RobinsonLabshomelab · ai · ops
CLUSTER · 16 GPU NODES · 40G RDMA
The archive
News · edition

Edition, Tue, 04 Aug 2026.

5 stories on the wire this day. Open any story for its sources. Summaries are written by software from the linked reporting and can be wrong. Spot an error? Email [email protected] and we will correct or remove it.

AI & Technology

The European software firm’s first purchase since its Nasdaq IPO signals a major consolidation in the enterprise data layer.

2 sources
Lead
this edition
Still developing
Anthropic debuts Mythos-class AI with record coding and science leapsClaude Fable 5 and restricted Mythos 5 set new benchmarks, solving a 50-million-line code migration in a day and generating novel drug candidates autonomously.The Decoder
OpenAI unveils Jalapeño, its first custom inference chip designed with BroadcomThe new processor aims to optimize performance-per-watt for large language model inference, reducing reliance on Nvidia GPUs.TechCrunch
ServiceNow hits $1B AI ACV and raises full-year outlook in strong Q2 2026 earningsServiceNow reported Q2 2026 revenue of $3.987B (up 24% YoY) and announced its AI platform crossed $1B in annual contract value, prompting a raised full-year subscription revenue forecast.Outlet
Meta raises 2026 AI infrastructure spending floor to $130 billion as stock dips on ROI questionsCEO Mark Zuckerberg cited confidence in the investment despite market pushback on the massive capital expenditure increase.Fortune
SpaceXAI launches Grok 4.5 as an Opus-class model at lower costThe new model excels at coding, agentic tasks, and knowledge work with twice the token efficiency of competitors.SpaceXAITechCrunch
xAI releases Grok 4.5 with 1.5 trillion parameter MoE architecture and competitive pricingThe new model launched July 9 alongside GPT-5.6, featuring training on Cursor data and achieving 83.3 percent on Terminal-Bench 2.1 while using approximately 25 percent fewer output tokens than Opus 4.8.AIapps
DeepSeek V4-Flash emerges as cheapest major AI model at 3 cents per testResearch firm Artificial Analysis found DeepSeek's V4-Flash costs roughly 3 cents to run through a full benchmark suite, making it by far the least expensive well-known model compared to 86 cents for Kimi K3, $1.86 for GPT-5.6 Sol, and $3.15 for Claude Fable 5.The Next WebYahoo Finance

AI & Technology

AI & Technology
The US regulatory body signals a deepening tech divide as the Trump administration pushes to protect domestic AI infrastructure.
2 sources
AI & Technology
The new open-weight model generates 15-second 2K videos but is barred from use in the US, EU, UK, and South Korea.
1 source
AI & Technology
One of the largest open models to date, it aims to push the frontier of accessible large-scale reasoning.
1 source
AI & Technology
CrowdStrike reports an 89% year-over-year rise in AI-driven cyber activity, including notable actions by North Korean operator “Stardust Chollima.”
1 source
AI & Technology

Bending Spoons acquires Airtable in a $1.285B all-cash deal

The European software firm’s first purchase since its Nasdaq IPO signals a major consolidation in the enterprise data layer.

The European software firm’s first purchase since its Nasdaq IPO signals a major consolidation in the enterprise data layer.

Bending Spoons has agreed to buy Airtable for $1.285 billion in cash, with the total equity value reaching approximately $2.25 billion. This marks the Italian company’s first acquisition following its IPO and is expected to close by the end of 2026. The deal reflects a broader trend of mature SaaS players seeking to expand their platform capabilities through strategic acquisitions rather than organic growth alone.

2 sources
AI Weekly
OrangeBot.AI
AI & Technology

FCC drafts proposed ban on Chinese AI data-center transceivers and humanoid robots

The US regulatory body signals a deepening tech divide as the Trump administration pushes to protect domestic AI infrastructure.

The US regulatory body signals a deepening tech divide as the Trump administration pushes to protect domestic AI infrastructure.

The Federal Communications Commission has proposed restricting imports of new Chinese-made optical transceivers for data centers, alongside a ban on new Chinese humanoid robots and grid inverters. This follows the White House’s strategic pivot to promote US AI models abroad while relaxing some open-source restrictions. Chinese officials reportedly view Anthropic’s “Mythos” model as a potential cyber weapon ahead of the upcoming summit between the US and China.

2 sources
AI Weekly
OrangeBot.AI
AI & Technology

MiniMax releases H3, a 33B video-generation model with strict regional licensing

The new open-weight model generates 15-second 2K videos but is barred from use in the US, EU, UK, and South Korea.

The new open-weight model generates 15-second 2K videos but is barred from use in the US, EU, UK, and South Korea.

MiniMax has launched H3, a 33-billion-parameter video generation model capable of producing up to 15 seconds of 2K-resolution video. Due to evolving regulations around video-generation AI, the community license explicitly prohibits use in the United States, European Union, United Kingdom, and South Korea. This release highlights both the rapid advancement in generative video capabilities and the increasing complexity of global licensing constraints.

1 source
AI Weekly
AI & Technology

LG ships K-EXAONE 2.0, a 750B open-weights Mixture-of-Experts model

One of the largest open models to date, it aims to push the frontier of accessible large-scale reasoning.

One of the largest open models to date, it aims to push the frontier of accessible large-scale reasoning.

LG has released K-EXAONE 2.0, a 750-billion-parameter Mixture-of-Experts (MoE) model available as open weights. This release continues the trend of leading institutions making increasingly large models available to the broader developer community, enabling more flexible deployment and fine-tuning options for specialized applications.

1 source
AI Weekly
AI & Technology

Cybersecurity threats surge with AI-enabled adversaries, prompting new regulatory and corporate responses

CrowdStrike reports an 89% year-over-year rise in AI-driven cyber activity, including notable actions by North Korean operator “Stardust Chollima.”

CrowdStrike reports an 89% year-over-year rise in AI-driven cyber activity, including notable actions by North Korean operator “Stardust Chollima.”

The 2026 Threat Hunting Report from CrowdStrike notes an 89% increase in AI-enabled cyber adversary activity. Among the most active groups are China-Nexus entities, which exploit critical vulnerabilities within 24 hours of public PoC release, and North Korea’s “Stardust Chollima,” which has poisoned 131 npm packages with malicious code. In parallel, Anthropic’s Claude models successfully engaged three real organizations during internal cyber evaluation tests, underscoring the growing intersection of AI capability and cyber security dynamics.

1 source
AI Weekly
Every edition

The wire publishes daily and keeps an archive of every edition, listed by date, newest first. If a summary is wrong, we correct or remove it.

Back to the archive Latest edition