The full desk, newest first.
Every story Pulse24 is tracking, in chronological order — newest first.
Antirez Releases DeepSeek V4 Engine
Antirez's `ds4` engine enables DeepSeek V4 Flash's 1 million token context to run locally on 128GB RAM MacBooks via Metal, using specialised 2-bit quantization and disk-based KV cache. This expands local inference capabilities for long-context AI tasks, reducing cloud reliance for developers.
Unsloth, NVIDIA Accelerate LLM Training
LLM development teams gain significant operational efficiency as Unsloth and NVIDIA accelerate training by 25%. This reduces compute costs and time for fine-tuning models, lowering the hardware barrier for platform engineers and founders deploying custom AI.
SpaceX Grants Anthropic Supercomputer Access
Anthropic gains access to SpaceX's Colossus 1 AI supercomputer, boosting Claude Pro and Max capacity and API limits. This deal highlights the intense competition for scarce compute and signals a potential shift towards orbital AI infrastructure to overcome terrestrial constraints.
Nvidia, Corning Advance AI Optics
Nvidia and Corning's partnership to build AI optical infrastructure in the US, including three new facilities and a tenfold capacity increase, signals a critical shift towards optical interconnects. This impacts data center design and procurement strategies for high-performance AI.
OpenAI, Partners Release AI Networking Protocol
OpenAI, AMD, and partners released Multipath Reliable Connection (MRC), an open protocol improving GPU networking for large AI training clusters. This standardises network reliability and efficiency, directly impacting training throughput and predictability for platform engineers.
DeepSeek Aims For $45B Valuation
DeepSeek's potential $45 billion valuation in its first funding round, led by China's state investment fund, signals a strategic national push for AI self-sufficiency. This accelerates the development of a vertically integrated AI stack, reducing reliance on foreign technology for Chinese platform engineers and procurement teams.
Tilde Launches Reversible AI Agent Sandboxes
Production teams can now deploy AI agents with reduced operational risk. Tilde.run's new platform treats every agent run as a reversible transaction within an isolated sandbox, preventing unintended data modifications and exfiltration, and allowing instant rollbacks.
Google AI Search Incorporates Forum Quotes
Google's AI search now integrates quotes from forums and social media, shifting the burden of source credibility onto users. This introduces new trust vectors for decision-makers, as AI-generated content now directly incorporates unverified public discussions.
Anthropic Settles Copyright, Faces Claims Issues
Anthropic's $1.5 billion copyright settlement faces significant operational issues, as authors struggle with a glitchy claims website. This highlights the practical challenges and reputational risks for legal and compliance teams managing large-scale AI training data liabilities.
OpenAI Violated Canadian Privacy Laws
Global AI model deployment now faces heightened regulatory enforcement after Canadian watchdogs found OpenAI violated privacy laws during ChatGPT training. This mandates immediate review of data sourcing and user rights mechanisms for all AI developers.
AI Designs Next-Generation Disinfectants
AI-driven research from Emory, George Mason, and Villanova Universities yielded 11 new disinfectants effective against "superbugs." This accelerates the discovery of critical compounds, offering a mechanism to rapidly identify and synthesise solutions against growing antimicrobial resistance.
SAP Acquires Prior Labs, Funds AI Lab
SAP acquired Prior Labs, committing €1 billion to an AI lab for tabular foundation models. This centralises control over enterprise AI access, mandating specific agent architectures and limiting choices for platform and procurement teams within SAP environments.
Google Expands Gemini File Search
Google's Gemini API File Search now processes multimodal data, including images, and adds custom metadata filtering and page citations. This reduces complexity for platform engineers building RAG systems, improving accuracy, speed, and verifiability for compliance and trust.
Nova Intelligence Secures SAP AI Funding
Nova Intelligence secured $31.5 million to accelerate SAP S/4HANA migrations. Its agentic AI platform reduces manual effort by 75% and costs by 50% for enterprise code modernization, addressing a mandatory 2030 deadline and an $89 billion services market.
Google Chrome Silently Installs 4GB AI Model
Google Chrome is silently installing a 4GB AI model, Gemini Nano, onto user devices without consent. This unconsented deployment raises significant privacy and environmental concerns, which the source argues breaches ePrivacy and GDPR, and incurs substantial CO2 emissions.
Anthropic Forms AI Services Venture
A new AI services company formed by Anthropic, Blackstone, Hellman & Friedman, and Goldman Sachs will accelerate AI adoption for mid-size enterprises. This initiative provides direct access to Anthropic's Claude models and engineering support, addressing a critical resource gap for smaller firms.
Apple Explores Intel, Samsung for Processors
Apple is exploring Intel and Samsung for processor manufacturing in the US, diversifying its supply chain beyond TSMC. This move responds to AI-driven demand and recent supply chain shortages, signalling a broader industry shift towards supply resilience for hardware architects and procurement teams.
YC Holds $5B OpenAI Stake
Y Combinator's undisclosed $5 billion stake in OpenAI, held since its 2016 seeding, creates a potential conflict of interest for YC co-founder Paul Graham when publicly endorsing Sam Altman. This highlights the need for transparent financial disclosures in leadership assessments.
Google Adds Gemini API Webhooks
Google's new Gemini API Webhooks reduce friction and latency for long-running agentic applications. Replacing continuous polling with push-based notifications, this update improves efficiency for developers and architects building complex AI workflows, aligning with Google's broader agentic push.
HUMAIN, AWS Launch AI Agent OS
Enterprise architects gain a unified operating system for deploying autonomous AI agents at scale, shifting from fragmented application ecosystems. Available globally via AWS Marketplace, it provides built-in security, data sovereignty, and regulatory compliance, addressing critical governance for regulated industries.