- 15
- Episodes
- Bi-weekly
- Cadence
- 2026
- First episode
About Still Early
Every other week, Niklas Stephenson and Peter Suhm hang out and talk about what's been going on in their businesses.
- Publisher
- Niklas Stephenson & Peter Suhm
- Category
- business · technology
- Language
- en
- Explicit
- No
- First episode
- 17 Apr 2026
- Latest episode
- 20 Oct 2026
Latest episodes
15 episodes in the feed.
20 Oct 2026
The podcast is in person now
In this episode: Why Phone.inc gives away phone numbers for free What’s getting built in October (call whispers, voice-changed greetings, maybe SMS) A bug report button that hands every report to a coding agent ChatGPT did almost all the work of submitting Phone.inc’s ChatGPT plugin Why they’re making their videos in-house instead of hiring an agency First episode recorded in person, in the Phone.inc office. A lot of people just need a phone number for something (the App Store, Stripe, registering a company), and Phone.inc’s free plan gives them one. Peter thinks it’s a huge use case that’s still up for grabs. With freemium you own the lead, as Patrick Campbell used to say, and Phone.inc literally has their number. Niklas cares less about why they sign up. If Phone.inc is where someone got their first business number, it’s probably where they’ll go the next time they need one. October is the first month that isn’t only foundational work. Call whispers give you a few words when you pick up (who’s calling, and probably why) based on caller context and previous calls. Welcome greetings get a voice changer, so you record the greeting yourself and it comes out in a different voice, with your pronunciation of the company name intact. The stretch goal is texting on business numbers, the most requested feature. It’s pretty easy in the US and UK, and a lot more work in Denmark. Niklas vibe coded a bug report button. Customers click the part of the page that’s broken, Sentry captures the session, and an Amp thread starts on a fix right away, all the way to a PR that’s ready to merge. In his first week as growth lead, Peter submitted the Phone.inc MCP server as a ChatGPT plugin. ChatGPT’s plugin creator did nearly all of it (test cases, checking his demo video, telling him to set up a test account with a call and a voicemail). He thinks that’s roughly how onboarding should feel. Peter wants Phone.inc to eventually make the kind of launch videos later-stage startups put out. They could hire an agency for that, but they’d rather get good at it themselves, which means shipping something every week and improving as they go. Companies that buy one great video often struggle to make the next one. And with so much AI-generated video around now, they want theirs to clearly be made by people. Still Early is sponsored by Phone.inc (https://www.phone.inc/): a business phone number, welcome greeting, call routing, and opening hours, all from an app on your personal phone.
6 Oct 2026
My first day at Phone.inc
In this episode: Peter's first day at Phone.inc (office breakfast, a design sync, and half an hour picking Lea's brain about the customer journey) He didn't have to set up a local dev environment, and made his first PR through Amp Jesse Hanley from Bento tweeted that he didn't want to be on the waitlist, which turned into a bunch of new signups Phone.inc hit the September goal of 100 companies (124 now), with about 15% of free users converting to paid Niklas hired a VP of sales, starting November 1 What does an activated user look like for a phone number that might never ring? Peter started full time at Phone.inc on Thursday. He filmed the whole day for a vlog (the coffee machine, the car door, the drive to Copenhagen in his right-hand drive manual from 2012). Lea had ordered breakfast, then came standup, the Thursday design sync that Katrine runs, and a priorities call with Niklas (who was home with his kids after three days of travelling). After lunch he spent half an hour asking Lea where customers are right now. The nicest surprise was that he probably doesn't need a local development environment at all. Everything was already set up in Amp, so he typed into a text box, got a tour of the app and the website, and had staging sites created for him. His first PR was adding the Bento logo to the website. On Friday Peter woke up early and saw that Jesse Hanley from Bento had tweeted that he needed a number and didn't want to be on the waitlist. Peter had gotten admin access the night before, so he invited Jesse and a few others from Twitter. That led to a lot of tweets back and forth, new followers and a bunch of signups. Jesse asked how this can be free. His use case is simple, Phone.inc runs its own voice infrastructure so the cost is low, and him tweeting about it is worth more than what he would've paid. Phone.inc hit the September goal of 100 companies and is at 124 now. Around 15% of free users convert to paid, and there are no gates yet (just an upgrade button). There are paying customers in the US, the UK and Finland, and the US has passed Denmark in the last 50 waitlist signups. Niklas spent two days in Stockholm at a go-to-market meetup one of their investors organized. A revenue leader from a fast-growing AI company talked about running self-serve and enterprise sales at the same time (most of their early enterprise deals started on the self-serve side). Niklas also hired a VP of sales, someone he's worked with before who builds direct sales teams from zero. He starts November 1, and there'll probably be a sales team from January. Peter is in the Copenhagen office two or three days a week for meetings, feedback and lunch, with remote days for deep work. He's still sorting out the logistics (a new electric car, a school carpool, the right train, and moving house in two weeks). Peter is responsible for activation now, and it's hard to define for a phone number. Jesse might never get a call and still be an active user. Niklas wants one simple metric (weekly active companies that made or received a call), plus softer signals like whether the number is on the company's website. They talked about a test call button, asking people at signup why they need a number, and what to do with free numbers nobody has used for six months. Neither of them thinks there's a right answer yet. Niklas keeps both his notes and his tasks in Linear, in a private team, and adds to it by talking to Claude in the car. Still Early is sponsored by Phone.inc (https://www.phone.inc/): a business phone number, welcome greeting, call routing, and opening hours, all from an app on your personal phone.
