Search the archive

AI CONTENDERS VS PRETENDERS IN ENTERPRISE

The enterprise AI buyer's filter has collapsed to one word: reliability.

MARC ANDREESSEN ON BUILDER CULTURE IN THE AGE OF AI (45 MINUTE LISTEN)

The doomer framing on jobs and productivity collapses once you observe what enterprise and consumer users actually do with the tools.

IT'S A BUYER'S MARKET | AND A SELLER'S NIGHTMARE

AI has flipped the leverage in software deals because more tools are sliding into the "nice-to-have" bucket. One vendor in the piece was charging $50K a year right up until a competitor released a free module that did 80% of the same work. The renewal playbook now works because buyers have a credible threat that the seller can't easily dismiss. "We'll build it ourselves with Claude." That sentence, three months before renewal, is what gets pricing to move.

THE NEXT BIGGEST MOAT IN AI

The most important companies are actually organizational inventions. They create a new kind of institution around a new kind of work, which makes a new kind of person possible. Great companies are structures that let a certain kind of talent finally express themselves. The opportunity now is in companies that have not been possible before, and the people who have been waiting for it to exist.

WHY ATTACKING YOUR COMPETITORS ONLINE IS DUMB

Public competitor attacks never work. Naming a competitor only gives them more exposure, and most people don't care what's going on between two companies. Instead, make fun of trends, publish details and honest comparisons, bid on competitor keywords, and correct factual errors privately, if they actually matter. Use the energy instead to outship competitors and build a better product.

HOW AI PRODUCTIVITY FAILS

Most AI users are only about 10% to 20% more productive, despite how 'game-changing' they claim the technology is. However, it is theoretically possible for 2x or even 10x+ productivity gains. Real transformation requires organizational refactoring and a change in personal practice. This article looks at personal and organizational pitfalls that bottleneck AI productivity. AI amplifies what you already do well and whatever your organization already enables.

MOMENTUM IS THE SAFEST MOAT

Building momentum safeguards businesses against AI disruption by enabling rapid execution and innovation. Quick, cohesive shipping, as seen with Claude Code, provides a strong competitive edge. Companies slow down when scaling, so maintaining speed and focus is crucial to outpace giants with traditional moats like economies of scale.

LOCAL-FIRST COST TRACKER FOR AI CODING AGENTS

budi is a local-first cost analytics tool for AI coding agents. It allows developers to see where tokens and money go across most popular AI coding tools, breaking down costs by repository, branch, ticket, and file. budi runs everything locally by default. It has an optional team dashboard that syncs aggregated metrics.

GOOGLE LAUNCHES GEMINI 3.1 FLASH-LITE IN GENERAL AVAILABILITY

Gemini 3.1 Flash-Lite is now generally available through Google Cloud platforms. The model provides ultra-low latency and can handle high-volume processing. It is the most cost-efficient Gemini 3 model. Gemini 3.1 Flash-Lite has multimodal capabilities and supports both text and image processing. The model can handle agentic tasks like tool calling and orchestration.

FIGMA'S COLLAPSE SHOULD TERRIFY SAAS LEADERS

Figma has lost roughly 85% of its market value since its IPO in August last year. The market has been pricing design-tool disruption for months, and the question now is not whether agents will replace design workflows but rather how fast the transition will happen. Figma's vulnerability was that its product was the workflow itself, something that AI agents excel in. This post discusses a three-layer framework for repositioning products over the next few years to help companies avoid succumbing to the same fate as Figma.

IN THE TRENCHES WITH 50 MIDWEST CIOS

The story in enterprise AI right now is that the bottleneck has moved off the model. Fifty CIOs from Midwest enterprises met in Chicago last week, and when asked who had agents at scale, no one raised a hand. Of 25 in a smaller breakout, only 5 had agents in production at all. The reason isn't model quality. It's that nobody has actually written down how the work gets done inside these companies, so there's nothing for an agent to operate on.

SLOP VS TASTE

Most people use "slop cannon" as an insult for a person making AI slop. The piece flips it. The slop cannon is a person using junky AI output as raw material to think with, then editing hard. Move fast in produce mode, slow down in selection mode. The piece also bets against the new OpenAI and Anthropic consulting joint ventures, on the grounds that most enterprise data isn't ready for the work being sold.

THE HARD THING ABOUT EASY THINGS

Management is more unpleasant when you're not defending humans, and when the blame you're absorbing is not on behalf of humans.

WHAT DO YOU HAVE TO BELIEVE?

If you can't describe what you'd do to kill your own company, you don't fully understand what you're building.

AI CHAT BOTS SHOULD BE DEPRECATED

Chatbots essentially offload interface design to users.

IS MY COMPANY "AI PILLED"?

