Philipp Tsipman

I gained 30 pounds in building my previous company. And let’s just say, it wasn’t muscle. So this time around, I decided to do it different. Not gonna turn this into a food blog. But will post some pics along the way. Let’s call it my Cooking for Founders series Let’s go win this thing. 😊✌️

21 July

Hi friends - I'm launching Outfox. If you do marketing for a living, Outfox should hopefully become your best work sidekick. :) At launch, it can deeply understand your website and brand, and analyze your Facebook and Instagram ads with incredible insights.

21 July

Raw Claude is only 90% accurate on marketing tasks. We built the framework and guardrails to boost marketing analytics accuracy to over >99%. If you know anyone who is running Facebook ads right now, please tag them or send this to them. We have a free tier to get started.

outfox.aiOutfox: Your growth marketing sidekickYour growth marketing sidekick — with creative strategy, ad generation, and performance analytics, Outfox helps you build winning ads faster.

21 July

This is the platform I've always wished existed. Outfox.ai -- welcome to the world! 😁

claude.aiBrowser Automation Evolution: From ChatGPT to Claude CodeExplore how browser automation frameworks shifted from thick wrappers to thin CDP tools. Compare Playwright, browser-use, and the new browser-harness approach for AI-driven web tasks.

20 Apr

Heya, productivity and automation nerds! Browser harness got released today. It’s the new kid in browser automation and it shifts the state of the art. You wanna check it out! Link down below. Here are the deets: claude.ai/public/arti...

20 Apr

Announcement — x.com/gregpr07/st...

19 Apr

“SaaS is dead, just vibe code it” — cool take, until your marketing AI agent blows an extra $1M on Meta ads overnight. 👀 Take something simple: running online ads. As a SaaS vendor, I’m obsessed by three things: • Performance — are our insights actually the best in the industry? Why?

19 Apr

• Security — the moment you log into Meta through us, every access token and cloud sandbox becomes ours to keep locked down. Same way your bank sweats every dollar in your account. • Guardrails — no “oops, Claude hallucinated the math and overspent by 6 figures”

19 Apr

Could your marketing team or agency vibe code a Meta ads AI themselves? Heck, absolutely. But I’m obsessed with it. And that obsession is the product.

As I shift focus within AI for marketing, I just wanted to look back at why AI for video marketing hasn’t panned out the way I thought it would. (Or, at least, hasn’t panned out yet!) I think it’s two things — 1. Tech 2. Market

So, net net, we are in this transitional moment where video AI is much more alchemy than automation. And much more creator economy than marketing. I believe that will change in the future. And I’m still very hopeful on how video AI evolves.

But for now, it's a market that others deserve to win more than I do.

But, however you think of it, when marketing has visible AI artifacts — it does remind people of arrogance, fears of job loss, and of Dead Internet theory.

So, the more valuable brands — brands where I have a sales advantage after working in marketing in New York for many years — are the ones most hesitant to be associated with AI.

That’s 12x-150x faster than humans can read. It often feels instant! Good video is much slower and clunkier. So, three years into this gen AI revolution — with video we are still at the very early stages. It's both here and not here yet! Good enough to fill your social feeds.

Not good enough to create outstanding ads yet. Market ------- But tech is not the only thing. To get here, the AI labs have run roughshod over all sorts of ethical and social norms. One can point to Chinese AI labs respecting copyright even less so.

b) AI can follow some directions much better than others LLMs have made huge leaps in 12 months in using tools, working like an agent, and reducing hallucinations. So, when you do research or code, AI can now run for 20+ minutes without making mistakes.

To stitch together and edit a video well, however, you have to pick the exact frame and exact right position within it. If AI is a great wench, video automation really requires a scalpel. c) Time and cost Not least is time & cost. AI can now generate world-class code at 50 - 700 words a second.

And at the same time, not be able to tell that if a person morphs into a tennis ball, there is something really wrong here. Human visual perception is so basic and yet incredible. It’s also amazing how AI can create *images* with such better taste than it does text, design, or video.

I think it's the training data. AI visual art often blows me away. But most AI copy sounds like slop. AI-generated websites are very meh. And AI is really good at generating videos in the style of Pixar, memes, and UGC — but not great at cinema using humans yet.

Tech ----- My aha moment for gen AI was first seeing Midjourney in 2022. You could tell even in those first, janky images that, within a very short time, this tech would be transformative to human creation. It's been fascinating, however, to see how uneven AI’s progress has been.

The most important missing parts still are that: a) AI doesn't reason well visually or have good taste It’s wild to see AI acing tests that are really hard for humans — programming, law, medicine.