22 Sept 2026
My first real job since 2012
In this episode: Tailwind is joining Shopify, and Peter isn't going with them Why he said no to pursuing Shopify (big company, no side projects, night calls from Europe) He's joining Phone.inc as product growth lead on October 1 Why Niklas wanted a founder type in the role instead of someone with a finished playbook The 2012 job Peter quit after four months, and the case for founders taking a job for a season Phone.inc shipped do not disturb and call transfer, and invites are open again Tailwind is joining Shopify. Adam Wathan announced it last week, and Peter's last day at Tailwind Labs was the day before. He'd been there four years, never full time (a couple of days a week as a contractor), but it was his main stable thing the whole time. He originally joined to build a job board, which he'd completely forgotten until recently. His own startup was near the end of its runway, and Adam asked if he wanted to start a business inside Tailwind Labs part time. It later turned into running partnerships. He's known about the acquisition for a long time, which made Laravel Live Denmark a month ago pretty strange (the perfect place to talk to everyone about it, and he couldn't). Following the team to Shopify was something Peter could have pursued, and there were a ton of good reasons to do it. But he's never really wanted to work at a big company, and if he ever did, he wanted it to be a decision he made on purpose. He also had too many things going on the side that he wasn't ready to let go of, and he'd have been the only person in Europe on a North American team. Adam told him he didn't have to come, and that it would still be a good outcome for him if he went and did his own thing. Instead, Peter is joining Phone.inc as product growth lead on October 1. He was already an investor and helping out as a contractor. The role covers marketing and the part of product that overlaps with it (the founder-led growth videos, onboarding, changes in the app itself). Niklas thinks of it as founder in residence, and Peter says it looks a lot like his job when he was starting Reform. A friend had been asking him for months why he was only there one day a week, since he already sounded like he was running growth. Peter isn't very employable (the last time he didn't work for his own company was 2012), and his real worry was joining when he knows he'll want to do his own thing again someday. Niklas said that's exactly who he's looking for. An experienced growth hire usually brings the one playbook that worked somewhere else, and at this stage Phone.inc needs someone who can go find the one that works for them. Peter also tells the story of his first job as a junior Rails developer. He quit after four months, and his boss offered to fund his startup if he stayed. Niklas makes the case (from his blog post "Seasons in life") that founders can take a job for a few years without it being an identity crisis. He did it at Trustpilot after Firmafon and says it was the best thing he's done career-wise. On the Phone.inc side, invites were paused in early September until two core features shipped: do not disturb and call transfer. Both went live yesterday. At Firmafon, transferring a call meant remembering a star code, so people just didn't (some customers literally threw the phone across the office instead). Now there's a transfer button next to end call, a live list of colleagues and their status, and you can brief the person before handing over while the caller listens to hold music Mikkel made in a day of jamming (Niklas says it's the best he's ever heard). Invites are back on, with ten new companies in a day and a goal of 100 live companies in two weeks. Phone.inc is also hiring someone to lead direct sales. The podcast will follow Phone.inc even more closely from here, hopefully recorded in person at the office. Still Early is sponsored by Phone.inc (https://www.phone.inc/): a business phone number, welcome greeting, call routing, and opening hours, all from an app on your personal phone.
8 Sept 2026
Working while I drive
In this episode: Peter went to Laravel Live Denmark, which grew out of a Copenhagen meetup he helped start in 2014 Everyone there assumed he works full time at Phone.inc now (the daily videos are doing their job) The form builder idea he never shipped at Reform: a form, plus an AI-built backend to actually process the replies Niklas said for months that Phone.inc wouldn't build call queuing, then built it in two days Claude's voice mode works with MCP servers on mobile, which turns a drive to Copenhagen into real work Peter wants to turn Memo into a microblogging platform (both ChatGPT and Claude told him it's procrastination) Peter went to Laravel Live Denmark in Copenhagen, which is a bit of a full circle thing. He and Visti started the Laravel meetup there in 2014 (he tried in 2013 but nobody replied to his forum posts), and Kasper has since turned it into a non-profit and a 300 person conference. Aaron Francis, Taylor Otwell and DHH all showed up this year. A lot of people assumed Peter now works full time at Phone.inc. He doesn't, he's still pretty part-time, but there's something on the channels every weekday now so it feels like he's always around. The batch of Q&A shorts they recorded before summer is what makes that possible (Joseph is editing them). Four are out, around a thousand views each on YouTube and roughly the same on TikTok. Niklas got recognized from the videos by a stranger in a swimming hall changing room, while naked, which he took as a good sign. The form builder came up again. Nanoform is the product Peter is most excited about, but the version he actually wants is the one he never built at Reform. A form is a small piece of software, and the replies always need a process behind them. So the AI builds the backend too, picking from real building blocks (a kanban board, a table, separate pages, whatever fits) instead of generating it from scratch. Niklas wants lead scoring and agentic steps that fire when a card moves status. Peter's phrasing: vibe coded forms, without the slop. The name is still wrong. Niklas had been saying for months that Phone.inc wouldn't build call queuing, since it's a feature for full time support teams and that's not who they're for. Then he looked at the data from the first 40 companies and found people waiting in line anyway (50 calls a day is fine until they all land during lunch). So he built the simplest possible version in two days. Every call is queued now, even the ones answered instantly, and there isn't a single new setting or checkbox anywhere. Do not disturb is next and falls out of the same model: turn it off and a waiting call rings straight through. They're also pausing invites for a few weeks. Forty companies in, there are four foundational things to tighten up first, and none of them were a surprise. The first number porting went through too. Niklas discovered that Claude's voice mode on mobile has all the MCP servers from his desktop, so an hour and a half of driving became actual work: Linear tickets, debugging calls, pulling product feedback out of customer service conversations. Peter can't stand talking to it (a clever door, he says), but they agree it's a bit mad that Siri isn't this. TechBBQ has grown into more of an innovation fair than a startup event, and Niklas would rather have something the size of Laravel Live Denmark, invite only, founders and investors. The side events around it were still great. And Peter wants to turn Memo into a microblogging platform. Longer than a tweet, shorter than a blog post, RSS, no algorithm, somewhere to document building the form builder. Both ChatGPT and Claude told him it's an elaborate form of procrastination. Niklas told him to build it anyway. Still Early is sponsored by Phone.inc (https://www.phone.inc/): a business phone number, welcome greeting, call routing, and opening hours, all from an app on your personal phone.
