AI agents are knocking. Is your API ready to answer? | GraphQL’s Matt DeBergalis
Open the door or get left behind.
The rise of AI agents is more than a tooling upgrade - it's a fundamental rewiring of the entire developer experience, with your APIs at the very center.
We're joined by Matt DeBergalis, co-founder and then-CTO-now-CEO (congrats Matt!) of Apollo GraphQL, to explore this massive transformation. He introduces the emerging concept of "agent experience," explaining why systems built for human developers are not ready for the unprecedented scale of AI calling APIs.
Matt argues that as the old rules of software development get re-evaluated, engineering leaders must rethink their entire stack. He presents a powerful analogy: a structured data layer like a graph is the perfect "left brain" for the "right brain" creativity of LLMs. This provides the semantic precision and guardrails needed for AI to act reliably, enabling a future where user experiences are personalized "to 11" and APIs become the core business asset. This conversation is a crucial guide for leaders on how to prepare by prioritizing higher-level system design, and why clear communication and architecture are becoming far more critical than handwriting code.
“[Software is] gonna make a lot more API calls. It's gonna do it in new ways with new combinations. I think the handwritten software was always kind of limited. You can only ship so many versions of your app every year.” - Matt DeBergalis
The Download
The Download is the top news from your favorite engineer’s favorite engineers. 🥇
1. Claude Code caps usage, so who ruined the fun? 🎭
Anthropic just capped usage of its Claude Code on its subscription plan, in another pricing shake-up for agentic coding tools. Some disgruntled devs took to Reddit to shine a light on some of the power users revealed behind Anthropic’s inference by sharing a viberank leaderboard, causing mixed backlash from both the AI provider and the community itself. The incident underscores the ongoing tension between AI providers and their users as the industry grapples with sustainable pricing models and the implications of transparent usage data.
Read: Anthropic unveils new rate limits to curb Claude Code power users
2. OpenAI's study mode provides a critical thinking boost 📚
OpenAI is rolling out a new study mode for ChatGPT, designed to encourage students to engage critically with their learning rather than just regurgitating answers. This shift could be a game changer in education, helping learners cultivate deeper understanding through Socratic questioning. This initiative marks a crucial step for the future of AI in edtech, bringing the vision of personalized tutors for every student closer to reality.
Read: Introducing study mode
3. Age verification rollouts and the parallels to shadow IT 👤
The UK has become a pioneer in implementing age verification for online services, but the rollout is proving to be a headache. Emma Roth at The Verge provided great coverage on why many companies are opting out of the region entirely, and many users are already finding clever ways to bypass the system. This mirrors challenges in software development teams where new administrative roadblocks often lead to shadow IT or distrust. When friction exists where it previously didn’t, devs just trying to do their job will find another way. The same goes for users at large.
Read: Ready or not, age verification is rolling out across the internet
Code reviews just got smarter (sponsored)
Stop treating code review like a chore. Start treating it like a strategic advantage:
🔎 Instant PR summaries that make context switching painless
⚡ Smart suggestions that go beyond linting
📬 Reviewer routing that gets the right eyes on the right code
The result? Higher-quality reviews, faster merges, and more time for actual engineering. LinearB AI reduces cognitive load, flags critical issues early, and frees reviewers to focus on architecture and business logic. Check out the button below if you’re ready to level up your reviews.
4. Atlassian's survey shows that AI redirects developers time instead of actually saving it ⏳
According to Atlassian’s latest survey, nearly all developers are saving knowledge working time with AI tools, yet they're simultaneously losing more time to organizational inefficiencies. With a growing empathy gap between developers and leadership (63% report that their leaders don’t understand their pain points) it's clear that simply implementing new tools and hoping for time saved isn’t enough.
Read: Atlassian research: AI adoption is rising, but friction persists
