👋 Howdy to the 1,029 new legends who joined since our last edition! You are now part of a 576,991-strong tribe outperforming the competition together.

LATEST POSTS 📚

If you’re new, not yet a subscriber, or just plain missed it, here are some of our recent editions.

📞 Why The Next Interface Can Hear & See You. An interview with Alex Smola, Co-Founder & CEO at Boson AI.
✨ My Favorite Startmate W26 Companies. A look back at a handful of the best and brightest from the latest accelerator cohort.
⚡️ Zapier’s AI Offer: Valuemaxxing. How the smartest AI companies are spending less on AI.

PARTNERS 💫

Accelerate global money movement with Airwallex Business Accounts. Manage all your funds centrally, convert at market-leading FX rates, and pay suppliers globally. Your AI agents can now provision accounts, move money, and optimize liquidity, all within the rules you set.

Framer is the AI website builder for teams that care about every detail. Agents work directly in the canvas, helping design pages, manage CMS content, write copy, add SEO, and audit sites. Trusted by teams at Miro and Perplexity.

• Build faster: Create, refine, and publish production-ready pages in minutes.

• Scale confidently: Manage content, localization, and metadata across your site.

• Stay in control: Review every change, collaborate in real time, and publish when ready.

_{Interested in sponsoring these emails? See our partnership options here.}

HOUSEKEEPING 📨

Sometimes I don’t really don’t have much to write about in my housekeeping notes. Today is one of those days. In cases like today, I pull out the trusted dog photo. This image below shows nine-year-young Ziggy (my dog), getting pestered by the puppy energy of six-month-old Bean.

It’s weird how the brain works. I’ve been a bit stressed and in my head over the last week or two, and when that happens, all my creative juices go out the window. Anyway, I’ve filled enough space with words to get it done. Over to you now. Enjoy today’s interview!

INTERVIEW 🎙️

Minh Cung, Co-Founder & CEO at Hyperlocalise

Minh Cung is the Co-Founder & CEO of Hyperlocalise, a Sydney-based AI-native localization platform whose agents translate and adapt product, marketing, and business content for new markets while keeping brand voice and local context intact. Before starting the company with co-founder Hans Bui, Minh spent four years building Canva's global localization infrastructure, including its one-click Magic Translate feature, and co-wrote the engineering deep dive on how Canva localizes continuously. He then led localization engineering at Heidi Health. Hyperlocalise is part of Startmate's Winter '26 Accelerator, which came with an AUD 120,000 investment, and the team has open-sourced a CLI that lets developers wire AI localization straight into their CI pipelines.

Minh is worth reading because he's lived the problem from both sides. He's multilingual, made his own cultural move to Australia, and spent six years watching companies discover their messaging rarely survives intact across borders. Localization has mostly been a Series B-and-up game, built on agencies, review loops, and in-house teams hired to manage them. Minh bets AI can shrink all of that to days and a fraction of the cost, so a seed-stage startup can go global from the start. The early signal came fast. One LinkedIn post about a waitlist drew sign-ups from Uber, OKX, and Canva, and during the accelerator he signed Heidi, WEEX, and Wall St. Rank as design partners, all in regulated industries. Along the way, he also learned the lesson most technical founders eventually face: selling to everyone means selling to no one.

What does localization actually mean for a startup or a scale-up?

Localization involves many things, but most people think of it as a translation layer. Let's say you have a product or a marketing campaign and you translate that into different languages. That's one of the easiest parts. You still need to involve humans, and there's a lot of project management here and there, but in simple terms, you can solve this with AI, which can come in and translate text very easily, at good enough quality. The second one, which is a bit harder, is what we call ‘transcreation’. Let's say you have a landing page, and you don't want to say the same thing to a different market. You want to say roughly the same concept, but with different wording and different imagery. Then you create something new from the source, and that's transcreation.

