Doug van Spronsen

For agentic systems founders and dev tools founders: People do not want to pay for raw markdown and they shouldn't have to. But they may pay for orchestration, hosting, updates, collaboration, portability, analytics, and managed execution. These can be great businesses.

@michaelmiraflor Gateway drug to the ecosystem

Who is this laptop for? Trying to wrap my head around the segment. For students? For people who want a second laptop? For people who wished the iPad had a keyboard? For white collar email warriors who travel a lot? It’s convenient but very under powered.

feels like such a wasted opportunity every moment your agents aren't running

👷Surviving the AI Capex Boom👷 Big Tech's AI buildout is transforming markets but history suggests caution. What should investors do? 🤖 AI Investment Boom 🚂 Echoes of Past Booms 📉 Rising Capex Firms Underperform 🏭 Magnificent 7: The New Utility? 🔎 Finding AI Early Adopters 🌊 Value Investing: Navigating Hype Cycles

@jliemandt @mbateman When will Timeback be in Canada 🇨🇦? 🙌🏻

I'm so excited that @mbateman is building a Montessori-based Alpha Timeback school. Working with the world's best educators to reimagine how kids should spend their first dozen years is the best part of this job.

@willahmed @willahmed love this. Will Canadian subscribers be able to upload data?

LAUNCHING TODAY: WHOOP ADVANCED LABS Many people only get bloodwork once a year. You receive a sheet of numbers, maybe a “you’re normal,” and that’s it. No context. No coaching. No understanding of how to move from acceptable to truly optimal. With WHOOP Advanced Labs, we’re changing that. Starting today, members can unlock two powerful ways to bring their blood biomarkers into the WHOOP experience: 1) Upload past bloodwork. Centralize your results from any lab, connect them with your 24/7 WHOOP data, and track long-term trends over time. This is included in all membership tiers for all WHOOP members and it will meaningfully improve the insights you get from all your data. 2) Book comprehensive testing. Through WHOOP, our U.S. members can now schedule curated blood panels that test 65 of the most impactful biomarkers, reviewed by a licensed clinician and seamlessly integrated into your app From there, WHOOP goes beyond a static PDF: -Context, not just numbers: With WHOOP, your bloodwork is no longer isolated. It’s integrated directly alongside your daily behaviors like sleep, recovery, training, and lifestyle choices. -From acceptable to optimal: Every result comes with clinician-reviewed insights that explain where you stand today and how you can actively improve each metric over time. When you fall out of range on a metric, we will coach you to not just get back in range but actually reach an optimal level for your gender and age. -AI-powered coaching: Our platform continuously analyzes your data and delivers personalized guidance, turning snapshots of lab results into a living feedback loop that makes your health dynamic rather than episodic. We’ve made significant investments in our A.I. infrastructure in the last six months, and you’ll see the benefits as your data comes together. Tell WHOOP what supplements you take, upload your blood tests, and await as the feedback gets much more useful. WHOOP Advanced Labs is more than a new feature. It’s a building block of a new Health Operating System that we’re building at WHOOP: a future where your body is continuously monitored across hundreds of dimensions, where feedback loops between biology and behavior are surfaced in real time, and where AI helps you take proactive steps to live longer, healthier lives. Already, more than 350,000 members joined the waitlist for WHOOP Advanced Labs. Today, the first members are gaining access in the U.S., with a global rollout to follow. Enjoy!

Substack’s feed product has evolved into old Twitter in a positive way that I didn’t expect. Worth revisiting if you wrote it off

It also allows the fintechs to build the profitable parts first before taking on the regulatory burden/cost of being the primary deposit taking institution (there is much less money in this) Banks work the other way, trying to increase share of wallet by selling wealth, credit etc

Interesting how trading platforms like Robinhood and Wealthsimple are expanding into banking. Feels like one of the smartest ways to build a neobank: start with trading, where users feel like they can earn money, then layer on banking. A very different relationship than traditional banks or neobanks, which mostly take your money.

This is on the nose. At the same time its also indicative of the current opportunity set available and people (rightfully) taking advantage of it. But it cant be culturally set as the "only" viable path because outcomes are never evenly distributed.

@GergelyOrosz 25 year head start = enterprise Lindy

How is it that Atlassian builds tools that: - SO MANY people use - SO MANY people dislike (JIRA, Confluence. Previously also HipChat) They are defying the conventional wisdom for business success that starts with "build something people will love"

“Please consider the environment before pasting this into GPT”

@patrick_oshag One of my all time favs. Vulnerable and inspiring

@Suhail Does an "always on" wearable from a frontier model co compliment or compete with AI-first browser? Different contextual memory, but likely more personal than what the browser connects?

