GA! It's Brett.
Dropped an essay yesterday on how Claude, Codex, Cursor and Micro are competing to become the super app of the west. TLDR - whoever figures out context will win it all. You can read it here.
Speaking of super apps, we just released an amazing feature in Micro for heavy LinkedIn users that has never been done before. Can’t say more but you can see for yourself in Micro (sign up for a free month below).
Got this forwarded? Join 35K+ founders & operators who start their week with us.

🔥 Launch videos are no longer optional if you want to win
a16z New Media just turned one year old and published a piece on what they've built and learned. The line that stuck with me:
"Launch Videos are a new (and we think, enduring) format for founders to make a statement at a critical moment, that's built for this new media environment where you don't have to ask permission to get on someone else's show or op-ed page: you can just make your own statement, however you like, and share it widely."
That's the whole thesis. You used to need a journalist, a podcast invite, or a PR firm to get your story out. Now you need a camera and a clear point of view.
The piece explains what a16z New Media actually does — in-house video and editorial teams, owned distribution channels with a million X followers and a quarter million newsletter subscribers, a concierge portco services team, and a talent network. The part worth sitting with: they're not just helping founders make content. They're helping founders win what they call the "battle for preferential attachment" — getting the right people to want to attach themselves to you, your idea, your product.
The insight underneath that: signal comes from people now, not brands. A founder with a great launch video and a point of view outperforms a company with a polished press release every single time. The founders who get this right build an audience that compounds. Every video adds subscribers, followers, brand. Every one that doesn't get made is a missed opportunity to tell your story before someone else tells it for you.
The piece also has a line worth sitting with: "One great launch video isn't a media strategy. Relentless volume is." Building a brand that compounds means showing up constantly, not just nailing one launch day.

Your prompts are leaving out 80% of what you're thinking.
When you type a prompt, you summarize. When you speak one, you explain. Wispr Flow captures your full reasoning — constraints, edge cases, examples, tone — and turns it into clean, structured text you paste into ChatGPT, Claude, or any AI tool. The difference shows up immediately. More context in, fewer follow-ups out.
89% of messages sent with zero edits. Used by teams at OpenAI, Vercel, and Clay. Try Wispr Flow free — works on Mac, Windows, and iPhone.

💥 Google just dropped its strongest model ever
Google released Gemini 2.5 Pro Deep Think and the benchmarks are genuinely strong.
The headline number: Gemini 2.5 Pro Deep Think scores 88.5% on GPQA Diamond, the hardest graduate-level reasoning benchmark. That's above GPT-5.5 Pro and above Claude Fable 5. On FrontierMath Tier 4 — the hardest math problems in existence — it scores 41.2%, also above both competitors.
The key features: a 2 million token context window (versus GPT-5.5's 200K), Deep Think reasoning mode that shows its work step-by-step, and native multimodal capability across text, images, audio, and video in a single call. It's live now on the API, AI Studio, and Vertex AI.
The context: this is Google's direct response to Claude Fable 5 and GPT-5.5. Both launched in the last two weeks. Gemini 2.5 Pro Deep Think benchmarks above both of them on the hardest evaluations. The model race has never been closer — or faster.
The question that matters for founders: does winning benchmarks translate to better outputs for your actual use case? The honest answer is usually "sometimes." The 2 million token context window is the most practically useful feature for anyone building on long documents, large codebases, or complex agentic workflows. That gap is real and it matters.
🧠 ON MY MIND
5 things on my mind this week:
OpenAI launched GPT-5.5-Cyber and Patch the Planet on June 22. GPT-5.5-Cyber scores 85.6% on CyberGym — the hardest offensive security benchmark. It's available under OpenAI's Trusted Access for Cyber program for vetted security teams. Patch the Planet, co-founded with Trail of Bits and HackerOne, uses the model to find and fix vulnerabilities in widely used open-source software. This is OpenAI's direct counter to Anthropic's Project Glasswing. Both companies are now racing to own the AI cybersecurity market. — Cyberpress
Wikipedia refuses to let AI edit articles. Co-founder Jimmy Wales told AFP this week that Wikipedia doesn't trust AI enough to let it play a direct role in editing. The reasoning: AI hallucinates, Wikipedia requires citations, and the combination is dangerous at scale. This is the most important institutional "no" to AI in recent memory — and a useful reminder that trust has to be earned, not assumed. — LLM Stats
SPCX fell 10% in week two. SpaceX's stock dropped from its first-week high as early investors took profits and Nasdaq-100 inclusion uncertainty weighed on the stock. Still up significantly from the $135 IPO price. The real test: whether the stock holds through the Anthropic and OpenAI IPO windows, which will compete for the same institutional capital. — AIToolsRecap
Fable 5's free window expired last night. Claude Fable 5 was free for Pro, Max, Team, and Enterprise subscribers through June 22. Starting today, all usage requires consumption credits at $10 input / $50 output per million tokens. The catch: the model was offline for six of those thirteen days due to the government-forced shutdown. Anthropic hasn't announced a refund or extension. — AI Update
SpaceX just acquired Cursor for $60 billion in stock. One week ago SpaceX announced an all-stock acquisition of Anysphere, the company behind Cursor — the largest acquisition of a venture-backed startup in history. The rationale: xAI's Grok has failed to get developer adoption in the coding market, while Cursor generates $4B in annualized revenue and reaches two-thirds of the Fortune 500. SpaceX and Cursor have already been jointly training a coding model on the Colossus supercluster in Memphis for months. Deal closes Q3 2026. — Build Fast with AI

👀 FROM THE FEED
Google just invested $75M into A24 to build AI tools for movie production. @DiscussingFilm


❓ AI GENERATED OR NOT


Last week's poll: 50% of you got it right. 33.3% voted Both, 16.7% voted Left. The right looked AI and it was. The thing is, real photos are getting harder to tell apart.



Homescreen is a collaboration between Micro and Launch House Ventures
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.
HOW WAS TODAY'S NEWSLETTER?

REACH 35K+ FOUNDERS, INVESTORS & OPERATORS
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”


