Hardware•Jul 8, 10:38 PM
China Weighs AI Export Curbs as DeepSeek Develops Own Chip
Beijing considers restricting AI model access abroad, mirroring US export controls, as domestic firms like DeepSeek push for self-sufficiency.
Share:
Newsletter
THE DAILY RIFT
Five shifts. Five minutes. No noise.
- •Verified technology intelligence
- •Emerging patterns before the hype
- •Direct action steps for builders & founders
No spam. Unsubscribe anytime. Powered by Beehiiv.
Today's Intelligence
FoundersAug 29
Hugging Face's Microduck Robot Tops $1M Sales Amidst Humanoid Hype
While venture capital floods humanoid robotics, Hugging Face's $399 open-source duck robot demonstrates a potent alternative business model.
●Builder Action:Developers can now access a fully open-source, trainable robot platform for under $400. This provides an excellent, low-cost entry point for experimenting with AI, reinforcement learning, and robotics, with the freedom to modify both hardware and software. The LeRobot ecosystem offers a standardized stack, reducing integration friction for custom projects.
CybersecurityAug 29
Sleepwalker: New Passive Backdoor Uses Custom Command Language
Researchers unveil 'Sleepwalker,' a stealthy backdoor malware that evades detection by using a unique, encrypted command and control language.
●Builder Action:Developers should be aware that custom, encrypted C2 protocols like Sleepwalker's can bypass traditional network security tools. Implement robust logging and anomaly detection within your applications and infrastructure. Consider building in self-checking mechanisms or out-of-band communication channels for critical services that could be compromised.
CybersecurityAug 29
GrapheneOS Ditches Pixel 11 Support for Hardware Memory Tagging
GrapheneOS removes support for Pixel 11's Memory Tagging, citing reliability and security concerns with the hardware.
●Builder Action:Developers relying on MTE for debugging memory corruption issues on Pixel 11 with GrapheneOS will need to find alternative debugging strategies. GrapheneOS's robust software mitigations remain, but the hardware-assisted MTE capabilities are now offline. This decision highlights the importance of validating hardware security features in real-world, hardened OS environments.
Original Deep Intelligence
RIFT REPORTDEEP INTELLIGENCE
Weekly Rift: AI Agents Mature, GPU Costs Soar, and Security Lapses Persist
AI agents are moving beyond simple tasks to complex workflows, while GPU shortages continue to drive up costs, and persistent security vulnerabilities plague the tech landscape.
2 minread time
Updated:24 Aug 2026
Funding Radar
GBP0Rolling
Seed and Venture CapitalEnterprise Ireland
EUR15.0MRolling
Start-up FundingEnterprise Ireland
EUR50kRolling
IRAP Youth Employment ProgramIRAP Canada
CAD60kRolling
Innovation Assistance Program (IAP)IRAP Canada
CAD500kRolling
Tool Radar
Enter ProAn AI-native platform for building and scaling applications.
paidDiscovered
GitNexusAn open-source kernel for coding agents, enabling AI-driven software development.
freeDiscovered
IQ RoutingTrajectory-aware LLM routing to reduce agent costs.
paidDiscovered
GLM-5.3-FlashThe first natively multimodal model in the GLM-5 series, offering advanced capabilities.
paidDiscovered
TracciaTraccia is a vendor-neutral AI Agent Control Plane for managing AI agents.
paidDiscovered
Latest Feed
AI Models•Jul 8, 10:38 PM
Vercel CEO: Separate AI Models from Agents for Production Efficiency
Guillermo Rauch argues for decoupling AI models from their agents to optimize for cost and performance in production environments.
Share:
Developers•Jul 8, 10:38 PM
GitHub Managed Settings and Ota: Distinct Roles in Repository Governance
GitHub Managed Settings handles platform-level repository configuration, while Ota focuses on execution governance for repositories.
Share:
AI Models•Jul 8, 10:38 PM
1,200-Year-Old Islamic Scholarship Method Inspires New AI Trust Framework
A novel approach adapts classical hadith verification techniques to assess reliability in multi-agent AI systems.
Share:
Developers•Jul 8, 10:38 PM
Meteor 3.5 Arrives with Major Overhaul of Reactive Data Layer
Meteor 3.5 ships a broad update, enhancing its reactive data layer, database drivers, auth middleware, and Node runtime.
Share:
AI Models•Jul 8, 10:38 PM
Occam's Razor: The 700-Year-Old Principle Guiding AI Development
The centuries-old principle of preferring simpler explanations is proving vital in the fight against AI overfitting.
Share:
AI Models•Jul 8, 10:38 PM
Anthropic Proposes 'Global Workspace' for LLMs
New research suggests a unified memory system could unlock more complex reasoning and collaboration in AI.
Share:
Founders•Jul 8, 10:38 PM
AI's $725B Spending Spree: Is It a Bubble or Sustainable Demand?
Massive AI investments face scrutiny over circular funding and revenue demands.
Share:
Developers•Jul 8, 10:38 PM
Digital Marketplace Launched: Next.js, Escrow, Auto-Verification
Production-ready multi-vendor marketplace with secure escrow and email receipt-based payment verification is available.
Share:
Founders•Jul 8, 10:38 PM
AI Law Startup Norm Secures $120M, Achieves $1.2B Unicorn Valuation
Norm, an AI-powered legal tech platform, closes a $120 million Series C led by Khosla Ventures, reaching a $1.2 billion valuation.
Share: