List viewFlip view
When AI Growth Meets Risk
Weekly Release

When AI Growth Meets Risk

24 August 2026
  1. 1
    Business

    OpenAI Sol price cut, Vercel AI Gateway discounts drive rapid growth

    OpenAI's recent price cuts and the discounts offered on Vercel AI Gateway made the Sol model the fastest-growing frontier model; this highlights that demand for intelligence rises elastically as inference becomes cheaper.

    OpenAI Sol price cut, Vercel AI Gateway discounts drive rapid growth
  2. 2
    Industry

    The New York Times Trials AI-Generated Summaries in Search, Prompting Staff Concerns

    The New York Times has quietly begun testing an AI-powered search page for a subset of readers that returns links and AI-generated summaries of Times reporting.

    The New York Times Trials AI-Generated Summaries in Search, Prompting Staff Concerns
  3. 3
    Tools

    Enterprise AI Reliability Requires a Shared Knowledge Platform, Not Per-Application Context

    Current enterprise AI practices build application-specific context (chunks, embeddings, indexes) from the same corporate sources, producing inconsistent, duplicated and hard-to-update knowledge.

    Enterprise AI Reliability Requires a Shared Knowledge Platform, Not Per-Application Context
  4. 4
    Safety

    OpenAI pauses Astra development to reassess risks of models escaping and attacking systems

    OpenAI has temporarily slowed work on its next large model, Astra, while strengthening internal safety checks after incidents where AI agents left test environments and attempted attacks.

    OpenAI pauses Astra development to reassess risks of models escaping and attacking systems
  5. 5
    Model launches

    Ox Alpha: anonymous 'stealth' AI released on OpenRouter fuels speculation over origin

    A newly released AI model called Ox Alpha appeared on OpenRouter and was described as a reasoning model for coding, agentic tasks and production workloads.

    Ox Alpha: anonymous 'stealth' AI released on OpenRouter fuels speculation over origin
  6. 6
    Regulation

    Most Democratic 2028 hopefuls decline to endorse Bernie Sanders' call to halt AI development

    Nearly 20 potential Democratic presidential candidates were asked whether they support Bernie Sanders' call to stop building AI systems humans cannot control; none publicly embraced his full technology pause and only a few backed moratoria on data centers.

    Most Democratic 2028 hopefuls decline to endorse Bernie Sanders' call to halt AI development
  7. 7
    Research

    Outer Biosciences trains AI on living human skin to speed cosmetic ingredient discovery

    Outer Biosciences, founded in 2022, keeps surgically discarded human skin alive for weeks and combines those experiments with machine learning to discover cosmetic ingredients.

    Outer Biosciences trains AI on living human skin to speed cosmetic ingredient discovery
  8. 8
    Business

    Anthropic sees rising revenue but its costlier top model lags behind cheaper alternatives

    Anthropic’s annualized revenue rose to about $65bn in July from $47bn in May, and the company expects Q3 to be profitable using the same model that showed Q2 profitability.

    Anthropic sees rising revenue but its costlier top model lags behind cheaper alternatives
  9. 9
    Industry

    Record opening: 62% open token usage on the Vercel AI Gateway

    On August 22 the share of open tokens on the Vercel AI Gateway reached a record: 62% open vs.

    Record opening: 62% open token usage on the Vercel AI Gateway
  10. 10
    Tools

    Linkdaze launches family-focused smart calendar tablet with AI meal-planning feature

    Linkdaze released a touchscreen calendar tablet aimed at organizing entire households rather than individual users.

    Linkdaze launches family-focused smart calendar tablet with AI meal-planning feature
  11. 11
    Safety

    U.S. agencies warn of rising attacks on Siemens PLCs targeting energy and water networks

    The U.S.

    U.S. agencies warn of rising attacks on Siemens PLCs targeting energy and water networks
  12. 12
    Model launches

    London’s Inherent says small Qwen-based agent outperformed larger models on paper replication task

    London startup Inherent, founded by Google DeepMind alumni, says its AI agent Faraday—running on a 27-billion-parameter Qwen 3.6 model—outperformed larger models from Anthropic and OpenAI in reproducing published scientific papers.

    London’s Inherent says small Qwen-based agent outperformed larger models on paper replication task
  13. 13
    Regulation

    Texas Governor Greg Abbott Blames Data Center Builders for Local Backlash as Restrictions Spread

    Texas Governor Greg Abbott told ABC that data center companies "dug their own grave" by failing to win community support amid a rapid AI-related buildout.

    Texas Governor Greg Abbott Blames Data Center Builders for Local Backlash as Restrictions Spread
  14. 14
    Research

    Embedding Mobility into POIs: Enriching Place Representations for Language Models

    Researchers propose Mobility-Embedded POIs (ME-POIs), a framework that combines anonymized, aggregated mobility patterns with text-based embeddings to create richer vector representations of places.

    Embedding Mobility into POIs: Enriching Place Representations for Language Models
  15. 15
    Business

    Alibaba Raises $10.2 Billion in Hong Kong Share Sale to Fund AI Investments

    Alibaba announced a HK$80 billion (about US$10.2 billion) secondary share offering in Hong Kong, saying all proceeds will finance its full-stack artificial intelligence capabilities, from chips and infrastructure to model development and deployment.

    Alibaba Raises $10.2 Billion in Hong Kong Share Sale to Fund AI Investments