TLDR Founders 2026-06-19

The AI cost crunch is now hitting the tools themselves. Microsoft is pulling its Experiences and Devices engineers off Claude Code by June 30 because the token bills were running about $2,000 per engineer a month, and moving them to flat-rate Copilot. Around the same time, GitHub flipped every Copilot plan to usage-based AI-credit billing on June 1.

Creating software used to require access to a small, highly specialized group of people concentrated in a handful of tech hubs. This meant that only ideas large enough to justify the costs ever made it into the world. That complexity and cost is collapsing. Now, domain experts of every kind can start solving problems they understand deeply. There will likely be an explosion of products from unexpected places built by local experts with experience.

Companies are essentially made up of a group of people and a collection of decisions. How these people make these decisions is the art of running a business. This post contains a collection of general principles on how to make decisions. They serve as frames, considerations, and shared practices to draw upon when making a decision.

Top-performing tech companies increasingly use founder-led content, customer education, community building, and product visibility to drive demand before traditional sales engagement. The playbook emphasizes creating trust and mindshare at scale through content, events, and social channels rather than relying solely on outbound sales motions.

AI promised to supercharge sales, and to be fair, it delivered the easy stuff. We can write great cold emails now, build account lists in seconds, and summarize every call. However, none of that actually wins deals. What wins is still the boring upstream call of deciding which accounts to chase and why now. Most teams are quietly letting an AI vendor make that call for them, so they end up with the same targeting as every competitor on the same tool.

Ads get ignored on social media but not in TLDR! Reach developers, PMs, marketers, founders and other tech leaders where they actually pay attention. Learn more about sponsorship opportunities.

Design, build, and manage websites with AI-assisted content creation, collaborative branching workflows, and a community platform for sharing and distributing site resources.

Copilot Cowork executes complex, long-running multi-tool tasks. Users define the work, and Cowork runs it end-to-end and returns a completed result. The tool is designed to be accurate, secure, and low-cost. The feature is now generally available for users on the Microsoft 365 Copilot User Subscription License.

Retool is a platform that allows developers to build software on whatever AI coding platform they want and ship it through a single, governed runtime that enforces authentication, permissions, and data governance automatically. Developers can deploy any React app on Retool, and it will map the app's data connections to resources the team has already configured and secured. Apps can be deployed on Retool Cloud, in a VPC, or fully on-prem. Retool handles versioning, release management, environment promotion, access changes, and monitoring. Free hosting, plus bonus AI credits on every paid plan, is available until July 1.

The smarter the models get, the less your software is worth alone, so every app company now has to become a data company and a fintech company. The data half is judgment, the corrections your users make that no benchmark captures, which is why xAI took a $60B option on Cursor mostly to get inside the token flow, not for the editor. The survival rule for the "what if Anthropic builds this" question is to let agents read but guard the writes, because the decision being made right now is the only thing a rival can't scrape.

Human workflows usually have checkpoints, like a review, that make coordination possible. Agents need the same kind of containment. Agents can do much more than just write code. A pull request would give agents a familiar framework to work with humans. It would allow people to review changes, accept useful work, reject weak work, and leave a record for the future.

This playbook remaps the four core stages of the startup journey (Idea, MVP, Launch, and Scale) according to the AI era.

AI startups should focus on resolving deep scientific challenges, as significant advancements will drive future technological frameworks.

Opal has introduced a chat-based onboarding UX that encourages users to interact with the product more.

Founders often excel at running their companies but struggle during board meetings due to different skill requirements.