The ultimate goal every company is trying to reach is hyperlocalization. That's where we got our name. This is the hardest part, because you don't just create something; you actually need to apply local elements across your entire product. For example, you support local payment gateways, offer different products, and run different A/B tests in different markets. The problem moves from just translating words to the entire company operating differently in different markets. So, in simple terms, localization involves three steps, but most companies still stick with transcreation.

How has localization worked historically? Is it in-house teams, agencies, or a mix?

Today, it's a bit of a hybrid approach. Sometimes a company has an in-house team that manages the agency. Funny enough, they have to manage the agency. The agency typically does the translation, but with AI, things have shifted toward what we call post-editing. AI does the translation, then the agency post-edits it. That's where we're at. Typically, you see this in companies with a lot of capital because it involves a lot of humans in the loop. You have to hire an agency, and you have to spend millions of dollars to actually get this done and go global. So the companies that invest in localization are typically Series B and above, or big enterprises, simply because the cost of entry is huge. With Hyperlocalise, we're trying to change that: it's not just for scale-ups. A normal startup can try, experiment, and go global faster. Then, if you have money later, you can invest more in it.

How does Hyperlocalise help companies roll out faster?

We built an AI-native platform that lets teams create, localize, and publish multilingual content across every market and medium in minutes. What we're offering right now is a platform for teams to localize, or in a sense, translate their websites, marketing assets, and apps into different languages in basically one click.

That shrinks down the time and the cost. As I said, people used to need an agency to review and translate the text, which took weeks or even months. We're trying to shrink that down to days, and effectively shrink the cost as well. So instead of spending millions of dollars on this, you might spend thousands, or even less, to go global. And you can experiment so much faster.

Where's the biggest opportunity?

The biggest opportunity for me right now is companies in regulated industries. A typical company like that will go global anyway, and when it does, changes in local laws, regulations, and frameworks create new needs. Translation alone doesn't work. You can't just do it in English. You have to do different things in different markets. Scaling is very hard for those companies. They have to hire a lot of people and figure out whether spending $1M on this will return on investment. We're trying to shrink that down for those companies first.

Was Magic Translate at Canva your first touchpoint with this problem?

This actually wasn't my first encounter with the problem. I was a localization engineer at Canva and, more recently, at Heidi. I built the infrastructure that let those teams go global, so everyone could build on it and launch features in different markets. Magic Translate was one of the customer-facing features. Users could do it with one click, and their design would translate into different languages. That's the gist of it. At the time, AI was very new, and that one-click experience felt magical to many people.

What I learned there is that language is the key to connecting with people. That's what we're doing right now in this interview. We speak English, and that's how we connect. For companies going global, if you use the language the locals speak, you connect with people straight away. That's why I feel that, when it comes to going global, we need to lower the barrier to entry. We need to let companies speak the language first, and then do the other things. That's what I learned.

How do you build context beyond word-for-word translation?

Let's step back and look at how companies do it right now. A typical company, like Canva, writes many documents on its brand voice and market playbooks. If you go to Japan, you should do certain things, and if you go to Europe, you should do others. With Hyperlocalise, I'm building AI that reads these documents and understands the context. So when it does the translation, it knows we're going to Japan, that Canva's voice is very uplifting and happy, and that in Japan you need certain wording.

That's how we've built it into our platform. Teams can just upload a PDF or any other guideline documents, and our agent keeps them in memory. Then, when it translates, it applies that knowledge.

Do you add local knowledge on top of company docs?

Yes, we also build out our own data points. A very common one is that if you go to Japan at Christmas time, you should talk about KFC. We build this kind of common knowledge so that when companies launch a marketing campaign or work on their GTM approach, the AI knows about it and can suggest translations or transcreations.

In general, though, a lot of the time the product messaging itself is very short. Things like save, cancel, and other short words. Typically, those don't change too much, so the AI, with guardrails like the brand voice, should do a good enough job. But that's not all we're trying to tackle. As I said, we're tackling the legal side and the GTM side for these companies as well, which means we're moving a bit more toward marketing. That's why this sort of common knowledge is crucial alongside the company's own knowledge base.

