Anthropic Expands Amazon Collaboration for 5GW Compute

#1 𝕏

Anthropic is expanding its collaboration with Amazon to secure up to 5 GW of compute for training and deploying Claude. Capacity begins coming online this quarter, with nearly 1 GW expected by the end of 2026.

#2 𝕏

Google AI now offers Pro and Ultra subscribers increased usage limits and direct access to Nano Banana Pro and Gemini Pro models in Google AI Studio—no API key required.

#3 𝕏

Rowan Cheung: Boston Dynamics has embedded Google DeepMind’s Gemini Robotics model into its Spot robot, giving it embodied reasoning—enabling autonomous tasks like reading pressure gauges—and multi-view environment analysis with built-in success detection.

#4 📝 Anthropic Engineering

Scaling Managed Agents: Decoupling the brain from the hands - Discusses architecture and design principles for scaling managed agents by separating the decision-making 'brain' from execution 'hands', enabling more robust, scalable agent systems. The article examines tradeoffs and system patterns for large-scale agent management.

#5 ▶️

How Intercom 2X'd engineering velocity with Claude Code | Brian Scanlan

How I AI Podcast

#6 𝕏

Cognition launched a parallelized Devin system: a main Devin spins up and coordinates multiple full Devin sessions—each with its own VM, terminal, browser, and testing environment—to break down and solve complex tasks concurrently.

#7 𝕏

There's An AI For That links to Netflix’s new void-model GitHub repo, open-sourcing their text-to-video diffusion framework complete with pretrained weights, Colab demos, and inference scripts.

#8 𝕏

Santiago unveiled Kimi K2.6, an open-source coding model hitting SOTA on HLE with tools (54.0), SWE-Bench Multilingual (76.7), and Pro (58.6). It also boosts front-end development, DevOps, performance optimizations, long-horizon coding, and cross-language generalization.

#9 𝕏

Guillermo Rauch clarifies that Vercel’s “All environments” setting encrypts secrets at rest but only decrypts them in development (via vc dev and vc env pull), and warns against reusing the same API key across dev, preview/staging, and production.

#10 𝕏

Claude now builds live dashboards and trackers in Cowork connected to your apps and files, refreshing with current data each time you open them.

#11 𝕏

Harrison Chase breaks down the end-to-end infrastructure for deploying long-horizon AI agents—covering task scheduling, state persistence, vector retrieval, prompt templating, and monitoring.

#12 ▶️

Hermes Agent: The New OpenClaw?

Greg Isenberg

#13 𝕏

bolt.new launched live multiplayer in Bolt, so you can open any project to see who’s in it, track typing in real time, and access the full shared chat history—one project, one thread, one source of truth.

#14 𝕏

Harrison Chase argues memory will be the ultimate lock-in and everyone’s racing to ship it first. He insists that memory—citing OpenAI’s Codex/Chronicle memories—should be open.

#15 𝕏

Cursor introduced the `/statusline` command, letting you customize your CLI status bar to surface the exact session and runtime signals you care about.

#16 𝕏

Dharmesh Shah has long expected a frontier model company to ship an AI memory/tool for coding and highlights OpenAI’s new Codex “Chronicle” feature, which persistently stores and recalls code context to boost developer productivity.

#17 𝕏

Lenny Rachitsky distills Nikhyl’s 10-point playbook for PM reinvention in the AI era, urging AI-first hiring, PMs as change agents, and rediscovering the joy of building to navigate the chaos ahead.

#18 in

Alana W. spent two days with ~80 EPD leaders exploring how AI is overturning the assumption that production is costly, giving us license to rethink sprints, handoffs and PRDs.

#19 𝕏

Lenny Rachitsky forecasts that over the next 12–24 months companies will conduct massive layoffs (e.g., shedding 30,000 roles) and then rehire a leaner, AI-first workforce of roughly 8,000 people.

#20 𝕏

Yann LeCun predicts AI will boost average annual GDP growth by about 0.6% (from 2.5% to 3.1%), compounding to 36% over ten years instead of 28% and leaving GDP 7.7% larger.

#21 in

Claire Vo shows how Brian Scanlan’s Intercom team doubled R&D productivity by setting AI-adoption targets, tracking leading indicators with dashboards, and using Claude telemetry to measure quality.

#22 𝕏

There's An AI For That: Netflix’s VOID AI model not only erases objects from video but also recalculates scene physics to seamlessly restore the background.

#23 𝕏

Anthropic secured an extra $5 billion investment from Amazon today, with commitments that could scale up to $20 billion more for future compute capacity.

#24 𝕏

Hugging Face announces Kimi.ai’s K2.6 SOTA open-source coding model, which tops benchmarks on long-horizon coding tasks and powers scalable agent swarms—available at platform.kimi.ai and api.moonshot.cn (see the full write-up at kimi.moonshot.cn/blog/k2-6).

#25 𝕏

Logan Kilpatrick announces that Google AI Pro and Ultra subscriptions now integrate with Google AI Studio, letting you code in the playground with higher rate limits—available now.