ON BRAND DESIGN WITH AI
A thread on how you can use GPT-4o's image generation to create professional branding.
Full-text search across article titles and summaries. Refine with sector, year, content type, and reading-time filters.
A thread on how you can use GPT-4o's image generation to create professional branding.
The most powerful lesson from Naval Ravikant's recent interview about Elon Musk is around his willingness to look foolish and return to zero.
Founders who chase looking busy instead of making real progress will sink their startups in record time, whether it's making perfect Notion docs instead of shipping product, saying yes to every user request without having a clear vision, trying to tackle too many markets at once, building flashy extras before core features work, or trusting spotty early data over gut instinct.
Sponsored
Moving to UBP isn't just a pricing change—it's a fundamental shift in how revenue is generated, products are designed, and teams are incentivized. Learn exactly how to do this from CEO and co-founder Scott Woody in Metronome's new usage-based pricing playbook [4]:
The venture capital market has changed since the 2022 boom. The median LP count for funds between $100-250M dropped nearly 50% (from 83 to 47). Remaining LPs now write larger anchor checks. Performance data shows declining returns across all vintages—funds from 2017-2020 have lower median IRRs, with the 2017 vintage falling from 16.8% in Q4 2021 to 12.0% by Q4 2024. Smaller funds show better top performance (2018 vintage funds under $10M have 4.03x TVPI at the 90th percentile versus 1.67x for funds over $100M), though many funds still struggle with distributions—half of 2018 vintage funds have yet to return any capital to investors.
AI startups grow faster but fade faster too. They can reach $50M ARR quickly but might shrink in year four as their lead vanishes. Unlike past tech eras where physical infrastructure (on-premise) or engineering teams (cloud) created barriers to entry, AI tools make it easy for competitors to catch up at any point. This creates a new risk curve where impressive growth numbers no longer guarantee lasting success as technology keeps shifting under everyone's feet.
CoreWeave, an AI infrastructure provider, priced below its target range at $40 per share. It raised $1.46 billion by selling 36.5 million shares, but missed its initial $47-55 range after concerns about datacenter overcapacity dampened enthusiasm. Meanwhile, inflation ticked up with February PCE hitting 2.5% year-over-year and core PCE reaching 2.8%—both above expectations and worrying given upcoming auto tariffs.
Small AI startups now achieve hypergrowth without massive funding or hiring sprees. Four ingredients make this possible: intense demand (when markets pull your product because it solves urgent problems better), a repeatable unit (one customer type, one product, one process), scalable pipeline (one reliable lead source plus word-of-mouth), and compression (constantly making your customer journey 50% faster and easier). Master these elements and you can build an efficient growth machine with minimal overhead.
This post explains how to fix pricing as you add products. Figma moved from charging separately for each product to a simpler system with three seat types based on user needs. This solved a common founder challenge—how to expand your product line without creating pricing chaos. The new model (Collab, Dev, and Full seats) kept things clear for customers while enabling a 20-35% price increase for power users and driving the adoption of newer tools like Slides.
Despite available VC dry powder, 2025 may see another wave of failed startups due to fundraising challenges. Startups will struggle to gain attention, especially non-AI companies. Early-stage founders often neglect effective storytelling. Founders must articulate compelling narratives to differentiate themselves, drive investment, and build credibility.
Pieces is a Long-Term Memory Agent that remembers everything you work on, capturing and resurfacing workflow details across your desktop.
Greta helps you build full-stack applications designed for growth—no coding required.
Quadratic is an AI-powered spreadsheet with built-in code and data connections for instant insights and interactive charts.
Stop drowning in team confusion with five practical fixes: cut unused "strategy layers" that nobody references, design simple formats for sharing updates, create different dashboards for different roles, accept that some teams move faster than others, and pick frameworks that show real problems instead of hiding them. Teams that follow these rules make better decisions because they see what's actually happening instead of getting lost in complicated planning documents that nobody reads.
The first shareholder letter sets the tone for a company's entire public journey. Steve Jobs emphasized Pixar's focus on "pictures, not quarters," Larry Page warned Google investors about "lumpy" results from long-term bets, and Jeff Bezos declared Amazon would sacrifice "GAAP accounting" for future cash flows. These letters reveal what truly matters to founders—Jamie Dimon's demand for "the highest standards" at Bank One and Steven Rales' insistence on "simple, understandable businesses" at Danaher showed their strategic DNA before success proved them right.
