Parse notes
Sponsored
Delve ran 4 ads across 2 TLDR newsletters and brought in 66 leads, with a meaningful chunk from enterprise. The result: $1M in attributed pipeline and a 52x ROI.
Databricks announced a new funding round that valued the company at over $100 billion, surpassing last year's $62 billion valuation. CEO Ali Ghodsi revealed the round exceeded $1 billion, enabling investment in AI-driven products. Rival Snowflake anticipates $4.5 billion in annual revenue, with a market cap of $65 billion.
Saronic is positioning itself in the defense-tech sector with autonomous surface vessels, securing $600 million in Series C funding and targeting military contracts amidst a US Navy shipbuilding crisis. Despite some skepticism about its strategy, Saronic focuses on landing smaller contracts and developing prototypes to eventually secure a major program of record. The company's revenue projections for 2025 have jumped to $400 million. It plans to expand beyond the DoD to clients like the Coast Guard and international allies.
Founders often mistakenly use their investor pitch deck for customer marketing, ignoring that investors and customers have different priorities. Effective early B2B marketing requires tailored messaging that focuses on a specific audience's immediate needs, not the broad vision in a pitch deck. Translate pitch deck messages into clear customer-focused website content by addressing specific pain points, differentiating the product, and building trust with social proof. Ensure communication is relevant, precise, and customer-centric.
Six founders who went viral share how they turned product launches into millions of views and instant waitlists. The pattern across all winners? Show, don't tell, make viewers say "I can't believe this didn't exist," and treat launches as weapons, not your end goal. Silicon Valley now treats B2B software like entertainment products because everyone's fighting for attention. Early investors literally write checks based on viral tweets - social media has become social currency.
Wispr Flow pivoted from a hardware focus to a software-centric voice dictation platform upon realizing there was limited market demand for its initial product. It cut its workforce substantially and launched Wispr Flow, which quickly gained traction with strong user adoption and a high paid conversion rate. The pivot emphasized listening to customer needs and moving swiftly to achieve product-market fit.
This post drops four prompts that actually move the needle for product teams. The first prompt uses computer control models to crawl your actual product and build a feature taxonomy from the user's perspective. The second prompt turns boring product strategies into narratives that sell your vision. The third prompt frames every PRD within your overall strategy so initiatives aren't just random features, but steps toward your vision. The fourth prompt evaluates your plans like a brutally honest CPO who'll call out second-order effects and hidden risks. These aren't just prompts, but a framework for decomposing tasks and bringing expertise to get outputs worth shipping.
AI's intelligence layer has become a balance-sheet arms race—capital expenditures in the hundreds of billions. The door is closed for new challengers. Investment returns will concentrate on a handful of at-scale companies.
After three years in one apartment, this author discovered what digital nomads miss: real connection only comes from dealing with a place's boring parts. He spent years hopping between cities, thinking each Instagram-worthy location would make him happy. However, constant movement meant never belonging anywhere. The case for staying put is simple - you can't build meaningful relationships or impact when you're always leaving. The daily annoyances, the negotiations with neighbors, the commitment to local problems - that's what creates actual community.
Chroma's founder says RAG is dead because it bundles three concepts poorly - the real job is context engineering, figuring out what belongs in the context window for each LLM generation. The best teams now use a two-stage approach: first-stage retrieval (vector search, text search, and metadata) culls 10,000 candidates down to 300, then LLMs re-rank those to the final 20-30. This matters because context rot is real - despite perfect needle-in-haystack marketing, models degrade as you add tokens. The future looks radically different: models will stay in latent space without converting back to text and retrieve continuously during generation instead of once at the start. What we call "memory" is just context engineering with better branding.
The revelation that you can finally build every idea you've ever had becomes a curse.
B2B executives fall into two camps: the "Cracked," working 70-80 hours to build the next billion-dollar company, or the "Slow Roll," choosing balance but building nothing “extraordinary.”
While many competitors rush to integrate AI, Obsidian remains hesitant, valuing user privacy and secure data storage.