This month, we looked at three patterns in what TLDR’s 7M+ readers are clicking on plus one big non-AI story. Our biggest single story pulled over 15,000 clicks, and the macro themes tell a clear story about where technical buyers’ heads are right now.

1. Anthropic had its biggest month ever: a Claude Code source leak, a high-cybersecurity-risk model called Mythos, and a $1 trillion valuation.
2. AI compute economics is realigning, and readers are doing the math.
3. Evals, agent harnesses, and governance are now what builders click on.
4. Apple’s hardware-AI pivot under new CEO Ternus pulled cross-newsletter audiences.

Let’s dive in.

April was Anthropic’s month start to finish. Three of the top six editorial stories across the entire network were Anthropic-related.

The story that captures this best is Lessons from Building Claude Code: How We Use Skills. It pulled 8,750+ clicks in TLDR Tech and another 5,400+ in TLDR AI, making it the most-clicked technical article this month. But it’s the broader pattern that matters.

• Claude Code’s source code appears to have leaked: 15.2k clicks (TLDR AI)
• Claude Mythos: 12.6k clicks (TLDR AI)
• Entire Claude Code CLI source code leaks thanks to exposed map file: 10.8k clicks (TLDR Tech)
• How Anthropic’s Claude Thinks: 10.6k clicks (TLDR Tech)
• We’re missing a much bigger point on Mythos: 9.3k clicks (TLDR Tech)
• Inside the Claude Code source: 9.0k clicks (TLDR Tech)

Plus Anthropic crossed OpenAI’s valuation at $1 trillion on April 24, and Claude Opus 4.7 shipped on April 17 with another 7.9k clicks on the launch story alone. Our readers couldn’t get enough of any of it.

💡 Takeaway for marketers: Anthropic-adjacent positioning had a tailwind in April that’s likely to keep running through Q2. If your product slots into how engineers use Claude in production (agent harnesses, Claude API tooling, Anthropic-compatible infra), this is the audience leaning in.

The most-clicked non-Anthropic AI stories this month were all about where the compute money actually goes, and how to spend less of it.

• Modular Post-Training: 12.8k clicks (TLDR AI)
• Google in talks with Marvell to build custom AI inference chips: 11.0k clicks (TLDR AI)
• Meta targets 20 May for 8,000 layoffs as it redirects billions toward AI infrastructure: 10.9k clicks (TLDR Tech)
• How Elon Musk Plans to Bypass the ASML Bottleneck to Build TERAFAB: 10.3k clicks
• OpenAI has effectively abandoned first-party Stargate data centers: 10.3k clicks

Three things stacking at the same time: hyperscaler infrastructure is being re-shopped, billing models are flipping to per-token, and the engineering bench is racing on extreme efficiency (1.58-bit models, KV cache sharing across datacenters, modular post-training).

💡 Takeaway for marketers: “TCO per token” and “where your inference dollar actually goes” land harder than feature pitches right now. Any product with a credible cost story is in tailwind territory. If you have a cost calculator, a benchmark, or a “we save you X on inference” claim with proof behind it, this is the moment.

March’s email noted that AI coding tools moved from interesting to essential. April makes the next move clear: builders are now organized around measuring whether their agents actually work.

• AI Governance Is the Bottleneck: 10.8k clicks (TLDR IT)
• Stash: 10.7k clicks (TLDR AI engineering, agent persistence)
• Measuring Scientific Discovery Agents: 9.7k clicks
• Gemini Robotics ER 1.6: Embodied Reasoning: 9.7k clicks
• The PR you would have opened yourself: 9.6k clicks
• Introspective Diffusion Language Models: 9.9k clicks
• Claw-Eval Benchmark for AI Agents: 9.0k clicks
• Windsurf 2.0 adds Devin and Agent Command Center: 9.1k clicks

The AI engineering category alone pulled 318.4k clicks in this window: second only to TLDR Tech’s quick links across the entire 12-newsletter network.

💡 Takeaway for marketers: For any product touching the agent stack (eval frameworks, observability, governance, harnesses, agent-friendly databases, secrets management for agents), TLDR AI is the highest-engagement venue on the network and “stop babysitting your agents” type messaging is what’s clicking. Show how teams instrument agents in production, not what your model can do in a demo.

The non-AI story of April was Apple. Six related stories cleared the click thresholds across three different newsletters.

• The Rise of Apple’s New CEO: A Hardware Expert Takes Over in the AI Era: 9.3k clicks (TLDR Tech)
• Apple AI Glasses Will Rival Meta’s With Several Styles, Oval Cameras: 8.7k clicks
• Apple’s Cook Gives Ternus a Pipeline of 10 Major New Product Categories: 8.7k clicks
• Apple’s Revamped Siri Interface in iOS 27 Is Hidden in WWDC Teaser: 6.7k clicks
• Apple approves drivers that let AMD and Nvidia eGPUs run on Mac: 5.0k clicks
• Apple celebrates its 50th anniversary with special animated homepage: 3.5k clicks (TLDR Design)

For the first time in a while, Apple is a multi-newsletter story. Tech, Design, and IT readers are all clicking on the same arc, and with iOS 27 and Apple Glasses on deck this is set to keep running through WWDC.

💡 Takeaway for marketers: Hardware brands, peripheral makers, accessory companies, and anyone selling to Apple developers had a real cross-newsletter audience moment in April. Multi-newsletter Tech and Design and IT runs ahead of WWDC let you reach the same buyer three times in a week. WWDC week itself fills fast.

May and June 2026 are mostly sold through with some slots still open. July and August have wide-open inventory across the network. Primary Placements for July windows are still available for all 13 TLDR newsletters but won’t stay open long as Q3 typically books out 4 to 6 weeks ahead.

💡 Where to run this on TLDR: Each of these themes maps to a specific audience your team can plan around. Anthropic-adjacent and agent-infra products fit TLDR AI Primary plus TLDR Tech programming. Cost-per-token and TCO positioning runs across TLDR AI, Tech, and DevOps. The eval / governance / harness segment is the busiest subset of the entire network by reader engagement, with TLDR AI engineering pulling 318.4k clicks in this window alone. And Apple-adjacent products land best as a multi-newsletter Tech + Design + IT pulse running into WWDC.

If you want help mapping your roadmap to a Q3 plan, your rep can put together a multi-week proposal that hits the right reader at the right moment.