Whoever owns the AI first browser will win AI memory long-term. The browser is the closest approximation of humanity's memory that we have.

@VibeMarketer_ Email

if you do cold email and are still writing emails manually in 2025, you're so behind. lets fix that... this n8n workflow writes and sends 10,000+ fully personalized cold emails for you: - scrapes ideal leads with apollo - uses GPT to analyze their site and LinkedIn - writes tailored outreach based on their business - sends the emails on autopilot using instantly built for freelancers, agencies, and leadgen killers want the full flow + setup SOP? comment “EMAIL”, follow, + repost and I’ll send it over (must be following so I can DM)

@kennandavison Icon

Introducing Icon, the world’s first AI CMO (Chief Marketing Officer): it can plan, create, & run 1000s of winning ads end-to-end. We're backed by Peter Thiel's Founders Fund & execs of frontier AI labs like OpenAI (ChatGPT), Cognition, & Pika. How it works: 1. Connect Icon to your Meta Ad Account & Google Drive. It’ll use your ad account data & existing assets for planning & making ads. 2. Icon finds 100 winning ad concepts daily. It does this by scanning websites (yours & competitors), competitor ads, customer reviews, ad account performance, & more. It comes up with 3 types of ads: "Competitor Clone," "New Concept," & "Winner Iteration." You can copy what worked for competitors, try fresh approaches, & scale your winners. 3. Icon turns winning concepts into ads better than ChatGPT. Say you’re trying to copy a competitor’s before-and-after style ad. ChatGPT would make up random before-and-after scenarios & get your product dimensions wrong. Icon has a reasoning layer on top of GPT-4o that deeply studies the ad for minutes. It understands that it needs to make before-and-after scenarios using real insights from your website, product descriptions, & reviews. Icon gets ad intent, while ChatGPT only sees pixels. 4. Launch ads into your ad accounts with 1 click for less than $1. Before Icon, you’d launch 10 ads per week. Now you can launch 1000 ads per week & get more winners. Everything in my life has led up to this. At 18, I had to support myself. I had to live with my parents to save money. And no one thought I’d go anywhere. I really hated this, so I worked 100-hour weeks until I destroyed coding interviews. Soon after, I dropped out to join Pinterest at 19. After being a minion at Pinterest, I quit cold turkey & started my 1st company Skio. I then fluked into Y Combinator as a solo founder & completely failed during the batch. Thankfully, pivoting worked well ($15M+ ARR & profitable). But it’s not over yet. I have a massive chip on my shoulder that I must fix with Icon. So far, so good: we went from $0 to $5M ARR in our first 30 days. I know I’m not supposed to say this publicly, but I want to make Icon the greatest company of all time. I want to deliver insane value to our customers. I want to create generational wealth for my team & investors. I want to break the $0 to $100M ARR world record. And I’m putting my money where my mouth is: I just bought icon.com for $12M. If you've read this far, I want to say thank you with a gift 👇 We have an internal Google Drive with 1000+ static winning ads & 1000+ winning video ads. These ads have driven >$1.3B revenue in the last 6 months. We trained AI CMO on these ads. I’m pretty sure the right person could make $100K+ by just copying them. Retweet this & comment “Icon” and I’ll send you the Google Drive link for free.

It's 2025 and most content is still written for humans instead of LLMs. 99.9% of attention is about to be LLM attention, not human attention. E.g. 99% of libraries still have docs that basically render to some pretty .html static pages assuming a human will click through them. In 2025 the docs should be a single your_project.md text file that is intended to go into the context window of an LLM. Repeat for everything.

@scottastevenson I think this is a function of utility vs novelty. High utility / integration / ecosystem usually wins over aesthetics, even if you are aesthetically driven

Getting sick of Perplexity in the same way I got sick of Arc:

@brunoswerneck @Plaid Is this feature coming to Canadian merchants who integrate with Plaid?

1/ At @Plaid we're seeing strong traction with pay by bank, particularly around bill pay use cases More info here > plaid.com/pay-by-bank/ We wrote about this here > plaid.com/blog/pay-by-ba…

It seems we are still treating AI like a tool for thought, instead of a tool for self sufficient exploration/agency. Innovation requires the right questions to be asked and for the most part, humans are still asking old questions in this cycle (it’s even embedded in the UI itself). I think this will look odd in retrospect.

When a big new thing appears, we tend to think of it initially in terms of the old thing. For example, people thought of cars and trains as mechanized carriages initially, before starting to see them as their own thing. How are we doing this with AI?