AI startups are shattering traditional SaaS growth models with exponential scaling. They achieve this through "Experimental Revenue," leveraging rapid technology evolution, novel applications, and customer behavior changes. Successful AI companies are now marketing-led from inception, focusing on transforming experimental revenue into stable production use while simultaneously running multiple growth initiatives.
Dropbox, now highly profitable post-IPO, is shifting focus to enterprise with its new Dash product while maintaining its mature file-sharing business.
Klarna has filed to go public, highlighting its evolution from a BNPL pioneer to a comprehensive commerce network with a strong focus on AI-driven efficiency.
Sponsored
Shiny demos are one thing. But getting an app or website to work, look, and scale in the way you envisioned is not something you can one-shot with a prompt.
The Bot Company, co-founded by former Cruise CEO Kyle Vogt, raised $150M in a Greenoaks-led round to develop robots for household chores. The startup launched in May with initial funding from prominent tech figures. This follows Vogt's resignation from Cruise after a pedestrian incident involving a self-driving car.
TechCrunch's accusations against AI SDR company 11x point to startup metric problems rather than clear fraud. Counting canceled contracts as ARR and displaying unauthorized logos appears problematic, but likely stems from poor processes rather than deliberate deception—especially with both lead investors publicly defending the company.
23andMe's journey from biotech darling to bankruptcy reveals the fatal flaws of building on novelty. Despite reaching a $6 billion valuation with its popular DNA testing kits, the company's business model collapsed due to four critical failures: its product was a one-time purchase with no repeat customers, a devastating 2023 data breach exposed 7 million users' genetic information, unclear privacy policies eroded consumer trust, and its attempted pivot to therapeutics failed to deliver a viable second act.
Raising a pre-seed round demands both strategy and mental toughness. After 100+ investor meetings in just 10 days, George Matelich found that thorough preparation is essential, founder intros are gold, and scheduling meetings close together creates momentum that turns into deals. The process involves many rejections, occasional elitist attitudes, and the reality that many "pre-seed" funds actually expect traction—making it crucial to position yourself as selecting partners rather than seeking handouts.
VCs don't fund ideas, they fund metrics. While investors love compelling stories, they rely on concrete numbers to gauge whether a startup is scalable and worth funding. The most critical metrics vary by stage—early investors focus on market potential and retention, Series A demands sustainable growth with $3-5M ARR and 100%+ Net Dollar Retention, while later rounds scrutinize efficiency indicators like burn rate and gross margins. The metrics that matter most also depend on your industry: SaaS investors prioritize ARR and NDR, marketplace VCs watch GMV and take rates, consumer apps need strong DAU/MAU ratios, and fintechs must show transaction volume with regulatory compliance.
B2B SaaS companies routinely sabotage their pricing pages through avoidable mistakes. Hiding prices behind "Contact Sales" buttons turns away 83% of procurement teams and excludes budget-conscious buyers who could become advocates, while poorly designed freemium plans fail to showcase core differentiators that enterprise evaluators need to see. Feature comparison tables create additional problems when they're optimized only for human eyes (making them invisible to AI procurement tools) or use the confusing "everything from lower plan, plus..." format that forces potential customers to read through multiple tiers to understand what they're getting.
MGX is a multi-agent AI dev team that follows real software SOPs, letting you collaborate with AI roles to build websites, apps, and more.
Teamble AI improves employee feedback on Slack and Teams, boosting learning and performance.
Reach simulates your LinkedIn audience, letting you test and improve content before posting.
The word "but" negates whatever comes before it, which is why most communication advice suggests avoiding it. Simply invert your sentence structure—instead of "The positive thing, but the negative thing," try "The negative thing, but the positive thing." This subtle change lets you speak directly about improvements while ending on a positive note, making your feedback more motivating and allowing you to disagree while still sounding collaborative, especially in situations where you need to balance honesty with maintaining relationships.
Marketplaces have been connecting buyers and sellers since ancient Sumerian cities 5,000 years ago, evolving from physical spaces to digital platforms that reshape industries by solving specific pain points. Each successful marketplace begins with an unmet need and builds trust through transparency, unique inventory, or superior discovery. While some fear AI will replace marketplaces, it's more likely to enhance them through better matching, simplified onboarding, reduced service costs, and improved fraud detection.
