SC

Stefano Corazza

Things Stefano Says on Podcasts

“People are usually all focused on the model and not the harness. But the importance of the harness cannot be overstated. … We're already on our third generation of harness for Brush. … We have really focused on making the harness thinner and smaller and more agile and simple. … It was more about removing the human knowledge and just letting the LLM run free to some extent.” — Stefano Corazza, The TWIML AI Podcast “You can fine-tune or train a model that is two orders of magnitude fewer parameters if you have the data and you can beat the bigger model. And then because it's smaller, you're going to be another magnitude faster and cheaper. … Right now we are at 40x cheaper than the most expensive frontier models.” — Stefano Corazza, The TWIML AI Podcast “There's a little bit of Silicon Valley bias that I've seen myself being part of, that you think the new tech comes and it's going to be single shot, perfect, and then people will just adopt it. But the reality is never the case, especially if the task has some creativity in it. … AI will get 70%, 80% there. It will never get to 100%. And so you have to factor in that for the last mile, you need a human in the loop and be able to iterate in a way that is fast and also fun.” — Stefano Corazza, The TWIML AI Podcast

Where to Find Them

Stefano Corazza has been a guest on The TWIML AI Podcast .

Recently: “Canva Is Tackling AI Design’s Most Frustrating Problem with Stefano Corazza- #784” on The TWIML AI Podcast (November 2026).

What They Said

“People are usually all focused on the model and not the harness. But the importance of the harness cannot be overstated. … We're already on our third generation of harness for Brush. … We have really focused on making the harness thinner and smaller and more agile and simple. … It was more about removing the human knowledge and just letting the LLM run free to some extent.” — Stefano Corazza, The TWIML AI Podcast

Asked what went into the custom harness around Brush, Canva's design agent, Stefano says the harness was the hardest part and the one with the most iterations. The lesson was to strip constraints out and let the model write code to operate design tools.

The TWIML AI Podcast · 2026-11-10 Permalink → Listen →
The TWIML AI Podcast Around 33:40 into the episode
Stefano Corazza

…will take, you know, a minute. And so that was not feasible. But now, you know, that's why speed is not just a nice to have. Speed is like a complete game changer because it can really increase the quality, it can increase the user experience. It can really make it like an iterative process.

Sam Charrington

Sounds like there's a simpler way to think about it. Craft is a model and Brush is a kind of an agent. And so in creating Brush, like I'm imagining part of the task here is like a custom harness. And I'm wondering what all that entailed. How did you approach that harness development?

Stefano Corazza

This is such a great question because I think people are like usually they are all focused on the model and not the harness. But I think the importance of the harness cannot be overstated. And for us, this probably was the hardest part and the ones, the part where we had the most iterations. In fact, we already like our third generation of harness for Brush. And really the high-level goal is basically to give as much freedom to the LLM to really run free of constraint and really be able to tap into the tools and have as much insight on what is able to generate. And so we have really focused on making the harness thinner and smaller and more agile and more and simple. And then at the end, Brush can write code to operate tools and make design operations. And so we were at the same time trying to leverage how good these LLMs are at writing code. And at the same time, simplifying the harness in the most basic possible way to really unleash the full power of the LLM. And so that's where we did a lot of iteration. And actually, it was more about removing the human knowledge and just letting the LLM like run free to some extent.

Sam Charrington

And when you talk about the code generation, what kind of code are we talking about? I can imagine a lot of what Brush might want to do is running a Python code, but I could also imagine you creating some design DSL that's like more natively represents the things that you care about.

Stefano Corazza

Yeah, it can be that, or it can be just writing code also to operate API calls and layout operations and that kind of stuff. So primarily

Sam Charrington

code or a little bit of both?

Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
“You can fine-tune or train a model that is two orders of magnitude fewer parameters if you have the data and you can beat the bigger model. And then because it's smaller, you're going to be another magnitude faster and cheaper. … Right now we are at 40x cheaper than the most expensive frontier models.” — Stefano Corazza, The TWIML AI Podcast

Asked why Canva bothered training its own models when frontier and open-weights models exist, Stefano Corazza argued that unique data plus vertical integration beats generic models on cost and speed. The 40x figure is his own claim.