Companies must embrace AI adoption to thrive, with success hinging on team expertise and innovation. Levels of AI implementation range from basic prompt usage to full process automation, impacting headcount and roles. Building a culture of adaptability, hosting hackathons, and investing in AI tools are crucial strategies to drive AI proficiency and maintain competitiveness.

THE "AI JOB APOCALYPSE" IS A COMPLETE FANTASY

Doomers keep predicting jobs will disappear in productivity revolutions, and they keep being wrong. Excel cost about 1M bookkeeping jobs and minted 1.5M financial analysts in the same window. The Atlanta Fed reports 90%+ of firms saw zero AI impact on headcount over three years. Earnings calls run 8:1 augmentation over substitution. The data isn't subtle. The narrative is just lagging.

AI APPLICATIONS HAVE TWO GROWTH CURVES

The companies winning the fast AI markets right now are not on the only growth curve available. Treating their trajectory as the benchmark for all AI is going to lead to a lot of misallocated capital. For the most valuable problems, the next phase of AI won't look like Cursor's growth chart. Slow growth is required to build companies that competitors can't easily copy in a quarter when the funding cycle turns.

DON'T SELL THE BUILD EFFORT

Avoid pricing software based on build effort, as it disconnects from optimal value-based pricing. Engineers and sales teams often struggle with pricing due to biases related to development effort and historical costs. Instead, focus on the value delivered to customers, with pricing decisions informed by a dedicated pricing team if possible.

KILL YOUR ONBOARDING: SELLING TO 10,000+ NEW USERS A DAY

The interesting analyses that businesses want to run today depend on events that happened two years ago. If these events weren't logged, no amount of cleverness will work. Businesses need to instrument event data on day one. The cost of not having historical events when needed is high and asymmetric.

THE BEST AI-NATIVE GTM PLAYS YOU'RE NOT RUNNING

O1C26IMLH to save. Limited availability. Register now.

GOOGLE ADDS AI-POWERED BIDDING AND DEMAND-LED BUDGETING TO SEARCH AND SHOPPING

Google is rolling out AI-driven bidding and budgeting features across Search, Shopping, and Performance Max. The update is aimed at helping advertisers capture more demand without increasing manual effort. Users will be able to help campaigns respond more dynamically to shifting consumer behavior. They will be able to capture more conversions by using AI to find demand and optimize spend in real-time. The new features will result in more value from the same budget while staying competitive in the fast-changing search landscape.

FLUE

Flue is an agent harness framework similar to Claude Code or Codex. It is 100% headless and programmable with no TUI or GUI - just TypeScript. Flue's agents act autonomously to solve problems and complete tasks with very little code. It enables anyone to write once, build, and deploy agents anywhere.

RAISING CAPITAL IS THE ONLY SKILL THAT MATTERS

Hashgraph Ventures' Dara Campbell argues capital raising sits above sourcing, diligence, and decks as the skill separating elite VCs and founders, citing data that founders pitch 58 investors for one seed yes, 75% of VC-backed companies never return cash, and 13% of GPs in 2024 had no plans to raise another fund. The skill is psychological, built on likability under stress, authenticity that allocators detect physically, patience that can take three years of relationship-building before a Middle East family office wires money, and the boldness to read the room and ask at the right moment for the right size.

DAVID HABER AT A16Z: IN VERTICAL AI, MARKET STRUCTURE IS THE NEW TAM

Market structure predicts success better than traditional TAM metrics in Vertical AI. Startups should adapt strategies based on whether markets are fragmented or concentrated, avoiding direct competition in entrenched markets. This approach enables Vertical AI companies to innovate and potentially dominate by building comprehensive, unique solutions.

THE STUFF NOBODY TELLS YOU ABOUT STARTUP MARKETING

Startups often overlook their existing activities as marketing.

STICKS, STONES, AND LIGHT SPEED

AI is reshaping the workforce, eliminating the need for middle-layer roles and emphasizing the importance of judgment and human connection.

DAU, WAU, AND MAU ARE THE NEW LIGHTHOUSE METRIC IN B2B + AI. HARVEY'S A GREAT CASE STUDY

Usage is arguably the most important metric in AI and B2B.

Learn more at [justworks.com]

THE CASE FOR STRATEGIC ILLEGIBILTY

Making your company legible to AI could result in the commoditization of your own moat. Founders need to be building two versions of their company right now: the company that humans work inside and a digital twin that AI agents can navigate, query, and operate in. They need to consider what things should be deliberately kept out of the AI system to protect their advantages. The highest-value things to keep partially illegible are the things that make the company weird, specific, and hard to copy.