25 Aug 2026
I killed my local dev environment
In this episode: Why Still Early and the Phone.inc vlog are two different shows, and who each one is for Phone.inc has 15-20 companies live on their own infrastructure, and the inbound is getting hard to keep up with The two kinds of partnership Niklas thinks are actually worth doing (and why most aren't) Niklas hired a head of engineering in about a day Peter got orbpilled: three days on Amp, a form builder and a kanban board, none of it built on a laptop Nanoform's entire interface is a prompt you paste into an agent Phone.inc has started letting people onto the platform, slowly, and it's all running on their own infrastructure now. 15-20 companies so far, a mix of people with a real need and friends of the house who are good at giving feedback. The whole team ran the signup flow and did a friction log, which produced 20-25 issues in Linear, and Niklas had Cursor agents fix the ones he agreed with (watching cards move across the board by themselves was apparently pretty great). The pull is different this time. Random people calling Niklas's private phone because they heard about it, 55 unanswered LinkedIn messages, an inbox he gave up on and archived. Lea pulled him aside to point out they never felt anything like this in all the Firmafon years. Partnerships too, which Niklas would normally tell founders to avoid (usually all motion, no customers, and Peter got burned this way with Branch). But two categories look real: vertical SaaS platforms where the customer data lives, so a call can come in with context attached, and incorporation platforms where you catch a company at day one and never get switched out. He also hired Marcel as head of engineering. They'd worked together at Trustpilot and Whimsical, the whole team met him the same day, signed eight minutes later. Thomas made the first Android call too, with the business logic shared in Swift and the apps still native. Then Peter, who has had a slightly unhinged week. He tried an OpenClaw-style setup, didn't click, and then Adam told him to try Amp's orbs. He's now completely orpilled. The point isn't the agent, it's not needing a dev environment at all: spin up an orb, get a URL, comment directly on the thing, review it later. He built a hosted version of Nanoform this way, where you never touch a form builder (you paste a prompt with a link to the agents file, the agent builds the form, you click claim). He also added JSON bodies and kanban boards to Memo, which agents can update by patching JSON. All of it from his phone, including one session from the corner of Max Burgers. They also get into where this breaks down: iOS builds, push notifications, FreeSWITCH. Niklas thinks Phone.inc will have to build those runners themselves. --- Still Early is sponsored by Phone.inc (https://www.phone.inc/): a business phone number, welcome greeting, call routing, and opening hours, all from an app on your personal phone.
11 Aug 2026
Why I'm cutting our runway
In this episode: Peter's first vacation in years without a computer (and using ChatGPT as a battlefield guide from a bench) Where memo is at: Peter is still user number one, and the aha moment problem he hasn't cracked Executable memos, and whether a reminder is really just agent orchestration Could an AI run OG Kit? Where to start, and wanting a layer above the pull request Niklas cut runway from eighteen months to thirteen so he can hire faster Rebuilding the Phone.inc website with Cursor and ui.sh, plus 25 integration landing pages built while he slept Both of them are back from vacation, so this one starts with camping. Niklas did Norway in a small van with two kids, two car seats, two bikes and a stroller (lovely, but next year will be less ambitious). Peter camped too and spent about half of it getting kids to sleep in a tent. He also left his computer at home for the first time in years, which went fine apart from the odd thought about what would happen to it if something happened to the house. The best moment was sitting on a bench looking across the water at one of the most famous battlefields in Danish history and getting ChatGPT to explain the whole thing from exactly where he was sitting (that hill over there is where the long range cannons were, that island is where the Danes retreated to). Peter's update is mostly memo. He's invited a bunch of people, most haven't really activated, and he's still using it a magnitude more than anyone else. He's made it automatic for himself with prompts and instructions, so Claude and ChatGPT just know everything lives in memo and they read, write and update memos on his behalf. The problem is that nobody else gets to the aha moment unless it's automatic for them too, where the agents fill up their memo in the background and one day it does something they didn't expect. He hasn't solved that yet, and he's not in a hurry, since he's learning a lot as user number one. The idea he's excited about is executable memos: a memo that runs at a certain point in the future. Start with reminders, and then notice that a reminder telling your Claude to do something in two weeks is just agent orchestration. That ties into OG Kit, which Peter isn't really working on but people still pay for and like (Phone.inc became a customer this week). The thought experiment is whether an AI could run it, roughly what Sahil is doing at Gumroad but at much lower stakes. He'd start with the maintenance he doesn't do (updates, bugs, the backup that failed the other day), while Niklas argues he should start with marketing since that's where he does nothing at all. What Peter really wants is a layer above the pull request: something that fixes a thing, notices the fix broke something else, and either fixes it or rolls it back without any of it becoming his problem. Which leads to the open question he's sitting with: what building blocks would memo need before you could build the OG Kit operating system on top of it? Niklas came back from holiday wanting to step harder on the pedal, cutting runway from eighteen months down to thirteen or fourteen and hiring faster (more product engineers, probably an engineering manager, and a marketing and growth lead before the public launch). He also made a roadmap, felt dirty about it, and then rebuilt the Phone.inc website with Cursor, ui.sh and Astro on Cloudflare Workers, feeding it 60 screenshots of sites he likes plus a monorepo with the app and the docs in it. Then he had an agent work for eight hours generating 25 integration landing pages (Shopify, QuickBooks and so on), and during a call with a potential integration partner he had one built, deployed and sent over as a link right after. He also did an AMA on the Danish startup subreddit that got 193 comments, 80,000 views and around 60 waitlist signups. --- Still Early is sponsored by Phone.inc (https://www.phone.inc/): a business phone number, welcome greeting, call routing, and opening hours, all from an app on your personal phone.