The TWIML AI Podcast · 2026-11-10 Permalink → Listen →
The TWIML AI Podcast Around 07:43 into the episode
Stefano Corazza

…the loop and be able to iterate in a way that is fast and also fun. So that's a key part, I think, of what we do. Until the moment when we will have telepathic helmets that you can wear and you can transfer the full amount of your intent and your vision, but that is a little bit far out, I think.

Sam Charrington

…you could have. You, Canva, could have chosen to build on. On the other hand, it's like, you know, maybe it's obvious that a Canva with all of the data and user interactions that you have access to would build a model. Do you think it's interesting, or do you think it's obvious that you did it?

Stefano Corazza

No, I think it's very interesting because I think it also applies to other companies, maybe of similar size, when basically there's two factors. So, first of all, you have unique data that nobody else has. And then the second part is deep vertical integration of AI Design’s in your stack is going to provide incredible performance and cost reduction. Those two factors are basically the driving input. And I'm sure there's a lot of companies that are like in the thousands of employees or even smaller that will do that and where it makes sense. And so for us, it was kind of the discovery of the fact that you can fine-tune or train a model that is two orders of magnitude fewer parameters if you have the data and you can beat the bigger model. And then because it's smaller, you're going to be like another magnitude faster and cheaper, right? So that is the belief that we had. And right now we are in the 40x cheaper than the most expensive frontier models. They can do the same design task, but they are using others or magnitude more parameters to do it. Instead, we have smaller models that…

Sam Charrington

In what sense? Without making it a full episode, in what sense?

Stefano Corazza

Something happened like about, I think, four months ago where some of the companies developing video models started releasing open weights versions. And that had a fantastic domino effect where it forced all the competitors to do it. And so now you have an ecosystem where a lot of the image models…

Sam Charrington

With Kraft, did you start from an open weights model and train it, tune it, or is it a from scratch train?

Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
“There's a little bit of Silicon Valley bias that I've seen myself being part of, that you think the new tech comes and it's going to be single shot, perfect, and then people will just adopt it. But the reality is never the case, especially if the task has some creativity in it. … AI will get 70%, 80% there. It will never get to 100%. And so you have to factor in that for the last mile, you need a human in the loop and be able to iterate in a way that is fast and also fun.” — Stefano Corazza, The TWIML AI Podcast

Canva's head of AI research explains why the company built its design models around iteration rather than one-shot generation. It came up after the host noted that delays stack up when you iterate on creative work.

The TWIML AI Podcast · 2026-11-10 Permalink → Listen →
The TWIML AI Podcast Around 06:19 into the episode
Stefano Corazza

…where something magical gets generated. But like for if you want to really be using the day-to-day, like that's not an option. You know, in game development, there's a rule, anything that takes more than 10 seconds is completely out of the question. And we wanted to get to like instantaneous AI.

Sam Charrington

Yeah, I mean, particularly with the way that you need to iterate when you're creating, like those delays really stack up.

Stefano Corazza

Yeah, absolutely. And people, I think there's a little bit of Silicon Valley bias that I've seen myself being part of that bias that you think like the new tech comes and it's going to be single shot, perfect, and then people will just adopt it. But the reality is never the case, especially if the task has some creativity in it. It's not just a purely technical task. And so you have to build around iteration. AI will get 70%, 80% there. I will never get to 100%. And so you have to factor in that for the last mile, you need a human in the loop and be able to iterate in a way that is fast and also fun. So that's a key part, I think, of what we do. Until the moment when we will have telepathic helmets that you can wear and you can transfer the full amount of your intent and your vision, but that is a little bit far out, I think.

Sam Charrington

We're going to dig into the models and what you've created there, talk a little bit about how. But I'm trying to decide how interesting it is that you even did this. Like on the one hand, there are lots of, you know, their frontier labs have models and they a big part of their businesses are…

Stefano Corazza

No, I think it's very interesting because I think it also applies to other companies, maybe of similar size, when basically there's two factors. So, first of all, you have unique data that nobody else has. And then the second part is deep vertical integration of AI Design’s in your stack is going…

Sam Charrington

In what sense? Without making it a full episode, in what sense?

Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.

Collections They Appear In

AI Builders AI Research & Frontier Labs