<p><strong>TLDR:</strong> Cursor quietly acquired Continue, the 34,300-star open-source Copilot alternative, in what looks like an acqui-hire — the product is shut down, users have until July 15 to export data, and co-founder Nate Sesti is joining Cursor. OpenAI expanded its Daybreak security initiative with a Codex Security plugin and GPT-5.5-Cyber, partnering with Trail of Bits to patch open-source vulnerabilities at machine speed. AWS launched Lambda MicroVMs built on Firecracker, offering VM-level isolation with near-instant startup and up to 8 hours of stateful runtime — a direct answer to the sandboxing problem for AI-generated code execution. Anthropic updated its privacy policy to require government ID and biometric verification from flagged users starting July 8, handled through Persona, a Peter Thiel-backed identity firm.</p>
<hr>
<h2>Cursor acquires Continue, shuts down the product</h2>
<p>Cursor has acquired Continue, the open-source AI coding assistant with 34,300 GitHub stars that positioned itself as a transparent alternative to GitHub Copilot. The deal is an acqui-hire: the product is dead, users have until July 15 to export their data, and the codebase has been handed to the community under its existing Apache 2.0 license. Co-founder Nate Sesti is joining Cursor; other team members have dispersed. This follows Cursor’s earlier acquisitions of Supermaven and Graphite, and coincides with SpaceX’s reported $60 billion investment in the company. <a href="https://daily.dev/posts/2jNQcpAln" target="_blank" rel="noopener nofollow">Read more</a></p>
<h2>OpenAI launches Patch the Planet with Trail of Bits, ships GPT-5.5-Cyber</h2>
<p>OpenAI expanded its Daybreak initiative with a Codex Security plugin for automated vulnerability patching and a new GPT-5.5-Cyber model focused on cybersecurity. The “Patch the Planet” program pairs Trail of Bits engineers directly with open-source maintainers to review code, develop patches and tests, and build reusable security workflows. It’s a direct competitive response to Anthropic’s security tooling, and the timing is notable given Anthropic’s own CVE disclosure work covered in yesterday’s digest. <a href="https://daily.dev/feed-by-ids?id=dfP6afOUQ&amp;id=CUVdBVyxk&amp;id=Rr3XFF0wQ&amp;id=TMwTNFDIT&amp;id=NkkffMsEy" target="_blank" rel="noopener nofollow">Read more</a></p>
<h2>AWS Lambda MicroVMs bring Firecracker isolation to serverless</h2>
<p>AWS launched Lambda MicroVMs, a new compute primitive built on Firecracker that gives each workload a dedicated VM with no shared kernel, near-instant launch via pre-initialized snapshots, and stateful execution for up to 8 hours. Specs top out at 16 vCPUs, 32 GB memory, and 32 GB disk, with each MicroVM getting its own HTTPS endpoint supporting HTTP/2, gRPC, and WebSockets. The obvious use case is running untrusted user- or AI-generated code in multi-tenant environments without the isolation tradeoffs of containers. Available now in five AWS regions on ARM64. <a href="https://daily.dev/feed-by-ids?id=ynJa57LqL&amp;id=X2cYRNPDz" target="_blank" rel="noopener nofollow">Read more</a></p>
<h2>Anthropic adds biometric ID verification for flagged Claude users</h2>
<p>Anthropic’s updated privacy policy, effective July 8, requires some users flagged for potential policy violations to submit a government-issued ID, selfie, and biometric face geometry data through Persona — a third-party identity platform backed by Peter Thiel’s Founders Fund. The biometric data qualifies as protected information under laws like Illinois’s BIPA. Anthropic says it affects only a small subset of users, but the timing draws scrutiny given ongoing tensions with the current US administration over access to its most powerful models. <a href="https://daily.dev/feed-by-ids?id=ls8HgQmok&amp;id=3L5mNJQv0" target="_blank" rel="noopener nofollow">Read more</a></p>
<hr>
<h2>Also notable</h2>
<ul>
<li><strong>Simon Willison ports Moebius 0.2B to ONNX and runs it in-browser with Claude Code:</strong> Using Claude Opus 4.8 in a parallel-agent workflow, Willison converted the Moebius 0.2B image inpainting model from PyTorch/CUDA to ONNX, published weights to Hugging Face, and shipped a working web UI on GitHub Pages — with the CacheStorage API handling the ~1.3 GB model file for browser-side caching. <a href="https://daily.dev/feed-by-ids?id=OAowH9jTK&amp;id=56A9G4PMc" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Reflection AI signs $150M/month compute deal with SpaceX through 2029:</strong> Reflection AI, an open-source lab founded by former Google DeepMind researchers, secured a contract worth up to $6.3 billion for access to Nvidia GB300 chips at SpaceX’s Colossus 2 data center near Memphis — following similar deals SpaceX struck with Anthropic ($1.25B/month) and Google ($920M/month). <a href="https://daily.dev/posts/D0SWKBZ1s" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Ai2 releases TMax 27B terminal agent hitting 42.7% on Terminal Bench 2.0:</strong> Ai2’s TMax 27B is available on Hugging Face and scores 42.7% on Terminal Bench 2.0, reportedly rivaling much larger models on terminal task performance. <a href="https://daily.dev/posts/LKIT2NooK" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AutoJack RCE chain patched in Microsoft AutoGen Studio:</strong> Researchers disclosed AutoJack, a three-vulnerability chain in AutoGen Studio’s graphical interface — localhost WebSocket trust, missing MCP API auth, and unsanitized base64 parameters — that lets a malicious webpage execute arbitrary commands on a developer’s machine via their AI agent; only developers building from the GitHub main branch during a brief window were exposed, and Microsoft has patched it. <a href="https://daily.dev/posts/uX7o821d6" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Airbnb reports 65% increase in PR throughput with 59% of code now primarily AI-authored:</strong> Airbnb achieved 97% weekly agentic AI adoption among engineers without mandates, driven by treating internal tooling as a product — the team built AirChat CLI, AirChat SDK, and AirChat Remote for async multi-agent workflows to support the shift. <a href="https://daily.dev/posts/yLDpUbc0P" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>


AI coding assistants, code generation tools, and vibe coding workflows.

Agentic Digest

A roundup of major AI developer tooling news: Cursor acquired Continue (the 34K-star open-source Copilot alternative) in an acqui-hire, shutting down the product with a July 15 data export deadline. OpenAI launched 'Patch the Planet' with Trail of Bits, shipping a Codex Security plugin and GPT-5.5-Cyber model for automated open-source vulnerability patching. AWS launched Lambda MicroVMs built on Firecracker, offering VM-level isolation with near-instant startup and up to 8 hours of stateful runtime for running untrusted/AI-generated code. Anthropic updated its privacy policy to require biometric ID verification for flagged users via Persona starting July 8. Also covered: Simon Willison's in-browser ONNX model port, Reflection AI's $150M/month SpaceX compute deal, Ai2's TMax 27B terminal agent, an AutoGen Studio RCE chain (AutoJack), and Airbnb's 65% PR throughput increase with 59% AI-authored code.

Cursor acquires Continue, OpenAI launches Patch the Planet with GPT-5.5-Cyber