Quick update - GLM 5.2 came out this week and I feel like this is a pivotal moment for Open Source AI..

I tested it for a few tasks and found it to be as good as Claude Opus 4.8 and GPT 5.5 in most tasks, and even better in some others… Even Qwen 3.7 is surprisingly good..

I think over the next 3-6 months we’re going to see these open source models getting even better and companies will start to migrate away from the frontier labs.. this will open up lots of new opportunities for us indie founders.. both for new business ideas, but also for how we run our businesses with small teams and low costs.. exciting times ahead!

For now, I’ve published my findings in 2 articles this week..

I tested the current best in class models on non-technical/non-coding work and published them here..

Do check them out -

Im AI Maxxing all my work as an entrepreneur, if you’re interested in more such studies and AI guides then let me know and I can share more such posts in the future..

For now, lets get back to today’s story - Hiro Analytics..

Hiro is an analytics and reporting platform built for email and SMS marketing agencies… the teams running Klaviyo, Postscript, and Attentive for e-commerce brands.

It pulls every client's data into one place and automates the weekly reports agencies used to build by hand.

$1M ARR in 19 months. Adding ~$10K MRR every single month.

Quick word from Apollo before we dive in -

Find, reach, and close your perfect deals — without juggling five tools or hiring more reps.

Apollo gives you everything you need to build real pipeline, fast. From inbound to outbound, first touch to close.

All in Apollo.

Sign up for free

Zach and Brendan didn't start with Hiro.

They met at a data-viz startup that got acquired by Amazon. They both worked there. Then left.

Started a newsletter business. It failed.

But through it, they learned email marketing.

That skill became an email marketing agency — which they ran for six years with ~30 clients.

Every week, the same grind: log into each client's platform, pull the numbers, format a report, repeat 30 times.

Brendan got fed up. So he built the tool he wished existed.

19 months later: $1M+ ARR. ~$96.2K MRR. The lowest churn Starter Story has ever seen.

Here's the twist — they don't sell to brands. They sell to the agencies that serve them.

One agency = 10-30 brands + dozens of paid seats. One sale does the work of 30.

From a failed newsletter to a $1M SaaS. By selling to the layer everyone else ignores.

Read the full Hiro Analytics case study →

That’s it for today

See you next week

Best

Ayush 🙏 (building Voibe)

PS: how you can be the top 1%