THE LONG BECOMING

The gap between AI-enabled and AI-native companies will look like the gap between cloud-enabled and cloud-native, where the two paths look identical for a while and then diverge. AI-native means rebuilding product, workflow, and org from the floor up, not bolting an assistant onto yesterday's process. The work shows up as a sequence of bottlenecks you solve in order. Adoption comes first, where every engineering manager is required to use Claude Code at roughly the median of their direct reports. Once people are actually using the tools, engineering velocity opens up. One CTO reports his top decile is 5x more productive while the median engineer is up 20%. After that, the rate limit is product taste - whether anyone has the judgment to decide what to build next.

HOW TO BUILD SERVICES-AS-SOFTWARE BUSINESS

The next trillion-dollar company will sell the work, not the tool. AI turns services budgets into something startups can actually attack. The head start belongs to the people who are willing to do the work by hand until the work tells them what to automate. Companies that look like services firms and earn like software firms are never accidents.

PRICING IS A PRODUCT SURFACE

Pricing is a surface because customers touch it more often than they touch most product features. Companies need to treat pricing as a hypothesis to test. They need to iterate on pricing as they learn what works. Being cemented in pricing will cause your business to bleed margin and cap growth, a dangerous combination.

HOW WE GOT 3.6M VIEWS & 9,000 DEMOS ON A MONDAY MORNING LAUNCH

LaunchVideo coordinated Helena's launch to 3.6M views, 10x homepage traffic, and 70%+ positive sentiment, taking founder Seijin Jung from 147 followers to 3,000+ in a week. The team rejected the "AI CMO" frame to avoid threatening buyer CMOs, anchored Helena as the "first autonomous AI marketer," and shifted the launch from the standard Tuesday 7 am PT window to Monday to dodge a crowded feed where competing launches all stayed under 3M views.

EVERYTHING I KNOW ABOUT FUNDRAISING

Slash recently raised a $100 million Series C at a $1.4 billion valuation. The company has raised over $160 million in venture capital across four rounds. This post details everything its founder knows about raising venture capital. Most of the lessons were learned through Y Combinator - while its advice is geared towards seed-stage companies, the concepts are still applicable at the growth stage.

trying to get found by AI, WordPress.com is ready

OPENAI LAUNCHES SELF-SERVE ADS MANAGER FOR CHATGPT

OpenAI has launched a self-serve Ads Manager as well as new CPC bidding and expanded measurement tools. Advertisers can now create and manage campaigns directly through OpenAI instead of relying on managed partnerships and agency relationships. Advertisers in the US can now register for access. The rollout is currently limited and still in beta.

MANAGE YOUR STORE INSIDE CHATGPT OR CLAUDE

Shopify users can now manage their stores inside ChatGPT or Claude. Users can simply prompt their chosen assistant to add products to their store. The integration allows them to get insights, look up orders, and more without ever leaving the chat. A video demo is available in the thread.

DON'T BECOME A WRAPPER

It has become a trend for engineers to turn themselves into agent wrappers. They interact with models to get them to generate their work, but they don't actually make decisions or do any reasoning. It is appealing to use tools that output work good enough that the failure modes don't bite right away. People need to consider what remains defensible when anyone with the same agent can get the same outputs.

YOUR TEAM ISN'T USING AI. HERE'S WHY THAT'S YOUR FAULT

The more people use AI, the better they will be at using it. Teams need experience in using AI to understand what the technology can do for them. Leaders need to give teams projects with tight deadlines to help teams learn how AI can benefit them. No amount of strategy can fix a lack of experience.

THE AGENT INTEGRATION TAX JUST BECAME THE BIGGEST MARKET IN ENTERPRISE AI

Salesforce now estimates that copying context across Slack, Salesforce, Workspace, and ServiceNow costs the average employee two hours of productivity a day. Google Cloud says 75% of its customers use its AI products and 330 of them processed more than a trillion tokens last year. ServiceNow runs 95 billion workflows annually and is moving agents into a governed registry with audit trails. Coordination across systems, not model intelligence, is what's getting in the way.

BECOME A CURATOR FOR TLDR FOUNDERS (3 HRS/WEEK)

TLDR is looking for someone who has founded a VC-backed startup to help write TLDR Founders for hundreds of thousands of subscribers. Our curators have been invited to major tech events, scouted for Tier 1 VCs, and get early access to unreleased TLDR products. Learn more [18].

WHEN EVERYONE IS A KEY PERSON IN YOUR COMPANY

The tradeoff at the heart of AI/labor ratio decisions is resiliency.

INSIDE JAPAN'S STARTUP BOOM: ELITE ENGINEERS, EXPLODING REVENUE, AND ALMOST NO FUNDING

Japanese startups have strong builders, launched products, and paying customers, but they lack access to capital.