RIP digital ads. Long live AI agentic advertising. So AI is about to change marketing. It's been bubbling up for 2 years. It's here now. And I'm shifting focus to fully bet on it. It's a bit of a wild bet. Here's why: :)

No matter how badly your ads are doing, Meta will show up to tell you to spend more! Oh great. I’m bullish on a new set of AI platforms that can play on the side of the advertiser. Not just DSPs — but platforms that can act across walled gardens, making sense of the entire marketing spend.

In fact, I'm so bullish that I'm founding a new company for it. More on this very soon. 🙂

It's less about arcane ad auction knowledge. And more about the true nature of marketing -- who, what, where, why. This new world is... also great for marketing tech. For the last 10 years, ad tech was confined to the shrinking open web. It no longer is.

There is a huge opportunity to add on top of what the curated platforms will do. If you look at Meta’s AI recommendations — firstly, they’re badly executed... struggling with basics like brand voice and pacing. And secondly, to be real, they’re built to lie.

Some are saying that ads are dead — that our robot butlers will order everything for us, no storytelling required. And, that Meta will just auto-generate ones you need. Let's go home. My bet is the opposite.

In a world that's ever more complex, we crave things that feel trusted, new, and meaningful. Shopify sales are up 30% in a year. Unique products and storytelling are more important than ever. And, for marketing teams at smaller companies, the curated web is actually easier to navigate.

Picture 2027, the open internet is dead. The only forms of entertainment are watching brainrot videos and chatting with your AI companion -- who really gets you. Screeeeech. Bump. No, screw that. The world is much richer and more interesting than that.

But the “open web” IS disappearing before our eyes. Traffic to top news sites is down up to 50% this year. The curated, "AOL 2.0" Internet vibes are very very strong. So what does that mean for ads and commerce?

9 Nov 2025

I’m about to hit 43. It feels very different now. Health is now a big focus. I work out 5 days a week. Shifting my diet to plant-based proteins (fun!) Gotta outcompete all the 20-year-olds. 🫡😃

6 Nov 2025

Gorgeous fall day in NYC. Come visit! 🍁😁

Startup story time! ✈️🌉 On August 11, I hit SF for @camcorderai.bsky.social 🌸📹 first fundraise. Just under 3 months post-pivot. The $3M qn - who will own the next wave of video AI for marketing? 😅 Here's our story. Watch us sprint into it: #buildinpublic 🧶

4. 💸 Money: The finest VC is your revenue. :) Microsoft was our first giant win -- it flipped the game. But in AI, VC is the 🚀 fuel. So much ahead here.

It's gonna be wiiild month.😅 Luckily, startups aren't built in a month. Excited for this sprint and for what's ahead! Onwards. Thank you for being a part of this. 🙏 😎

This flipped our product. It's now video intelligence as much as creation. We built the best on the market Deep Search for video -- our benchmarks are off the charts. Next up: a marketing agent that can think, plan, and use hundreds of tools. “Toto, we’re not in Kansas anymore!” So stoked 😎

3. 💪 Team: I was lucky to get Justin to join as co-founder 1 year ago -- the stars aligned. I was lucky to meet Pavel through Twitter. I was lucky to bring Lewis, who I worked with 10 years ago, to pitch in. We're hiring for growth marketing, AI eng & frontend. Small team, big impact, join us. :)

What’s next on sales? • Focus on mid‑tier for the rest of the year and own it. • Then go for even smaller, local shops. • Finally, swing back up -- and get Target, Nike, MLB, HBO Max, Comcast over the finish.

2. 🚢 Product/Eng: Video AI is now in overdrive. :) When RunwayML launched 2 years ago, this felt imminent. But it took until 2 months ago for the big leap: agents turned real and video gen got finally good.

Startups learn through speed. So, in January, we pivoted hard. Now, building for creative teams at mid‑size consumer brands and agencies. This challenged us in every possible way. I've built and sold into enterprise for most of my career. @justinjware.bsky.social‬ is a consultative salesperson.

Here, the game is a product with a one-click "wow". And white‑glove care for small, early fans. So grateful for the brave customers who bet on us first. 😅

around... it's really fascinating to me that we have a new magical computer and it's helping me boil an egg.

It's not helping the government do something really crazy like some military ballistics or some special technology. Indeed, corporations are governments are lagging behind the adoption of all of us, of all of these technologies. So, it's just backwards."

1. 🎯 Customers: First, why pivot? What did we discover in market? I was deeply wrong on how quickly enterprises will adopt AI. They've moved much slower than smaller co's. This tech wave is so different from most.