28 Jul 2026
The business books we keep re-reading
In this episode: Summer special: Peter and Niklas each brought five business books off the shelf (plus a pile of honorable mentions) Positioning, knowing when to quit, OKRs, marketing channels, and how to stop fooling yourself about your own idea Why Getting Real still holds up twenty years later The Mom Test and Deploy Empathy as a pair (and why they'll probably kill your idea) Radical Focus vs Measure What Matters: the startup version of OKRs and the enterprise one A reading list for the summer, since Niklas is driving to Norway No updates this week (they already recorded once), so it's just books. Peter's five: April Dunford's Obviously Awesome (April was a mentor at TinySeed when Peter was in the accelerator, and his advice is to actually do the steps, not just read them), Seth Godin's The Dip (he re-reads it every time he's about to quit or sell something, to double check it's the right call), Poor Charlie's Almanack in the Stripe Press edition, John Doerr's Measure What Matters, and Traction by Gabriel Weinberg and Justin Mares (19 channels and the Bullseye framework, which Peter has read so many times he doesn't really need to open it anymore). Niklas's five: Build by Tony Fadell (his "get fired up" category, next to The Social Network and the Isaacson Jobs biography), Getting Real from 37signals (his startup bible, published in 2006, free on the website, and he's given away more than a hundred printed copies), The Mom Test, Inspired by Marty Cagan (his scaling bible, and the thinking behind getting engineers in front of customers at Trustpilot), and Radical Focus, which is basically the startup-sized version of OKRs. Both of them gravitate toward books that work like debugging tools. The Mom Test and Michele Hansen's Deploy Empathy pair up really well (one is short and hands you the questions, the other is dense and gives you the whole customer interview process). Peter's take: if you're reading The Mom Test and The Dip at the same time to solve the same problem, you probably should have read The Mom Test a lot earlier. Honorable mentions: Purple Cow, The War of Art, Scaling People, Derek Sivers' Anything You Want (Adam sent Peter four Sivers books in the mail when he started at Tailwind), The Entrepreneur's Guide to Keeping Your Sht Together* by Sherry and Rob Walling (Rob's TinySeed was the first investor in Peter's previous company), Business Model Generation, and Founder-Led Sales Explained, which Niklas is a third of the way through. On that last one, Peter also recommends Jen Abel's two episodes of Lenny's Podcast (in order). They've helped him a lot with the Tailwind CSS partner program, where he's the founding salesperson and pulls in Adam when a deal needs the founder in the room. --- Still Early is sponsored by Phone.inc (https://www.phone.inc/): a business phone number, welcome greeting, call routing, and opening hours, all from an app on your personal phone.
14 Jul 2026
Are building blocks the new end product?
