👋 Howdy to the 1,365 new legends who joined since our last edition! You are now part of a 571,352-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.

🚧 Selling Safety, Pricing Blind & Scaling Infrastructure AI. An interview with Shelley Copsey, Founder & CEO at FYLD.
🧑🏻‍🏫 Why Every CEO Needs Someone To Report To. An interview with Alexis d'Amecourt, CEO at Mochary Coaching.
🥣 Unfiltered: Granola Ate The Competition. A decade of AI note-takers, all beaten by the one that refused to join your call.

PARTNERS 💫

Meetings aren’t the problem. The challenge is everything that comes after them. With Granola, the AI Notepad for people with back-to-back meetings, it's no longer about scrambling to keep up.

Take notes the way you always have. Granola works in the background, turning conversations into clear summaries, action items, and next steps. Before, during, or after a meeting, you can chat with your notes to quickly understand what needs to happen, write follow-ups, or share context with others.

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.

• 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.

Trusted by teams at Miro and Perplexity.

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

HOUSEKEEPING 📨

One of my buddies just joined Glitch Capital to launch an a16z-style media team to build the brand and promote their portfolio companies. I’m Australian, back living in Australia, conscious that we have a way to go to grow our little slice of the tech ecosystem. Will (said buddy) and the team are polished and producing great stuff like this interview below with one of the lords of Australian tech, Dom Dym (also features twice in Open Source: here and here).

I think this is exactly how brands should be thinking. We’ve been building our negative CAC funnel at Athyna, and I think this is the VC equivalent. More deal flow, more LPs, and free earned media for the portfolio companies and broader ecosystem. Unreal.

INTERVIEW 🎙️

Alex Teichman, Founder & CEO at Happenstance

Alex Teichman is the Founder & CEO of Happenstance, a San Francisco startup building agentic network search for job seeking, sales, hiring, fundraising, and more. Connect your email, calendar, contacts, and social accounts, then give it a goal like "fill the three engineering roles on our careers page," and it searches your network until it finds the right people. More than 500,000 people use it, including teams at Perplexity, Greylock, Brex, Cognition, Accel, and Lovable. Garry Tan has called it the best social network search he has ever seen. Before Happenstance, Alex did his PhD at Stanford and co-founded Lighthouse AI, a 3D-sensing home camera, which was acquired by Apple.

Happenstance runs on the observation that the person you need for whatever you're trying to do is often already in your network. But there's no easy way to see who you, your friends, and your colleagues really know, so most networks go untapped. Happenstance changes that by indexing all the people you know and making them searchable to you. It's ChatGPT for working your network. Add friends or join a group for your team or community, and their connections open up to you (and vice versa), multiplying your luck surface area. The largest groups on Happenstance, like the All YC group, make millions of connections searchable. What Alex is building toward now is the version of Happenstance where you don't have to ask at all. He calls it ‘background luck generation.’

Why does the world need more luck?

A huge amount of human progress depends on people finding each other: your next great hire, your ideal next boss, the perfect early customer or design partner, a great investor for your startup, new deal flow in a space you're excited about, and so much more. Humanity is a distributed intelligence. It gets things done when its nodes connect with each other and exchange information. If you can make that process of finding each other ten times faster or better, that's a big deal. It meaningfully accelerates everything people are trying to do.

I used to do consumer electronics. We shipped a very sophisticated computer vision system with 3D sensing and custom optics. Finding the right manufacturing partners was genuinely hard, and working our investors' networks was essential to getting it done. Before AI, the general pattern involved finding the right little corner of the internet, shouting into the void, and hoping the right person happened to be listening. Sometimes that's a tweet asking if anyone knows a certified crane operator.

That's insanely inefficient compared to what's possible with AI. Making ten times more luck in the world isn't just achievable. It's inevitable.

What does Happenstance do that LinkedIn search cannot?

Here's an example. You open Happenstance Chat. It's agentic, so it feels like ChatGPT for working your network. You point it at your careers page and tell it to run an exhaustive search for each open position until it can't find any more good matches in your network, then do web searches for more detail on the best candidates, then reprioritize the list based on who looks open to something new, then draft outreach to your colleagues and friends asking what they think of each candidate and whether they could make an intro.

It does all of that in one shot. All you need to do is give it an objective and Chat will figure out what to search and who to filter out. At the end, you have lots of great warm leads from your network for that role. The same thing works for running your sales process, finding a new job, sourcing deal flow, fundraising, or whatever else you're doing.

What was the moment you knew you'd cracked product-market fit?

There was a moment early on, before we'd even incorporated. Someone in My Climate Journey, a community made up mostly of founders and operators in the sustainability world, needed to find the right person at PG&E to talk to about DERs. At that point, Happenstance was pretty primitive: some Python scripts and GPT-3.5 or GPT-4 working with data on my laptop. There was no website and no product. I would type a request in at the command line, press go, and see what came out.

So I ran his ask through the system and posted the results back into the community under my own name, as though a service like Happenstance already existed. It worked. He found exactly the right person at PG&E to talk to about DERs. That was the moment it became clear the world needed a lot more of this sort of thing.

What did Y Combinator change about how you build the company?

That one really made me stop and think. Surely somebody in that group knew exactly the right person, but there was no way to find them before AI. So I started messing around with demos in there, and an old friend from a past life in self-driving cars told me to take it to Garry Tan. I still didn't think I was going to do YC, but then I talked to Garry and discovered that I did, in fact, need to do YC. His argument, delivered with a fist on the table, was that Facebook started with Harvard, Columbia, and Stanford, and Happenstance was starting with Bookface, Post Exit Founders, and a few other entrepreneurial communities we were experimenting in.

