Gemini 3.5 Flash Rolls Out on Databricks

#1 𝕏

Sam Altman ships the new Codex today, delivering an AI-driven coding assistant to help developers accelerate their workflows.

#2 𝕏

Ali Ghodsi rolled out Gemini 3.5 Flash on Databricks, offering blazing-fast AI inference and smart capabilities directly within the platform.

#3 𝕏

Qwen launched Qwen3.7-Max, a flagship agent-first foundation that delivers end-to-end coding, multi-agent MCP productivity integrations, 35-hour autonomous workflows, and scaffold-agnostic toolchain support.

#4 𝕏

Boris Cherny rolled out a new `/usage` command in Claude Code’s CLI to break down token usage by Skills, Agents, MCPs, and Plugins. Desktop support is coming next.

#5 𝕏

Rowan Cheung sits down with Sundar Pichai to unpack Google I/O 2025’s big reveals—Omni, a “Nano Banana for video” personalized cross-device AI, plus Spark agents—and explores the future of YouTube, mainstream Gemini adoption, and AI’s three-year roadmap.

#6 𝕏

NVIDIA AI shipped NVIDIA-Verified Agent Skills, offering transparent skill cards that detail each skill’s function, origin, risks, and integrity. Built on an open specification, these verified skills run reliably across Claude, OpenAI Codex, and Cursor.ai.

#7 𝕏

Google DeepMind launched Science Skills for Google @Antigravity, a new AI toolkit integrating insights from 30+ life-science sources—like UniProt and the AlphaFold Database—to accelerate day-to-day research workflows.

#8 𝕏

Philipp Schmid built a GitHub Issue Triage Agent using a single curl to the Gemini API.

#9 𝕏

Santiago demos a step-by-step video to build a voice agent from scratch using Claude Code and @AssemblyAI’s new Voice Agent API. This approach cuts out the need to stitch together 12 separate components.

#10 𝕏

Philipp Schmid announced that you can now create custom agents using Google’s Gemini API managed-agents quickstart (https://ai.google.dev/gemini-api/docs/managed-agents-quickstart) and teased upcoming support for AIS invocation.

#11 𝕏

v0 now offers a “Fix PR Conflicts” feature that automatically merges the base branch and uses an AI agent to resolve pull-request merge conflicts for you.

#12 𝕏

LlamaIndex 🦙 launched Latency Metrics in LlamaParse, offering queue, processing, and total latency breakdowns by tier. It also adds a Metrics tab with latency scatter plots and job volume histograms for pattern analysis over time.

#13 𝕏

xAI now lets @grok or X Premium subscribers tap the Grok Build model in @opencode for high-speed codebase intelligence.

#14 𝕏

Sebastian Raschka hails Gated DeltaNet-2 as a standout hybrid-attention innovation in the transformer stack and has added its paper to his reading list. He also shares his primer on the original Gated DeltaNet for background.

#15 𝕏

Cognition launched Devin on Windows, letting PM Builders map, build, and modernize Windows applications. It also migrates .NET Framework to .NET Core with built-in QA/testing and leverages Devin’s computer-use capabilities for end-to-end app testing.

#16 𝕏

Garry Tan shipped full iOS QA in GStack, enabling automated testing on real iPhones and simulators and unlocking /qa features for iPhone and iPad apps.

#17 📝 PromptLayer Blog

The 7 Best Prompt Management Tools in 2026 — Tested and Compared - Overview of top prompt management tools for 2026, explaining why prompt management becomes essential as prompts move from simple strings to production artifacts. The article defines what a prompt management tool does—versioning, testing, monitoring, and regression protection.

#18 📝 Ampcode Chronicle

Rush, 2.0 - Amp released Rush 2.0, which now uses GPT-5.5 with no reasoning and is tuned for small, focused code edits using shell_command for search/read/verify and apply_patch for edits, with task subagents sharing the same config and several redundant tools removed. In internal evals Rush 2.0 solved 44% of tasks (up from 39%), cost $0.58 per task vs $0.44 previously (+18%), and averaged 1:32 runtime (down from 2:59), and the team recommends pairing Rush with the slower, read-only Oracle for planning or review.

#19 📝 Claude Code Blog

Claude now works with more security and compliance tools - Claude now integrates with additional security and compliance tools through partnerships and APIs, making it easier for enterprises to connect Claude to existing security and compliance workflows. The update highlights expanded platform and enterprise capabilities for regulated environments.

#20 📝 Simon Willison

Datasette Agent - Simon announced the first release of Datasette Agent, an extensible AI assistant for Datasette that integrates his llm Python library with Datasette. He expresses excitement about bringing LLMs and Datasette together and links to a longer post.

#21 𝕏

Yann LeCun – Professor at NYU & Executive Chairman at AMI Labs says next-gen AI will rely on non-generative architectures like JEPA and world models that map (state, action) to the next state in representation space.

#22 𝕏

claire vo 🖤 – building @chatprd, 3x CPTO, PM + AI + startups highlights ClickUp CEO’s vision for AI-native orgs: lean teams of engineers, PMs and agent managers (merged with design) delivering 100× leverage by replacing any inefficient node with an agent, backed by $1 M salar...

#23 𝕏

Lenny Rachitsky reports ClickUp cut 22% of its workforce per Zeb Evans (@DJ_CURFEW) to rebuild teams around AI-driven workflows, aiming to forge a “100x org” through automation and lean roles.

#24 𝕏

Lenny Rachitsky points out that in large, fast-moving companies, PMs shouldn’t spend much of their limited bandwidth on shipping to production. Instead, their highest-leverage work is unblocking teams, aligning stakeholders, and defining the strategic direction.

#25 in

Dan Shipper automated every process at Every Inc. with AI agents yet saw headcount jump from 4 to 30 since GPT-3. His new report argues that as AI makes expert competence cheap, demand for human experts only grows—and this cycle will intensify as we near AGI.