In this episode: Peter's thesis for the AI era: build building blocks, not end solutions Memo.sh update: 43 waitlist signups without a launch tweet, and a new workspace instructions feature The value hacking phase (Mike Maples Jr.) and betting on yourself as customer zero Phone.inc's new infrastructure: deadline day, three smartphones, and a 50-bullet testing list Why the phone system should be primitives, and everything around it wide open The shorts factory experiment: 30-something shorts in an hour and a half Peter's been heads down on memo.sh (https://memo.sh). The waitlist is at 43 people, and he hasn't even done the "hey, this is what I'm building" tweet yet (just leaving the link in regular tweets, which keeps converting one or two people at a time). He also added a follow-up form after signup with three questions, and almost everyone describes the same problem they hope Memo solves. That's a good sign. Next step is a demo video (the new Osmo Pocket 4 gets its first job), then inviting the first users. He's leaning towards charging them from day one, mostly as a filter. He's thinking about all of this through Mike Maples Jr.'s value hacking framework: a startup is a group of people trying to productize a unique insight, and before you worry about growth, you nail the value. Peter's bet is that he's customer zero for Memo, and that today's AI power users look a lot like tomorrow's normal users. That leads into the big idea of the episode: in the age of AI, you should probably be building building blocks, not end solutions. Everyone can now assemble their own workflows with AI, so nobody wants something too opinionated. They want great, slightly opinionated blocks they can put together themselves. Peter's newest Memo feature doubles down on this: per-workspace instructions for the agents connecting to it, which turns Memo into a flexible storage and search layer people can build things with that he hadn't imagined. Niklas pushes back a bit (he still believes in super opinionated products, and thinks the blocks will need to get more opinionated as non-builders arrive), but they mostly land in the same place. Niklas is recording on the day of Phone.inc's internal infrastructure deadline, surrounded by three smartphones and a 50-bullet manual testing list (testing telephony is a project in itself). The new setup is built around conference calls instead of logical phone lines, which makes the whole call life cycle simpler and way more responsive. They're moving their own number over now and customers after the summer holiday. His conclusion fits the building blocks theme: the phone system itself should be a simple set of primitives, and everything around it (MCP servers, APIs, webhooks) should be wide open. One prospect on a recent call has written 450,000 lines of code with AI in six months and just wants his phone calls in that pipeline. They also debrief the shorts factory experiment (30-something shorts in 90 minutes), and Lea reports that the vlogs are changing Phone.inc's hiring: candidates show up already knowing the company. Niklas's takeaway: when you judge content like this, count customers, hires, partners, and investors, not just views. --- Still Early is sponsored by Phone.inc (https://www.phone.inc/): a business phone number, welcome greeting, call routing, and opening hours, all from an app on your personal phone.
30 Jun 2026
I bought a $1500 domain name
In this episode: - Peter bought memo.sh (a $1500 domain) about fifteen minutes before recording, mostly so he could announce it on the pod - What he's building: a shared memory layer for AI tools, so Claude, ChatGPT, Codex and the rest stop making him repeat himself - His latest solo retreat, the Suhm ApS "board letter", and why doubling down on network beat doubling down on software - Niklas is back from San Francisco energized, hiring another engineer, and pulling his ops hire's start date forward to Monday - Why terms and privacy policies actually matter when you're venture-funded (it's the due diligence, not just the text) - AEO/GEO: how comparison pages and FAQs are quietly how AI decides what to recommend --- Niklas makes a new podcast rule on the spot: if someone just bought a domain, that's where the episode starts. Handy, because Peter had spent $1500 on memo.sh maybe fifteen minutes before recording, partly so he could talk about it here. The idea came out of a recurring annoyance. He uses a bunch of different AI tools (ChatGPT on forest walks because the transcription's better, Cursor or Codex or Claude Code back at the desk) and they never share context, so he keeps explaining the same thing twice. On his last retreat he vibe-coded a fix he calls Memo: a shared space where his agents write stuff down and read it back later. Codex one-shotted a Laravel app off a spec he'd built with Fable, wired up to the MCP and Passport packages he worked on at Laravel, and it just worked. He used it nonstop for two days and it stuck. The waitlist is up now at memo.sh. Niklas thinks fixing context across agents is going to be huge no matter who wins, and got a Turbopuffer plug in while he was at it. The other big thread is the retreat. Peter does these solo (a hotel, ten thousand words of notes, Claude helping him spot the patterns) and writes himself a board letter to Suhm ApS, where the board is just him. The takeaway, again: every time he thought the opportunity was software, the thing that actually paid off was network and personal brand (the conference circuit, the speaking, working with Niklas, this podcast). OG Kit stays the perfect side project, not the main bet. Niklas suggests publishing a public version of the letters, Berkshire-style. Then it's Niklas's turn. He came home from SF fired up, mostly from seeing companies further along and wanting to pick up speed. They're wrapping a big telephony infrastructure project (important, but invisible to customers), hiring another engineer, and his new ops person now starts Monday instead of after summer. He also gets into why legal isn't optional when you're venture-funded: terms and privacy policies are the contracts that get picked apart in due diligence, and AI-generated ones come back to bite you. Plus a rant about the FCC robocall program sending him more spam than it stops. They close on AEO (answer engine optimization), following up from last time. The short version: AI recommends products off bite-sized comparison pages, FAQs, and testimonials with a clear source of authority. Tactics that are dead for SEO but work great for LLMs. Probably something both OG Kit and Phone.inc should build. --- Still Early is sponsored by Phone.inc (https://www.phone.inc/): a business phone number, welcome greeting, call routing, and opening hours, all from an app on your personal phone.
