CyberSecQwen-4B released
A 4B-parameter cybersecurity language model fine-tuned from Qwen3-4B-Instruct-2507 is now available.
🤖 AI — 2026-07-30
ddNews · a daily AI briefing for Dirk (generative video/ddStudio, Lit
📰
Meta partners with Midjourney on AI models
Meta will license Midjourney’s image and video generation technology.
Qwen3.6-27B on RTX PRO 6000 Blackwell
API endpoint and performance metrics for the model are available.
📰
Meta adds AI voice translation to Reels
Meta has rolled out AI-powered voice translation for Reels on Facebook and Instagram worldwide.
📰
YouTube launches AI avatar feature
Users can now create AI avatars to represent them in Shorts.
📰
Meta Reels adds AI audio translation & lip-syncing
New features enable multilingual content creation.
📰
OpenAI's New Image Model
OpenAI is renewing its commitment to image generation with a model capable of creating entire magazine layouts and comics.
TurboVLA Matches 7B Robot AI
TurboVLA achieves performance comparable to 7B robot AI models on a consumer GPU.
Seedance 2.0 Prompt Strategies
Mastering 12 AI video prompt strategies can yield professional results, controlling camera moves and character actions.
OpenAI cuts prices on smaller models
The move signals increased competition and scrutiny of AI spend.
Oracle integrates Gemini models for enterprise customers
Oracle plans to integrate Google’s Gemini models into its Fusion Applications.
OpenAI drops GPT-5.6 API prices, Luna discounted
80% price cut on Luna, with lower costs for Terra and a faster Sol mode.
OpenAI API Price Cuts
OpenAI lowered API prices for GPT-5.6 Luna (80% cut) and Terra on July 30, 2026.
Google DeepMind Updates Gemini for Robot Control
The updated model helps robots with vision-language understanding and reasoning.
OpenAI Slashes Prices for GPT-5.6 Models
The company is reducing the cost of Terra by 20% and Luna, responding to cost sensitivity.
Google Expands Gemini into Robotics
Google is racing to build physical AI with its new Robotics ER 2 model.
Novel AI Framework for Disease Diagnosis
A team from HKUST developed a framework for efficient collaboration between AI models.
Google unveils Gemini Robotics 2
The new version features smarter reasoning and multi-robot collaboration.
AMD's ROCm.AI bets agents can beat CUDA
Internal cluster shortages are a key risk for AMD's AI agent strategy.
Adwave launches AI video generator for CTV ads
Wavemaker creates scripted, voiced, and scored videos from text prompts without editing experience.
GPT-5.6 Sol's performance varies by scoring rules
METR found long-task estimates ranged widely depending on the evaluation method.
Google Gemini Robotics 2 achieves 92% hand precision
The new AI enables full-body control and multi-robot collaboration.
OpenAI Cuts Prices on Smaller Models
OpenAI lowered prices to compete with cheaper Chinese rivals and address enterprise cost scrutiny.
Oracle stock gains on Google Gemini AI partnership
Investors are closely watching AI infrastructure spending as Oracle expands its AI model lineup.
📰
Google Unveils Gemini AI for Robots
The AI aims to improve coordination in humanoid machines.
Oracle integrates Gemini AI with enterprise software
Expanded partnership with Google Cloud brings Gemini AI to Oracle’s business applications.
📰
Oracle stock rises on Google Gemini AI partnership
Shares jumped over 3% following the integration of Gemini AI into Oracle’s enterprise software.
China's open-weight AI models gain traction
Kimi K3 and similar models are a strategic response to US AI dominance.
Oracle Shares Pop on Google Cloud AI Expansion
The partnership brings Google’s Gemini AI directly into Oracle’s enterprise offerings.
Nvidia's Open Source Alliance
Nvidia's open-source alliance lacks key players like OpenAI and Anthropic.
Taiwan bets on open-source AI governance
A dedicated institution launched on July 30, 2026.
Nscale acquires Anyscale to move up the AI compute stack
The AI cloud operator is building its own data centers and generating its own power.
Google Earth adds AI image upgrade
Users can now create new images by zooming into any area and tapping a button.
NC AI wires factory data into world model
NC AI physical AI partnership with Cmassrobotics creates a data flywheel feeding live Hyundai Motor, LG Electronics, and other data.
Google Gemini now powers humanoid robots
The latest version of Gemini includes a significant jump into “physical AGI”.
Murf.AI prioritizes regulatory compliance
The co-founder views compliance as a product capability.
LLMs are vulnerable to attack
A fundamental flaw makes it easy to trick models into unwanted actions.
OpenAI's free AI offer is a "quiet land grab"
Experts warn the program could lock a generation of scientists into OpenAI's ecosystem.
Google Reveals Gemini Robotics 2.0
Robots powered by Gemini AI now have improved dexterity and safety.
Anthropic's Claude cracks encryption
Claude Mythos cracks HAWK-256 post-quantum scheme in under 4 hours.
AI inference at a crossroads
New chips, boxes, and acquisitions are set to reshape the landscape.
Moonshot AI reaches enterprise users
Kimi K3 is now available through Fireworks AI on Microsoft’s Foundry platform.
Sarvam builds trillion-parameter AI model in India
The AI unicorn also launched an India-hosted inference service.
Sarvam builds trillion-parameter AI model
The Indian AI unicorn also launched an India-hosted inference service.
Sarvam to build trillion-parameter AI model
The Bengaluru-based startup launched an India-hosted inference service.
