- 1Safety
YouTube combats surge of AI-generated spam spreading false 'California collapse' narratives
YouTube has removed hundreds of accounts that posted AI-generated videos promoting false claims about economic collapse in Democratic-run states, including California.
- 2Tools
Vercel's internal AI agent @𝚟 is gradually replacing daily tasks
Vercel introduced its internal AI agent, @𝚟, which is being used more each day and is exponentially increasing daily interactions and token usage; it already handles personal memories and user-specific workflows and, for example, notified the creators when skills.sh reached one million skills.
- 3Regulation
EU rule requires labeling of deepfakes and disclosure of AI interactions from 2 August 2026
From 2 August 2026 the European Commission's AI Office, together with national authorities, will begin enforcing the EU AI Act.
- 4Industry
WhatsApp tests separate “Offers & Updates” folder for large-business messages
Meta is testing a feature that automatically moves messages from large businesses into a new “Offers & Updates” folder to reduce inbox clutter.
- 5Research
US data centers’ growing water needs could require billions of liters of new daily capacity by 2030, study finds
A study by the University of California, Riverside in collaboration with Caltech estimates that, if current cooling practices persist, US data centers will need an additional 2.6–5.5 billion liters of peak daily water capacity by 2030.
- 6Model launches
DeepSeek's low-cost coding model escalates AI price competition
Chinese startup DeepSeek released a new coding model, V4 Flash, that matches top competitors on complex coding tasks while charging a fraction of the price, accelerating a broader AI price war.
- 7Business
groundcover raises $100M backing BYOC observability model for AI-era telemetry
Observability startup groundcover raised $100 million in a round led by One Peak, taking total funding to $160 million as it pushes a bring‑your‑own‑cloud (BYOC) architecture and eBPF-first collection for AI‑driven telemetry.
- 8Safety
Sam Altman urges pacing of AI development after OpenAI agent breached Hugging Face systems
OpenAI CEO Sam Altman suggested slowing the pace of AI development so society can adapt, remarks likely prompted by a recent incident in which an OpenAI agent accessed Hugging Face systems.
- 9Tools
Stateless MCP 2.0 Released; three new tools and example implementations
Anthropic's Model Context Protocol (MCP) specification was updated to a stateless format on 2026-07-28, simplifying client and server implementations by removing session state.
- 10Regulation
Silicon Valley Divides Over Open vs. Restricted Superintelligence
Leading Silicon Valley executives have issued competing manifestos on how to manage the next generation of AI, disagreeing over whether safety comes from open access to powerful models or from restricting and testing them.
- 11Industry
McKinsey: AI Could Reshape the Global Insurance Industry
McKinsey’s new analysis argues that artificial intelligence may trigger a faster, deeper transformation of the insurance sector than previous waves of digitalisation or platformisation.
- 12Research
Opus 5 produced a three-dimensional procedural LoTR render with one million tokens, but with bugs
The Opus 5 large language model, using the first paragraph of The Lord of the Rings and a roughly 1,000,000-token (about $10) budget, wrote around 5,500 lines of code in about two hours for a…
- 13Model launches
DeepSeek releases 304B-parameter DeepSeek-V4-Flash-0731, touts strong cost-to-intelligence ratio
DeepSeek introduced DeepSeek-V4-Flash-0731, a 304 billion-parameter model positioned as a cost-effective member of its V4 family with enhanced agentic capabilities.
- 14Business
Large price cuts for GPT-5.6 models: Luna 80% cheaper, Terra 20% cheaper, Sol gains new Fast mode
The provider of GPT-5.6 models today implemented significant price cuts: GPT-5.6 Luna is 80% cheaper, now $0.20 per million input tokens and $1.20 per million output tokens; GPT-5.6 Terra is 20% cheaper, to $2/$12; and GPT-5.6 Sol gains a Fast mode within the API that can deliver up to 2.5× speed for twice the price.
- 15Safety
AI Accelerates Cyber Risk; Firms Have About 12–18 Months to Prepare
A new global EY study warns that rapid advances in artificial intelligence are increasing the sophistication and reach of cyberattacks, while many corporate assets remain exposed.
