Full-text search across article titles and summaries. Refine with sector, year, content type, and reading-time filters.
Getting value from data is still incredibly hard, making data businesses terrible VC investments. Auren Hoffman predicted exponential demand, but fewer than 100 hedge funds buy alternative data today. After 1,000 SaaS unicorns, there's been one DaaS unicorn: ZoomInfo, which never took VC. Data companies grow slowly with predictable revenue - perfect for PE, not VCs needing explosive growth.
OpenAI's latest experimental reasoning model solved 5 of 6 problems on the 2025 International Math Olympiad, earning 35/42 points. The model won't be released for several months.
Anysphere acquired AI-powered CRM startup Koala to develop a dedicated enterprise-readiness team as it battles GitHub Copilot for large Fortune 500 contracts.
Working harder on the wrong thing is why you're stuck - systems move at the speed of their slowest part.
Good investors expect pushback - accepting garbage terms makes you look clueless.
Figma is targeting a $16.4 billion valuation in its NYSE IPO, aiming to raise up to $1.03 billion by selling shares between $25 and $28.
OpenAI plans to launch GPT-5 in early August, offering integrated o3 reasoning capabilities and mini/nano versions available through its API. GPT-5 aims to unify OpenAI's language models and simplify usage across applications, although it won't reach AGI status immediately. The anticipated declaration of AGI could impact OpenAI's revenue agreement with Microsoft.
Vanta's valuation surged to $4 billion after closing a $150 million funding round backed by CrowdStrike Ventures. The startup, which helps companies manage compliance with cybersecurity frameworks like SOC 2 and ISO 27001, serves over 12,000 customers, including tech giants Atlassian and Snowflake. More investors included Wellington Management, Atlassian Ventures, JPMorgan Chase, and Sequoia Capital.
AI answer engines are becoming the most influential gatekeepers in B2B buying, with ChatGPT hitting a billion users and 2.5 billion daily prompts. Buyers now ask AI what to buy instead of researching themselves - if your product isn't mentioned, it might as well not exist. Josh Blyskal from Profound found only 11% of citations overlap between ChatGPT and Perplexity, meaning you need platform-specific strategies. Reddit is now the most-cited domain at 6% of citations, and one-in-three citations come from comparative listicles. Almost half of cited domains shift within a single month.
Revenue leakage is the gap between what you should earn and what you actually collect - the ghost quietly draining dollars through missed invoices, zombie accounts still getting service, and pilot pricing that never graduates. It's a systems problem disguised as a people problem, living in the cracks between sales handoffs and billing processes. Common culprits include auto-renewals that silently fail, discounts hardcoded forever, and churned customers still accessing your product months later.
AI chatbots should be reframed as self-service tools rather than human support replacements due to their limitations. Achieving this involves clearly defining their purpose and implementing technical guardrails to ensure alignment. Understanding AI's capabilities and limits helps in effectively integrating them into customer support.
Sponsored
This episode of Metronome's _Future of SaaS Pricing_ features Kyle Poyar, Co-Founder & Operating Partner at Tremont. Kyle shares 2025 pricing trends [9] and how to stay competitive in the rapidly evolving
Substack surveyed 2,000 publishers and found 45% use AI - mostly men over 45 in tech/business categories who rely on it for research and editing, not generating full posts.
AI Meeting Notes by Notion is a tool that transcribes and organizes meeting notes directly within Notion, synced with your calendar.
PRDKit is an AI tool that generates structured product specs, wireframes, and go-to-market assets from feature ideas.
College graduates now have higher unemployment rates than the overall workforce for the first time since 2018 - the opposite of how this is supposed to work. Young men are struggling most, with 50,000 healthcare jobs going to women graduates while men gained far fewer positions across all sectors. People are applying to 1,000 jobs and getting rejected by algorithms in what's become a casino economy where getting hired feels like winning the lottery. This broken system feeds zero-sum thinking - the belief that my success requires your failure - which has replaced the Sputnik generation's positive-sum worldview where everyone could grow together.
Enshittification isn't just about social media - it's the natural lifecycle of any business with network effects. Companies offer great value to build their user base, then exploit that base once locked in. Uber's "earnings inflection strategy" jumping from losses to profits by squeezing customers proves this was always the plan.
Founders are river people, not villagers - they thrive in uncertainty while others choose stability. They see differently: they're fearless enough to hear "no" constantly but keep asking, they actively seek contrarians to challenge their thinking, and they know execution beats strategy. They ignore the right feedback, focus obsessively on recruiting A-players, and carry unshakable optimism despite knowing 90% of startups fail.
The internet's content distribution is about to mirror online advertising's auction system, but without the ads.
Rising executives often freeze when asked "What should we do?" because they're afraid of being wrong.
Fay built a platform that connects tens of thousands of clients with dietitians covered by insurance and raised $75M from Goldman Sachs and General Catalyst.
Sponsored
One user query triggers 50 model calls. You suddenly realize your "per request" pricing doesn't make sense. Users get hit with surprise bills, while your usage dashboard is still stuck two hours in the past.
Elad Gil argues the fog has lifted. Anthropic, OpenAI, Google, Meta, and other giants now dominate foundation models while Harvey leads legal and Cursor rules coding. Customer support converges on Decagon and Sierra and Perplexity stakes search. Capital and compute now block new LLM entrants, yet openings remain in accounting, compliance, and security. The next breakout will hinge on bold go to market, agentic workflows, and roll‑up strategies.
Figma just set its IPO range at $25–$28 a share, valuing the company at around $16B—less than Adobe once offered, but above its 2024 secondary round. It's not a moonshot, but it's solid: $1B run rate, strong margins, and a clear growth story. While most coverage focuses on past performance, Figma is quietly packing AI features into its product, possibly adding upside for retail investors who still believe in narrative.