Do you run on one model or a multi-model setup?

People in the industry have mixed opinions about which model to use. Some companies want to use Claude Sonnet, for example. We use OpenAI because it's cost-effective. So we're very open, in the sense that we allow teams to use whichever model they want. We built what we call a harness, so whichever model a team uses runs on the same infrastructure and with the same method. In general, we also try to stay platform-agnostic. For example, many teams already keep their knowledge base in Notion. So it's not just the model layer. On the integration layer, we're agnostic too, so we can fit every use case we can find.

What made you want to leave Canva and build Hyperlocalise?

I just wrote a LinkedIn post saying I was building an AI-native localization platform, and anyone who wanted to see it could sign up for a waitlist. That one LinkedIn post gathered so many people from big enterprises like Uber, OKX, Canva, and Agora, to name a few. So I thought, 'Shit (sorry for my language), this is something big.' Bigger than just the problem I used to work on at Canva.

I thought about it a little bit, and at the same time, the Startmate program was coming up. So I applied for the Pitch Night. I pitched it as a one-click way to translate or localize every piece of content into different languages, showed them my waitlist, and they accepted me. Being the opportunist I am, I saw it as something I needed to do. Timing is very important. If I don't do it now, someone else will probably pick it up. So here's some money, here's some traction with the waitlist, let's do it. That's why I moved on from Heidi to work on this.

What does a typical week inside Startmate’s 12-week Accelerator look like?

It's very busy. As with any early-stage company, you need to spend a lot of time on customer discovery calls. A typical day for me is three or four calls, plus a lot of back-and-forth LinkedIn messages to discuss the problem and understand what we're trying to build.

We do so much at the same time. And now, toward the end of the program, we have to spend a lot of time on pitch decks and investor calls. You'll see a lot of people out there saying fundraising is a full-time job, and yes, it is. But I still have to do the discovery calls, and I still have to do the product. So, very busy.

What meaningful belief did you pivot on between week one and week 12?

From my background, I knew the problem was huge and applied to everyone. But one thing I learned in Startmate is that selling to everyone is selling to no one. So the problem stayed the same: one click to translate and do everything globally, but we pivoted our positioning so we weren't targeting everyone. We're going after regulated industries and global enterprises that need to stay up to date on regulatory requirements.

Another thing I learned is that leadership cares more about ROI. Obviously, localization costs money, so they want to see a return in the next three to six months, not the typical one or two years. We're focusing on the GTM motion of different companies too. We have a few tools that let GTM teams create localized assets beyond English. They can create a Japanese sales deck or Japanese video graphics that feel relatable. We're trying to empower small teams of one or two people, like a remote satellite team in a different country. So that's another way we pivoted our positioning a little bit. In a sense, the product is the same, and what we're trying to solve is the same. It's just the ICP we're solving for and selling to that's slightly different.

Where was traction on day one, and where is it now?

At week zero, we had a waitlist full of enterprises. By week four, we'd finished our phase of discovery calls. We basically called all of them. We also posted more on LinkedIn, not just that one post, and generated much more interest. From week four to week 10, we signed three design partnerships. One is Heidi, another is WEEX, a crypto platform, and another is Wall St. Rank, a financial intelligence company. As you can see, they're all in regulated industries, where they want to be mindful about what they're saying, so we can see our motion is working pretty well there.

We've seen many other large companies and enterprises enter our pipeline. Atlassian is very interested in our self-serve GTM motion. We've seen Finder, we've seen Sitemate, and many other Australian companies and enterprises that are very interested in us. Startmate basically accelerated our traction in that sense. Hopefully, we'll have more by Demo Day, but for now, I'm pretty happy with where things are. By Demo Day, we'll try to show you all the technological breakthroughs we've made.

What's your North Star right now?