16 Jun 2026
Fix the inputs, not the output
In this episode: The Phone.inc iPhone app is out with its first users, and there's a real call on the platform every hour during Danish business hours Why Phone.inc decided to roll its own voice infrastructure (and how a senior engineer reimplemented a SIP library in an afternoon with Claude) Niklas on his biggest worry: scaling the commercial side, and his investors telling him he's trying to fix it too early Splitting the podcast and the vlog into two channels, and a short that hit 22,000 views Peter's partner-description workflow and the lesson that stuck: when the writing is bad, fix the inputs, not the output Getting OG Kit recommended by LLMs, and learning GEO from scratch --- Recorded remotely while Niklas holds down the fort solo (his wife is in China for the week, and he's flying to San Francisco right after she's back). Niklas kicks off with the big news: the Phone.inc iPhone app is out with its first handful of users. There's a Slack channel called Pulse that pings on every call, and these days there's a real call on the platform pretty much every hour during Danish business hours. The first version just handles making and receiving calls, but the goal is for the app to become the leading interface so you don't need the web app at all. He also shares a fun moment with Katrine, their designer, who doesn't have a Figma license and works directly with Claude instead. He nudged her to stop writing a detailed Linear ticket and just build the thing, and by lunch she'd shipped it. A few days later she messaged him at half past midnight saying she couldn't sleep because building was too fun. The bigger Phone.inc decision: they're rolling their own voice infrastructure instead of buying it. Fully loaded (staff, office, hosting) it's still 3-5x cheaper than buying from a supplier, which opens the door to something Niklas hasn't seen before, a freemium phone system. It pushes the public launch back a couple of months, but it's worth it. There's a great story here about SIP, the core telephony protocol. Good non-GPL open source SIP libraries for iOS basically don't exist, and the commercial licenses run $10-50k a year for old C code with patchy support. Their engineer Thomas, who has 25 years of experience but none in telephony, looked at it, looked at Claude, and reimplemented just the parts they needed in a few hours. It got both of them thinking about what open source even looks like now that you can reimplement the 40 lines you need instead of pulling in an 8,000-line dependency. Niklas then opens up about his biggest worry with Phone.inc: scaling the commercial side. He's completely confident they can get to hundreds or thousands of happy customers, but getting to millions requires finding a channel that really works, and that's the part that's always been hardest for him across every company he's built. His investors gave him good feedback that he's trying to fix it prematurely, before there's even a product, pricing, or a single US or UK customer. He reframes it as a channel-fit worry rather than a product-market-fit worry, since he's hit product-market fit in this exact problem space before. The conclusion he's sitting with: accept that commercial scale is a later problem, focus on the first 50-100 happy customers, and carry the worry a bit longer. On Peter's side, the content setup has changed. The podcast (this one) stays on the Still Early channel, but the vlog has moved to a new [Phone.inc](https://www.phone.inc/) branded channel, since everything they filmed for the vlog happened when they met up to work on Phone.inc anyway. Still Early is really for the people who already know them, while the shorts reach a much wider audience. One short hit 22,000 views and is still picking up views weeks later, which is why those branding impressions belong on the Phone.inc brand. Peter wants to build the muscle organically first, then test putting money behind whatever clips generate real interest. Over on the Tailwind side, Peter is shipping individual pages for each sponsor on the partners page. Writing ~70 partner descriptions by hand was never going to happen, so he built it with Claude, but the first attempts kept coming out slightly off. The research phase would latch onto something and overweight it, and trying to fix the writing afterwards just made things worse. The fix was to split it into two skills: one that researches the company and pulls quotes, where Peter manually deletes anything he can tell the AI will fixate on, and a second that writes the description from that clean input. With the right context, it one-shot the description seven times out of ten. The lesson: when the writing is bad, don't fix the writing, fix the inputs. Niklas connects it to the shift from prompt engineering to context engineering, and how he's never seen a hallucination when he hands a model a flawed call transcript and asks for a summary, because the input is right there. They wrap on OG Kit, which Peter's feeling excited about again (Geocodio and Laracon AU both started using it). It's barely making money right now because it has no distribution, but the obvious distribution channel is LLMs recommending it when people build websites, the way Tailwind, Resend, and Tally get recommended. That's pushed Peter into learning GEO (generative engine optimization) from scratch. --- Still Early is sponsored by Phone.inc (https://www.phone.inc/): a business phone number, welcome greeting, call routing, and opening hours, all from an app on your personal phone.
2 Jun 2026
LinkedIn broke my SaaS
In this episode: The OG Kit LinkedIn saga: three days of debugging why Open Graph images stopped rendering, and the weirdly simple fix Why paying for a tool like OG Kit is worth it (someone else deals with this stuff so you don't have to) Tailwind's revamped partners page and the thinking behind turning it into a filterable directory Phone.inc's infrastructure decision (build vs. buy), the team settling in, and the first TestFlight invites going out The case for outbound sales in SaaS, and why Peter is weirdly excited to make cold calls --- Recorded outside in Niklas' garden, with his robot lawnmower making a threatening cameo in the background. Peter kicks off with the OG Kit story he'd been promising. A customer emailed to say they thought OG Kit was banned on LinkedIn: images rendered fine when hosted as static files, but not when generated through OG Kit. LinkedIn is a big chunk of the audience, so this was stressful. Peter walks through three days of debugging, testing different domains and URL structures, talking to Claude, trying a Cloudflare worker proxy, and looking into Cloudflare for SaaS (which would've eaten about 10% of OG Kit's revenue in per-domain fees). The actual fix turned out to be much simpler: LinkedIn's bot was choking on his robots.txt, and explicitly allowing the LinkedIn bot at the top of the file made everything work again, on the OG Kit domain directly. No custom domains needed. Huge relief, especially since the whole point of OG Kit is to stay small and simple. It's also a good argument for why paying for a tool like this is worth it: someone else absorbs the random platform breakage so you don't have to. On the Tailwind side, Peter shipped a revamp of the partners page (formerly the sponsors page). The big idea is to make it work more like a directory, so you can filter the ~70 companies by category. That's better for people browsing, and it also makes the sponsorships more valuable and easier to sell: being the third email company on a filtered page beats being the 71st logo in a giant cloud. The page now includes a testimonial section with companies speaking to specific parts of the program, like Shopify on supporting independent open source, CodeRabbit on tailored partnership opportunities, Resend on brand promotion, and Vercel on direct access to the team. Next up (not shipped yet): giving each company its own page on the high-traffic tailwindcss.com domain. Peter is also candid about churn, since a lot of companies signed up during the January news cycle and some are now rolling off, so the focus has shifted to making the program more valuable for the ones who stick around. Niklas gives updates on [Phone.inc](https://www.phone.inc/). Infrastructure is still the big open question, with deep negotiations on pricing, terms, and build vs. buy. He's managed to cut about 75% off the price from the main supplier they're talking to. The team is settling in after onboarding a new colleague pretty much every week, and the first TestFlight invites for the iPhone app go out Monday, with Android close behind. Katrine, the designer, has fully ditched Figma and is now working directly in Xcode with Claude Code, shipping pull requests, which has been a wild but exciting shift to watch. They also built a first Shopify integration, so when a call comes in you get a voice prompt with context about the caller (like "it's Peter, he placed an order five days ago"), which fits their ICP of small teams where the person answering the phone is also packing orders, not full-time support staff. The episode wraps with a good conversation about outbound sales. Niklas makes the case that hardcore cold calling is unfashionable in SaaS but genuinely works, pointing to companies like FlatPay, Trustpilot, and Firmafon back in the day. Even at Phone.inc's fairly low price points, the unit economics might support a small calling team. --- Still Early is sponsored by Phone.inc (https://www.phone.inc/): a business phone number, welcome greeting, call routing, and opening hours, all from an app on your personal phone.
19 May 2026
My VC told me not to be cheap
In this episode: Peter's Tailwind ui.sh support inbox after inviting 30k people to early access The first vlog landed well on click-through but had a watch-time problem (probably because the clip we pushed was also the first half-minute of the video) OG Kit got hit by Meta's new scraper, which means another entry in Peter's Cloudflare block list (currently around 20 scrapers) Niklas onboarding three new full-time hires and getting a much higher ambition push from the new London VC than he's used to from Danish ones Telco infrastructure strategy musings: going one layer deeper in the supply chain makes phone numbers 27x cheaper and changes what commercial offerings are even possible --- Peter's week started a day late after a road trip to Germany with his oldest, which sounds nice but also means feeling behind for the rest of the week. Most of Peter's Tailwind work has been ui.sh support. They invited around 30,000 people to early access, so the inbox just doesn't really stop right now. OG Kit got hit by Meta's new scraper, which is how Peter usually finds out big companies are training models (Apple recently, now Meta). He has a Cloudflare setup a friend helped him build that lets him add new scrapers to a block list, currently around 20 of them. The big customers with hundreds of thousands of pages are actually useful here because they get scraped first, so the system gets hardened before normal users notice anything. Peter and Niklas recorded the second and third vlogs in Copenhagen (you can check those out on YouTube). The first one did well on click-through but watch time was lower than expected. Peter ran the YouTube analytics through Claude and ChatGPT and both pointed at the same thing: the clip they'd pushed hard on social before release was also the opening of the video, so people clicked, saw what they'd already seen, and bounced. Obvious in hindsight. The plan now is to release clips after the video, not before. Niklas's week has been onboarding new full-time hires, signing infrastructure contracts, and a first check-in with the new lead investor in London. Big takeaway: the ambition level from a London VC is genuinely different from what he's used to from Danish VCs. They pushed him to hire someone he wanted but couldn't quite justify yet (the answer being basically "if you'll need them in six months, hiring later will cost you more than the salary"). They also told him to take the most expensive serviced office option as long as it doesn't pull focus. The infrastructure piece is the fun part. Phone.inc currently builds on one provider that's great but premium (think AWS-tier pricing for telco). Going one layer deeper in the supply chain makes phone numbers around 27x cheaper and makes incoming call traffic free, but means building more in-house. That unlocks commercial models that don't otherwise work, like sending every new Danish company a free business number. The build-vs-buy decision is also a go-to-market decision. The episode wraps with Niklas asking Peter to start looking into paid acquisition experiments for Phone.inc. Peter has zero practical experience with paid ads but is interested, and the goal isn't to crack it, just to figure out where to even start. Phone.inc is sort of a prosumer product (selling to the business owner who just started their first company, not a procurement department), so the usual SaaS playbook probably doesn't apply. Instagram, TikTok, founder stories, visually compelling small businesses as mascots. They'll see what sticks. --- Still Early is sponsored by Phone.inc (https://www.phone.inc/): a business phone number, welcome greeting, call routing, and opening hours, all from an app on your personal phone.