Claude launched AI-powered apps with a clever twist - users pay with their own Claude subscription, not the creator's. A teacher can build a SAT prep tool and share it with students, but each student must sign up for Claude to use it. No API keys, no hosting costs, no authentication headaches. The constraint becomes the opportunity. Each app spawns its own growth loop as project tools drive team invites and calculators spread through forums. While Cursor and Bolt chase developers, Claude went the opposite direction - maximum constraints for minimum friction, turning every creator into a distribution channel.
Citizen, a real time safety alerts app, was losing users after it muted push notifications. Julia Knight tackled the issue by hand-collecting local emergencies, dropping headlines into a Figma template, and blasting fifteen tailored emails daily. The clicks proved so strong that the team automated triggers that now send forty million incident emails each month, bringing dormant accounts back for free, cutting paid acquisition, and even lowering iterable costs through renegotiation.
Antispace is an AI operating system that predicts, coordinates, and executes actions across your workspace to automate repetitive tasks.
HyperArc is an AI-native business intelligence platform that delivers instant, auditable insights from natural language queries.
A list of funds that invest in female founders.
In the 1950s, your software arrived in a semi‑truck bolted to its hardware. Seventy‑five years and six pricing waves later, vendors bill for shelf‑less shelfware, meter compute like electricity, and blend seats with usage in hybrid plans. This author traces the journey from boxed SKUs to SaaS, the ASP flame‑out, CFOs fighting surprise overages, and why clear value metrics will decide how we price AI next.
Now that every enterprise tool claims to be "AI-powered," what actually moves the needle? This playbook keeps it simple: pick a weird industry with painful, overlooked problems. Build for real workflows, not just demo decks. Prove your product works with clear, objective results, especially in edge cases. Let AI power more than just the product: use it across onboarding, support, sales, and GTM.
Most teams use AI to save money. Carta used it to save time where it mattered most: judgment-heavy accounting tasks that bottlenecked client service. Instead of replacing humans, it built in-house agents that could gather context, reason through complex workflows, and act autonomously. One agent now handles 25,000 monthly tasks in seconds, saving 3,500+ hours.
Picture an AI growth hacker that boosts clicks in days while accidentally killing sign‑ups.
Charta Health raised $22M in Series A funding led by Bain Capital Ventures to enhance healthcare operations with AI.
YC investments have increasingly focused on cybersecurity, industrial/manufacturing, and B2B sectors since 2020.
Sponsored
Business productivity shouldn't mean going all-in on Google or Microsoft. Proton Mail is email + calendar that's built for teams that care about privacy and security [4]. It's trusted by over 50k businesses and 100 million people around the world.
Google paid $2.4B to Windsurf just to get its AI talent fast. Meta is poaching AI researchers for NBA superstar money. The race shows how much A-players matter at every level. Netflix treats companies like sports teams, not families, ruthlessly keeping the best talent. Real performers are the "drivers," while everyone else is just a passenger along for the ride. One hiring rule stands out: pick the motivated hard worker over the brilliant slacker every time.
Founders are starting to treat LLMs like new distribution rails, not just search engines. Instead of competing for top spots on Google, they're optimizing to appear directly in answers from ChatGPT, Claude, and Perplexity. These models reward schema, clarity, citations, and presence across trusted sources, not just backlinks and keywords. When done right, it's like owning a seat in the AI's memory. LLM SEO isn't just a trend, it's an overlooked edge with compounding returns.
Mercury CEO Immad Akhund focused on scaling while maintaining high product quality by enhancing risk and compliance teams and launching new financial tools like invoicing and spend management. They created small, independent product teams to innovate while keeping core products strong, exemplified by the successful launch of Mercury Personal for consumer accounts. Akhund emphasized hiring passionate talent, managing "product debt" effectively, and leveraging existing customer behavior to expand services.
Keyword research is like writing bad survey questions - it assumes what people want before asking. A pizza shop targeting "Domino's" keywords gets clicks from people who specifically want Domino's, not any pizza. This seems obvious once pointed out, but most SEO starts with chasing high-volume keywords anyway. The author's client proved this wrong. Instead of writing another generic "best parenting tips" article (huge search volume), they wrote about the psychological cost of giving screens to toddlers. Zero people searched for that exact phrase, but parents loved it so much that it created its own search demand.
Effortless voice dictation in every application.
Simulated users for customer interviews.
Motherboard is a tool that lets founders track key business metrics from any website in one private, local dashboard.
Delphi Digital has launched Delphi Intelligence, a dedicated research platform focused on AI, robotics, and emerging tech. Backed by its experience in crypto, it has already invested in 20+ AI x crypto startups, incubated projects through AI-focused accelerators, and built a growing research team. The platform offers in-depth, public reports on key AI developments, with the aim of becoming a leading hub for high-signal analysis in the space.
Sub-$1M ARR startups went from 13 people in 2022 to just 7 today, while companies over $50M ARR saw efficiency drop 34%. Public SaaS now takes 57 months to recover sales costs - nearly five years. Early-stage teams figured out what big companies missed by hiring former founders who do everything. One person runs content and paid ads, GTM hires code like engineers, and each employee manages AI agents instead of people. Big companies keep adding specialists and management layers while startups just give everyone AI and let them run wild.
Everyone thought AI would simplify software - turns out we were wrong. Google's 2015 paper showed machine learning was just a tiny black box surrounded by massive infrastructure. AI was supposed to fix this - drop in an LLM and watch it work. Makes you wonder - are we making progress or just shifting complexity around?