List viewFlip view
AI, Power, and Fragility
Weekly Release

AI, Power, and Fragility

29 June 2026
Vienna to Open Center for Digital Humanism to Institutionalize Human-Centered AI
Regulation

Vienna to Open Center for Digital Humanism to Institutionalize Human-Centered AI

Vienna, together with Austrian federal bodies and several universities, will establish the Center for Digital Humanism Vienna (CDH Vienna) to promote democratic, ethical and human-centered approaches to digital technology.

Read the story
High memory prices driven by AI push Valve’s new Steam Machine into premium price bracket
Industry

High memory prices driven by AI push Valve’s new Steam Machine into premium price bracket

Valve revealed retail prices for the Steam Machine in a Monday update, but high component costs—especially for memory driven up by demand from AI data centers—have pushed the small-PC towards a much higher price than many expected.

Read the story
Mythos / Sol's offensive and defensive cyber capabilities both pose risks
Safety

Mythos / Sol's offensive and defensive cyber capabilities both pose risks

The cyber defense capabilities developed by Mythos / Sol can be used for both offensive and defensive purposes; if adversaries obtain similar offensive capabilities, this could endanger American companies that do not recognize their hidden vulnerabilities.

Read the story
Apple Seeks U.S. License to Buy Memory Chips from Blacklisted Chinese Supplier
Business

Apple Seeks U.S. License to Buy Memory Chips from Blacklisted Chinese Supplier

Apple has requested a special U.S.

Read the story
Vercel paid special attention to the observability of AI agents
Tools

Vercel paid special attention to the observability of AI agents

The Vercel team made built-in observability for AI agents a priority because such systems consist of non-deterministic AI models and multi-actor, distributed API calls that are prone to errors and…

Read the story
ESMFold2 uses transformer embeddings to predict structures of proteins, DNA, RNA and ligands
Research

ESMFold2 uses transformer embeddings to predict structures of proteins, DNA, RNA and ligands

ESMFold2, released by Biohub and EvolutionaryScale, is an open-source structure prediction system that treats biological sequences and molecular descriptions like language by using transformer-based embeddings.

Read the story
Dani Rodrik on the social risks of leaving technological transition to giant corporations
Regulation

Dani Rodrik on the social risks of leaving technological transition to giant corporations

Dani Rodrik, Harvard Kennedy School professor of international political economy, warned in Budapest that allowing tech giants to steer technological change would deepen social inequalities and produce worse shocks than past globalization.

Read the story
Google limits Meta's access to Gemini models amid compute shortages
Industry

Google limits Meta's access to Gemini models amid compute shortages

Google told Meta in March that it could not fully meet the social network's demand for access to its Gemini AI models, and has since applied restrictions that have delayed several Meta internal AI projects.

Read the story
Anthropic: Claude Mythos 5 restorable for U.S. critical infrastructure, work underway on Fable 5 access
Safety

Anthropic: Claude Mythos 5 restorable for U.S. critical infrastructure, work underway on Fable 5 access

Since June 12, Anthropic has been cooperating with the U.S.

Read the story
OpenAI introduces the GPT‑5.6 Sol model with reinforced safety protections
Model launches

OpenAI introduces the GPT‑5.6 Sol model with reinforced safety protections

OpenAI announced the GPT‑5.6 Sol model, for which they have reinforced real-time protection against high-risk cyber activity and repeated abuse; the system was strengthened with human red teaming and…

Read the story
SoftBank’s Market Value Now Hinges Largely on Its OpenAI Stake
Business

SoftBank’s Market Value Now Hinges Largely on Its OpenAI Stake

SoftBank’s Tokyo-listed shares plunged as much as 13% after the New York Times reported OpenAI may delay its IPO to next year.

Read the story
How we teach coding agents our design guidelines
Tools

How we teach coding agents our design guidelines

The authors of the article present how coding agents can be designed and governed to automatically follow an organization’s design and coding guidelines.

Read the story
Shared AI Hiring Assessments Remove ‘Clean Slate’ for Jobseekers
Research

Shared AI Hiring Assessments Remove ‘Clean Slate’ for Jobseekers

A team from Stanford HAI, Chapman University and Northeastern University released the largest audit to date of algorithmic hiring tools, analysing 3.37 million applicants and 4.19 million applications processed by Pymetrics.

Read the story
Swiss watchdog warns AI is amplifying banks' cyber risk
Regulation

Swiss watchdog warns AI is amplifying banks' cyber risk

Marlene Amstad, president of the Swiss Financial Market Supervisory Authority (FINMA), warned that artificial intelligence is significantly increasing cyber and national security risks for banks and urged faster adoption of supervisory technologies.

Read the story
1 / 16