5 May 2026
Building a phone company for the second time
In this episode: What Peter and Niklas are each working on right now (Tailwind, OG Kit, Phone.inc) Niklas hired his first employee and had to buy a laptop at Denmark's terrible excuse for an Apple Store Phone.inc's second paying customer and the challenge of marrying old telco infrastructure with a modern sign-up experience Peter's day in Copenhagen (coworking, Laravel meetup, the Royal Library, vibe coding a revenue dashboard) The content plan for Still Early going forward (podcast and vlogs) Peter and Niklas are back for episode two, this time recording remotely from their respective towns. They use this one to properly introduce what they're each working on and catch everyone up on the week. They start with a quick intro to what they're each working on. Peter works with Tailwind Labs, where the team is building ui.sh (https://ui.sh/), which is basically trying to figure out how to get AI agents to become good at design. Peter also runs the Tailwind partner program and builds OG Kit on the side, a SaaS tool for dynamic open graph images that he describes as the perfect side project (tiny revenue, but super fun and near to his heart). Niklas gives the backstory on Phone.inc (https://www.phone.inc/). He co-founded Firmafon back in 2009 (now called Relatel), which became the biggest SMB telco in Denmark. Around 10-15% of all Danish businesses use it today. But three things that didn't exist back then have completely changed: apps with built-in IP telephony, global telco infrastructure you don't have to build yourself, and AI that can actually access the contents of phone conversations. Phone.inc is basically Firmafon rebuilt from scratch as an app. The domain was just available, which is kind of unbelievable, and it was honestly the thing that made Peter get excited about the idea. The big update this week is that Niklas hired his first employee, a designer and product manager named Katrine who he used to manage at Trustpilot. It was his first day having a coworker in over a year, and the onboarding included a pretty rough trip to Denmark's closest thing to an Apple Store where they got the wrong laptop, couldn't get the phone they wanted, and walked out angry. Niklas has a message for Tim Cook. Phone.inc also got its second paying customer, someone from the waitlist who'd been banging down their door. The biggest challenge right now is negotiating with telephony infrastructure providers who want five days of compliance review before a customer can go live, which doesn't work when the goal is to have people up and running within a minute of signing up. Niklas compares the wow moment they're going for to Square's original product video, where suddenly your phone becomes a payment terminal. Peter's week included a trip to Copenhagen for the Laravel meetup (which he originally started back in 2014, now run by someone else). He stopped by Niklas' coworking space, bumped into a potential Phone.inc customer on the street, worked from his favorite spot at the Royal Library, helped send out 500+ UI.sh invites, and vibe coded some MRR tracking into Tailwind's internal dashboard. Got home past one, pretty tired. They also lay out the content plan for Still Early: podcast episodes every other week, vlogs on YouTube in between. They're getting help with editing and trying to keep everything as simple as possible. --- Still Early is sponsored by Phone.inc (https://www.phone.inc/): a business phone number, welcome greeting, call routing, and opening hours, all from an app on your personal phone.
23 Apr 2026
Why are we starting a podcast?
In this episode: How Peter and Niklas met and why their previous podcast attempts didn't work out What Niklas is building with Phone.inc (https://www.phone.inc/) and why distribution is the biggest challenge right now Founder-led marketing and content strategy in the age of AI Why vlog-style video is probably the future but still feels super cringe If you prefer video, subscribe to our channel on YouTube: https://www.youtube.com/watch?v=Hz2TxxEkW9k (https://www.youtube.com/watch?v=Hz2TxxEkW9k) Peter and Niklas have been trying to start a podcast for years. They met back in 2019 when Niklas was listening to Peter's podcast Out of Beta and asked if he wanted to give feedback on his startup Legal Monster. Peter came by the office, they hit it off, and Peter literally ended up moving into Niklas' office. They've tried starting a podcast together multiple times since then — including a Danish-language show called "Store Planer" that produced a few genuinely great episodes before the production effort killed it. This time around, Niklas is building Phone.inc (https://www.phone.inc/), a modern business phone system for small businesses and startups. He's got a first paying customer (who basically wouldn't stop banging on the door until he got access) and over 120 companies on the waitlist. But unlike Peter's work with Tailwind Labs, where the team is sitting on massive built-in audiences, Phone.inc is starting from zero on distribution. Peter is helping out as an advisor and small investor, and one of the things they agreed on pretty quickly is that founder-led marketing is the move right now. So part of Peter's role is helping Niklas figure out the content side — editing, promoting, keeping things moving week to week. Every week they just talk about whatever's top of mind — Phone.inc, Tailwind, Peter's stuff with OG Kit, the startup journey in general. No heavy research, no rigid scripts. Just two founders checking in every week and figuring out the format as they go. There's also an interesting tension around content strategy in the AI era. Phone.inc is still in semi-stealth mode — not ready for self-serve, carefully controlling who gets access. But a podcast generates thousands of words per episode, and those transcripts eventually get consumed by language models as they're trained. So it's a way to build awareness with both humans and AI simultaneously, without prematurely blasting out marketing copy on the website. They also talk about vlog-style content and how that's probably what actually works for discovery on YouTube. Niklas had a recent day full of potential content (office tour, lawyer visit, VC skybox event at a football match) and he had his camera in his backpack the entire time. Never pulled it out once. The cringe barrier is real, but they're working on it. Still Early is sponsored by Phone.inc (https://www.phone.inc/): a business phone number, welcome greeting, call routing, and opening hours — all from an app on your personal phone.
17 Apr 2026
Still Early Teaser
The first episode of Still Early is dropping on April 23nd. Subscribe now to get notified!
Contact Still Early
- Guest appearances
- Does not typically book guests
Based on episode analysis; this does not confirm that the show is currently accepting guests.
Host of Still Early?
Claim your podcast to manage its listing and keep your show details accurate.
Pod Engine is an independent podcast discovery and analytics service and is not affiliated with or endorsed by this podcast. Artwork and show content belong to their owners. Full legal notice.
Explore this show
Podcast research with Pod Engine