. But before you know it, within 15 minutes, I've got end-to-end from a topical map to content briefs, to writing the content, to the images being created, to the coding of the website, to the schema being done, to the internal linking being done, all that is running parallel. It's not just having one single lane being run, going, okay, I've developed the website, now I'm going to come and then start setting up schema. Okay. It's like having one member of staff do everything and that one member of staff burning out. You've got all these running side by side in parallel to each other. I think for me, that for the speed, not only is it great on better quality where you can cross reference against each other, but the speed and how fast things are being done with these agentic IDEs.
How much faster now is it for you yourself, Kasra, where you've got all these autonomous agents and this agentic IDE? How much faster is it for you to deploy a new website in comparison to, let's say, six months ago, where you would just ChatGPT prompt in and out, and then having to then orchestrate everything yourself from start to finish?
Kasra Dash (3:53)
A lot faster. So previously, what I was doing, and I thought that this was pretty advanced, like at the very beginning when people had just started using ChatGPT for coding, you would basically give it a screenshot of a section of your website, and you would say, like, hey, code this, for example, and it would basically code that, and you would need to copy all of that code, and then paste it into your WordPress Element or your WordPress Gutenberg design. I thought that was pretty advanced. Then that's when a couple of months after that, you could end up creating WordPress themes on the fly. Again, I thought that was pretty advanced, but now, literally, you can give any one of these agentic IDEs access to your website, it can like upload themes, it can download your current theme, it can make any changes.
2 more minutes of transcript below
Thousands of transcripts fetched by people building searchable podcast archives
Fetch the whole transcript
The demo key returns a sample episode in full, no card needed:
curl -H "x-api-key: pt_demo" \
https://spoken.md/transcripts/1000651996090Markdown with the speakers named, for your notes, your knowledge base, or anything that makes HTTP calls.
From $0.10 per transcript. No subscription. Credits never expire. Prices exclude VAT, added at checkout for EU customers. Not what you expected? Email us within 14 days with 20 or fewer credits used and we refund the pack in full.
Using your own key:
curl -H "x-api-key: YOUR_KEY" \
https://spoken.md/transcripts/1000791964928