TSMC has proposed a joint venture with Nvidia, AMD, and Broadcom to operate Intel's chip foundries, following a request by former President Trump. The proposal, which faces potential challenges, would be contingent on approval from Intel's new CEO Lip-Bu Tan, who is committed to keeping Intel's design and manufacturing integrated. Tan's appointment has boosted investor confidence, although Intel's turnaround remains a significant challenge.
True distinctiveness comes with a cost.
The trendy "GEO" (Generative Engine Optimization) rebrand misses the point by continuing SEO's worst mistake—optimizing for engines instead of users.
The majority of unicorn founders come from tech giants, elite universities, and military organizations, with many being serial entrepreneurs or having backgrounds in tech roles like engineering.
Sponsored
INTERVIEWER: Tell us about a challenge you've encountered at work.
Founders who maintain direct selling involvement throughout early growth stages gain market insights that are impossible to delegate. The most successful vertical software startups treat founder-led sales as a superpower that shapes product development, refines ICP targeting, and establishes repeatable motions, avoiding the common "product-market fit mirage" where promising initial traction tempts teams into premature scaling decisions that compromise long-term success.
Google DeepMind's new "thinking model" represents a leap in AI reasoning. Gemini 2.5 Pro has topped benchmarks with its ability to analyze problems before responding, showing superior performance in math, science, and coding without costly test-time techniques while maintaining Gemini's strengths in multimodality and expanding its context window to 2 million tokens.
In our information-saturated era, we're witnessing a paradoxical phenomenon where people voluntarily surrender their informational edge by making themselves increasingly easy to understand. Whereas information was once the highest form of alpha, accessible only through extraordinary effort (like scaling cliffs to decipher ancient inscriptions), today we're incentivized through game theory to broadcast our best ideas, trading secrets for attention in a world where knowledge is commoditized and standing out requires making yourself hyperlegible to algorithms, investors, customers, and peers.
Despite the hype around AI-driven sales, the reality shows more nuanced results. While 48% of teams use AI for GTM and it can boost outbound ROI by 30-50%, only 4% of companies rank AI-driven outbound among their top pipeline sources, and 46% of customers still prefer human agents even when bots are faster. The most successful implementations use AI to handle routine tasks (with chatbots resolving 70% of conversations autonomously) while keeping humans for complex interactions.
Setting realistic sales quotas requires looking beyond spreadsheet projections to what a human can actually accomplish. Quota design must account for the operational constraints of rep capacity, as demonstrated by a cautionary tale where an $800K quota translates to 60 opportunities requiring 360 total meetings per year or 7-8 meetings every week—before accounting for internal syncs, admin work, and the basic reality that reps only have 4-5 hours of actual selling time per day.
AI tools can now pick and use third-party services on their own, creating a major shift in B2B infrastructure. Frameworks like Model Context Protocol (MCP) allow companies like Supabase to grow rapidly by making their services AI-friendly, raising key questions about how founders should design products for AI selection, what makes a service AI-attractive, and whether we're seeing the birth of a new business model where AI systems become primary buyers.
Rapport brings AI-driven avatars to life, enabling real-time voice conversations with ChatGPT and other AIs on a scalable cloud platform.
Lemni lets you deploy AI agents for customer interactions across calls, emails, and outreach, scaling support without increasing headcount.
Platus is an AI-powered workspace for legal teams that offers a chat assistant and AI agents backed by your company's knowledge base.
Movie theaters didn't randomly set their popcorn sizes and prices—they're strategically using the Goldilocks Effect to guide you toward the medium option. Research shows we naturally avoid extremes when making decisions, which is why companies like Williams-Sonoma doubled sales of its $279 bread maker simply by introducing a $429 version, while streaming services and SaaS companies design their three-tier pricing to make the middle option feel "just right" through visual emphasis, social proof badges, and carefully calibrated price jumps.
Silna launched with $27M in funding to tackle healthcare administrative waste through its Care Readiness Platform. The solution optimizes insurance workflows like prior authorizations and benefit checks, significantly reducing delays in patient care. Already deployed in 45 states, Silna is a leading platform transforming administrative processes for healthcare providers.
Emergence has launched a $1 billion fund, Emergence VII, to support founders innovating in the AI and workplace technology space. The firm has a strong track record, with more than 1-in-10 early-stage investments going public and substantial realized returns. Emergence is notable for its hands-on and deep partnership model, which supports notable companies like Salesforce, Veeva, and Zoom. The company is now focusing on AI-driven startups like Together.ai and Unify.