AI Briefing
TLDR AI news
The latest AI and developer news from TLDR AI, with short summaries.
Visit website →Cognition welcomes the TierZero team
Cognition has acquired incident response automation startup TierZero and will integrate its capabilities into Devin.
Increasing AI spend doesn't improve your supply chain
Pallet has launched a Custom Model dedicated to supply chain teams, stating it will solve the cost and control problems of frontier AI dependence.
Anthropic's Conway Test Set to End — Broader Rollout Possible
Anthropic's internal always-on agent experiment Conway will end on July 24, raising questions about whether it will be fully discontinued or transition into a full launch.
Z.AI Completes Gigawatt-Scale AI Data Center Using Only Chinese-Made Chips
Chinese AI company Z.AI has completed a 1-gigawatt-scale AI data center built entirely with domestically made chips.
Why token prices are falling but AI bills are rising
Per-token prices have plummeted, but actual AI costs have risen instead due to a surge in usage driven by agent architectures.
Xiaomi-Robotics-1
Xiaomi has unveiled Xiaomi-Robotics-1, a robot policy model built on 100,000 hours of embodiment-agnostic pretraining.
Introducing Cosmos 3 Edge
NVIDIA has unveiled Cosmos 3 Edge, a 4B-parameter open world model for edge devices.
Language Model Harnesses Enable Compositional Generalization
Experiments demonstrate that a well-designed LLM harness (RLM) can generalize from training on short tasks alone to tasks 8-32x longer and to new domains.
Online Learning for Cost-Efficient LLM Routing
Ramp cut AI costs by over 25% and reduced error rates using dynamic LLM routing based on online learning.
Kimi K3: Performance and the Controversies That Come With It (70-minute read)
Kimi K3 is the strongest open model available today, but it still trails the closed-model frontier by several months.
Google's New Chip for Gemini
Google is developing a 'Frozen v2' server chip that could be up to 10 times more efficient than existing AI chips, targeting a 2028 release.
AMD's Helios
AMD will ship its first rack-scale AI system, Helios, this year, with Microsoft, Meta, OpenAI, Oracle, and others confirming adoption.
Kimi Work (website)
Moonshot AI has launched Kimi Work, a desktop AI agent that supports local file integration, web automation, and 24-hour scheduling.
Claude Fable 5 to Be Included in All Max and Team Premium Plans
Starting July 20, Claude Fable 5 will be included in Max and Team Premium plans, while Pro and Team Standard users can access it via credits.
Google Preparing to Launch Gemini Live and Skills Features on Web Version
Google is preparing to expand Gemini Live, Gemini's real-time voice mode, and Skills, its user-customization feature, to the web and desktop environment.
Chinese AI startup nears $1 billion in revenue while releasing top-tier models for free
Chinese AI startup Z.ai is boosting its market share by open-sourcing its models and is on the verge of hitting $1 billion in annual revenue.
Alibaba Open-Sources AI Chip Software Stack to Break Free from NVIDIA CUDA Dependency
Alibaba has open-sourced SAIL, its software stack for its own AI chips, in an effort to break the dominance of NVIDIA's CUDA ecosystem.
Kimi Code CLI (GitHub Repository)
Moonshot AI has released Kimi Code CLI, an AI coding agent that can edit code and execute commands in the terminal.
Distributing Responsibility: A Learning Method That Complies with Dale's Principle Without Weight Transport
A new neural network learning mechanism has been proposed that enables training while following Dale's principle, without the Weight Transport problem.
Fable 5 vs GPT-5.6 Sol on an NP-Hard Problem: Does /goal Mode Help
Claude Fable 5 showed superior performance and consistency over GPT-5.6 Sol on an NP-Hard optimization problem.