🏠 Own your AI or someone else will

GA! It's Brett.

We're halfway through 2026 and I genuinely cannot believe how much has changed since January. SpaceX went public. Anthropic's best model got shut down by the government three days after launch. The Knicks won in five.

Made a video about that last one. Three hours, no budget, just AI tools. Breakdown is below.

Also excited to share that Micro just landed its first official partnership. Slash customers now get 3 months of Micro free — if you use Slash, go grab it.

Sign up for Micro below and get 1 month free!

Posted this on June 10: watch it here

my mayor is muslim
my bagel is jewish
my cream cheese is chive
my Knicks in five

They did win in five, by the way.

Made a 26-second video to go with it. The whole thing took about 3 hours, no production team, no budget.

The process: generate the song in Suno, drop the audio clip plus reference images plus a prompt into Higgsfield (Seeddance 2.0) to generate 10-20 clips, put it all in CapCut and align it until it feels good, apply text and effects. How I made it.

A few things worth breaking down about why this works:

The poem came first. Four lines, two minutes to write. That's the brief — a specific cultural moment, a specific city, a specific feeling. The clearer the brief, the better every downstream tool performs. Suno needs a mood and a reference. Higgsfield needs visual direction. CapCut needs a rhythm to cut to. All of that comes from having a clear idea before you open a single tool.

The song Suno generated on the first try was about 80% there. I ran it twice more with small adjustments to the prompt and landed on one that felt right. Total time: maybe 20 minutes. Old way: hire a producer, wait two weeks, pay $500 minimum.

Higgsfield is where most of the time went — generating 10-20 clips gives you enough raw material to be selective. You keep the 5 that are great, throw out the rest. The Seeddance 2.0 model handles motion in a way that feels cinematic without looking artificial.

The creative unlock AI gave me isn't about having better tools. It's about the gap between idea and execution finally closing. I had the idea for this video in about 30 seconds. Three hours later it existed. That gap used to be months of learning, money, or both.

If you want to try it yourself: Suno for music, Higgsfield for video, CapCut to put it together. Start with something you actually care about. The taste part is on you.

Out Of Home advertising has long been effective but hard to scale—until now. AdQuick makes it simple to plan, deploy, and measure campaigns with the same efficiency and insight you expect from online marketing tools.

Marketers agree: OOH is powerful for brand growth, driving new customers, and reinforcing messaging. AdQuick makes it easy, intuitive, and data-driven—so you can treat real-world campaigns like any other digital channel.

Learn more, visit AdQuick.com

Satya Nadella said something this week that every founder building on top of AI needs to sit with.

The argument: companies must own their AI "learning loops" to compound human and token capital, or risk ceding all value to a handful of frontier models.

In plain language: if your product just calls OpenAI or Anthropic's API and returns the result, you're not building a business, you're building a reseller. The model gets smarter. Your product stays the same. Over time, as models improve, the value of what you've built approaches zero because the foundation model is doing more and more of the work.

The companies that win are the ones that use foundation models as a starting point and build proprietary feedback loops on top. Every interaction with your product teaches it something the base model doesn't know. Your customers' data, your domain context, your fine-tuning, that's the moat. Not the model itself.

The practical question for every founder: where in your product does data flow back in and make the system smarter over time? If the answer is nowhere, you're renting intelligence instead of building it.

The companies that figure this out in the next 12 months will be the ones that can't be replicated by a GPT wrapper. The ones that don't will spend the next five years watching the base models eat their margins., via @satyanadella on X

5 things on my mind this week:

@BullTheoryio: The UK wants to scan every photo, video, and message on every phone in the country. 10.6M views. Worth reading, but take it with a grain of salt. The proposal includes an opt-out for adults, which the post leaves out.

Neon corridor. Perfect symmetry. Pink to teal gradient. Looks like someone typed "futuristic hallway" into Midjourney.

Last week's poll: Real. 66.7% of you got it right. The 33.3% who voted AI were fooled by the perfect symmetry, the pink to teal gradient, the whole thing screaming "someone typed this into Midjourney." It was a real installation, shot with a phone. Real life keeps looking more like the model.

Micro is a CRM that doesn’t feel like homework and an email client that knows what matters. Simply put, it’s an all-in-one productivity tool that organizes itself so you don’t have to.

LHV is a community-driven, early-stage venture fund backing ambitious founders at the earliest stages.

If you’re building tools that help founders, investors and other professionals who want to be 100x more productive with 100x less work, and you’re interested in advertising with us, send an email over to [email protected] with the subject “Homescreen Ads”