We're a B2B platform, so we need to sign as many enterprises as possible as design partners. That's one of the key things. But beyond revenue and traction, we want to see product usage and the value we bring to customers. We really care about customer feedback and customer love.

Every single week, we have a call with each design partner, and they all give us feedback. We track what they like and why they love us, and we also look at usage to see whether they like it. We're seeing weekly usage, actually daily usage, and the volume of translations is going very high. But again, that's not really our North Star. Our North Star is more about customer love and what they like.

How do you measure customer love this early on?

Our initial product offering is AI agents that help teams translate, work, and coordinate jobs better. So we track usage, like how many times they chat with the agent or use it, and we look at the number of outputs. If the output is very high and they use it a lot, that's great. The other part is the calls. If we hear them tell us they love this or love that, and we see the actual usage correlate with it, we think that's very important. But honestly, the underlying love is also about seeing them gain ROI. Wall St. Rank, for example, integrated everything in under two weeks and had everything ready to go global. Another metric we want is how long it takes a smaller company like that to go global, and for big enterprises, how fast we can solve the problem. And then, hopefully, they use it a lot as well.

As you scale past Startmate, what breaks first?

In general, our team and our product are trying to solve something everybody wants. Evidently, many people are on the waitlist, and there's a lot of interest. What worries us is the localization industry itself. As you said at the beginning, it's a bit of a mystery. People know about translation, but what else? A typical company will say it can just use AI to translate however it wants. That worries us a little, because most of the time, those companies do it wrong, simply because of the language's nuances. Not everyone has the knowledge and understanding of how to get it right. So I think it's the sales cycle.

Companies will do their own thing, realize it's not working, and then come to us, or maybe go somewhere else. That realization period is quite long for a lot of companies. Some companies are very interested in us because we promise they can go global very fast. But you'll see, especially in the US, people will say AI can solve this for them, and it takes them a long time to realize it won't, at least not at first.

How do you get the best out of yourself?

I think you need to find something fun to do as part of your routine. In general, your day will change a lot. For us, the routine is quite simple. We take the kids to school, go to work, and obviously pick them up. But the thing is, we need to pick something we want to do on that particular day. For me, it's creating LinkedIn posts, which is very fun, and creating creative images or videos. That keeps me happy and gives me positive energy.

Outside of work, you obviously need to do other things. If you have a family, spend some time with them and understand what their day is like. And sport. I think sport is important because a healthy brain needs a healthy body. If you don't play sport, you can go for a walk or a run, or touch grass, you know. Just something simple like that. Also, build up habits like not drinking and not overdoing things. Coffee is fine, but don't drink too much. Basically, try not to overstimulate yourself. Spend time in nature, go for walks, and help your body feel healthier.

Extra reading

• From Translation Management to Translation Intelligence - July, 2026

• Speaking everyone's language: In Conversation with Minh Cung, Hyperlocalise - June, 2026

• AI Translation Is Not Enough: Why Global Teams Need Context-Aware Localization - June, 2026

And that’s it! You can also follow Minh on LinkedIn and X, and don’t forget to check out Hyperlocalise to keep up with what they’re building.

BRAIN FOOD 🧠 

TWEETS OF THE WEEK 🐣 

TOOLS WE RECOMMEND 🛠️

Every week, we highlight tools we like and those we actually use inside our business and give them an honest review. Today, we are highlighting Svix—a webhooks-as-a-service platform that helps developers send secure, reliable, and scalable webhooks.

See the full set of tools we use inside Athyna & Open Source CEO here.

HOW I CAN HELP 🥳

P.S. Want to work together?

1. Hiring global talent: If you’re hiring tech, business or ops talent and want to do it 80% less, check out my startup, Athyna. 🌏

2. See my tech stack: Find our suite of tools & resources for both this newsletter and Athyna here. 🧰 

3. Reach an audience of tech leaders: Advertise with us if you want to get in front of founders, investors and leaders in tech. 👀