And he was right. It was an incredible network to start from, and YC has been really fantastic. We've been in every batch since, and Happenstance is now an important part of how YC founders find their early customers, design partners, and investors.

How has your job changed since the pre-seed? And what does your week look like now?

In the earliest days, it was just me writing code and putting demos in front of people. As the team grew and AI got vastly more capable, the job became much more about causing things to ship than writing every line of code myself. And now that our user base is so much larger, it's less about showing demos to individuals and more about shipping A/B tests and measuring the results.

Nowadays, my days are super varied, and I love that. Some days I'm deep in the code on vector database architecture or retrieval evals. Other days I'm hiring and interviewing, talking to investors, or having conversations like this one with you. A lot of it right now is jamming on product with the team and going through the whole stack: what we're actually trying to accomplish, what the user experience should be, which metrics matter, and how to build it so it's efficient on both compute and cost.

Do you run the whole hiring process yourself at this stage?

It's not just me anymore. We don't have an in-house recruiter at our size, so hiring is a team effort. A few people on the team have gone out and found great candidates on their own, and a couple of posts here and there picked up real traction.

At one point, we got a thousand applications in a week. Broadly, though, it's two routes: using Happenstance to search our own networks, and people turning up in my inbox.

How do you build culture in a small team?

One of our recent hires asked whether we had anything written down about how we operate. I realized that what's obvious to me isn't always obvious to everyone else, and that it's worth writing down and talking about. Here are a few of the things I sent back.

Something I learned from my old co-founder: you can go 2x faster by working more and 10x faster by knowing which corners to cut. You should do both and go 20x faster. Every click and every hundred milliseconds of a user's attention is precious and must be treated that way. If a word, a component, or a button isn't necessary, cut it. Be ruthless about it. As for actually passing these on to a team, there's no magic. It comes down to leading by example, talking about them, writing them down somewhere visible, and referring back to them whenever the moment calls for it.

How do you use AI as a team day-to-day?

Work that would have been a full two-week sprint for a team of ten great people at my old startup, we now routinely do with one or two people using AI in a day or two. It's crazy. You have to know how to use it, of course. No slop grenades, no meat proxies. One of the most important skills now is knowing which parts of the code require your deep understanding, which just require the right tests, and which you can sanity-check just the results and not worry about the details. Once you have that innate sense, where you can review code and see which category each block falls into, wow, you can go fast.

The concrete piece is our own internal harness, hpn-agent, and it's a delight. Claude Code, Codex, or whichever agent we want drops inside it with access to everything it needs, in a way that's programmatically and cryptographically safe. It can't break anything, but it can read from all the appropriate places. It's the same setup you'd give an intern. They shouldn't have write access to your production Postgres, but they should be able to use a read replica to run an analysis. At the AWS and GitHub level, our agents can't push to the branches that really matter, but they can make their own feature branches and do whatever they want there.

What was the hardest technical problem in the early days, and what is it now?

In the earliest days, the AI models just weren't very good. You could barely torture decent results out of them. This was roughly the GPT-3.5 and GPT-4 era, and you could make it work, but it took a lot of bending over backward. Today, the hardest problem is scalability. There's a lot of data (several hundred million vectors in a vector database, for example) and a lot of small architecture decisions you have to get right, or it just won't work.

What's maybe not obvious is how many agents there are across Happenstance. Chat is agentic, and the search underneath it is agentic, so both need their own evals. There's also an enormous amount of AI work you don't see, like determining with high confidence that an email address belongs to a certain person, or finding someone's public social posts, personal website, and news coverage on the web and turning it all into something users can search.

Where is search heading now?

One thing I'm very excited about is the shift from you asking us for luck to us making it for you, based on whatever you choose to share. Our Granola integration is the first real example. If you've connected Granola to Happenstance, then the moment a meeting ends, we look at whether there's a way you could help the person on the other side of it. If there is, we suggest the searches without being asked.

That's the category I want much more of, and the short name for it is background luck generation. Our bread and butter so far has been agentic chat, where you go and work your network yourself. That's the tagline on our site: make your own luck. What comes next is luck being made for you and the people around you, continuously, without any of you having to think of it.

How do you get the best out of yourself?

The most important thing is to work on something you're super excited about. That makes everything else easy. I'm very fortunate in that way. Beyond that, it's doing the boring stuff: sleep enough, eat well, exercise. None of that is controversial anymore, though there was a time when people really pushed the idea that you should sacrifice sleep for work. If you want your brain working well, get enough sleep.

A less obvious one is making a conscious effort to control what's in your context window. Anything political makes it harder to work, so I mostly don't let it in my brain. Reading sci-fi about synthetic serendipity has the opposite effect: it leaves me inspired to ship more of it into the world and help our users more.

Extra reading

• Granola: Healthy Breakfast & Killer AI Notepad - August, 2026

• The consumer AI apps breaking out in 2026 - March, 2026

• AI Agents & The New Developer Workflow - January, 2026

And that’s it! You can also follow Alex on Happenstance, LinkedIn, and Twitter, and don’t forget to check out Happenstance’s website 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 Airwallex—their agentic business accounts instantly provision accounts with local bank details to receive payments around the world.

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. 👀