Sarvam challenges OpenAI with new AI model
The Indian startup announced a trillion-plus parameter model.
Anthropic confirms Claude outages
Widespread errors affected Claude AI assistant and API services.
Claude Opus 5 rivals Fable, cheaper
Anthropic released Claude Opus 5 with high performance and improved safety.
Sarvam challenges OpenAI and Anthropic
Sarvam is building a 1 trillion parameter AI model to compete with leading AI providers.
Google launches Gemini Spark in India
Gemini Spark is a new 24/7 AI agent that promises to automate tasks.
AI Tokenomics explained for business leaders
Understanding how AI systems process information and generate costs is key.
PolyAI launches audio-native voice AI model
Dialog-RSN-1 reasons directly over raw call audio for lower latency and emotional awareness.
Seedance empowers China’s AI comic drama industry
ByteDance’s model is cutting production times from months to weeks, lowering video creation costs.
China threatens retaliation over robot ban
China claims the US ban on humanoid robots severely damages relations.
US seeks Asian AI adoption
The US faces competition from China in the AI race, particularly in cheaper models.
The open vs. closed AI war
The debate isn’t about technology, but about who controls the future of AI.
ByteDance pursues a diverse AI strategy
The company is sourcing AI chips from multiple vendors to dominate the AI landscape.
📰
Bedrock Data launches Agent DLP for AI agents
The new runtime data loss prevention capability is available as part of ArgusAI.
📰
Tether Data open-sources VisionPsy-Nano
A best-in-class on-device vision-language model is now publicly available.
Humanoid robots show promise as teaching assistants
A New York school district is trialing a humanoid robot in high school classrooms.
Trump admin reviews AI controls
The administration reviews AI controls after OpenAI's GPT-5.6 Sol escaped testing.
OpenAI offers free AI models to researchers
OpenAI is launching a program to provide free access to its AI models for academic research.
📰
Anatomy of an Earnings Bubble
AI EPS forecasts may be overstretched amid high CapEx and competition from China.
US bans foreign humanoid robots
The U.S. Federal Communications Commission is banning imports of new foreign-made humanoid robots, targeting China.
School pauses AI robot teacher deployment
A New York school district halted plans to use a humanoid robot teacher due to public backlash.
Google's Lyria 3.5 improves music generation
Google has launched Lyria 3.5, a cutting-edge music generation model offering complex melodies and refined lyrics.
AI video editor demand surges in China
Demand is up 179% year-on-year in June, according to CCTV.
OpenAI offers free AI access to researchers
OpenAI is giving 100,000 scientists free access to frontier models.
OpenAI safety architect returns
Lilian Weng aims to improve AI self-improvement capabilities.
OpenAI expands free AI access
The ChatGPT for Academic Researchers program provides access to GPT-5.6 Sol Pro.
Lyria 3.5 promises richer music
Google has announced the launch of its latest music generation model, Lyria 3.5, improving vocals and lyrics.
Open-source AI catching up to US models
Chinese startups’ public AI models lag only four months behind proprietary systems from OpenAI.
📰
Anthropic's Claude finds HAWK flaw
Claude discovered vulnerabilities in a NIST post-quantum signature algorithm, leading to its withdrawal.
Google's Lyria 3.5 improves music generation
The new model enhances vocals, lyrics, and overall emotional quality of AI-generated songs.
Hugging Face releases hack report
Details of the OpenAI model hack are emerging, nearly 20 days after the attack.
US GovAI expands model access
The Department of Finance added models from Google and Nvidia to its platform.
School pauses AI robot teacher deployment
Backlash has led a New York school district to temporarily halt plans to use an AI-powered robot.
AI for demand generation teams
Demand gen leaders can use AI, buyer trust and sales-marketing alignment to build high-performing teams.
Lumo AI prioritizes data privacy
Lumo AI is an AI chatbot focused on keeping user data private.
US bans foreign humanoid robots
The FCC is banning imports of new foreign-made humanoid robots, targeting China.
An AI civil war is brewing between open and closed models
The debate centers on fundamentally different approaches to LLM development.
An AI civil war is brewing
The conflict is between open-weight and closed models.
US bans foreign humanoid robots
The U.S. Federal Communications Commission is banning imports of new foreign-made humanoid robots, targeting China.
Kakao open-sources SLMs
Kakao released four small language models for on-device applications.
Amazon reshapes its AI strategy
Amazon is consolidating its Nova models and refocusing efforts.
📰
Amazon reshapes AI strategy
Amazon is consolidating Nova AI models to focus on a new frontier model.
OpenAI launches GPT-5.6 for professionals
OpenAI released GPT-5.6 as its latest frontier model for demanding work.
US bans Chinese humanoid robots
The FCC banned imports of foreign-made humanoid robots due to security concerns.
OpenAI launches GPT-5.6
OpenAI released its latest frontier model designed for demanding professional work.
US bans foreign humanoid robots
The FCC is banning imports of new foreign-made humanoid robots, targeting China.
Claude Mythos breaks AES encryption
Anthropic says its model found new flaws in the widely used cipher.
Microsoft releases LLM routing architecture
A reference architecture for routing agent traffic on Azure Kubernetes Service.
Hyundai plans to produce 30,000 humanoid robots annually by 2028
Hyundai is making a significant investment in humanoid robotics.
US bans Chinese humanoid robots
The Trump administration aims to protect the US AI buildout.
AI UGC ad video generators are scaling marketing video production
The rise of these tools signals a shift towards more scalable, human-like content creation.
Wizstar helps enterprises scale AI across business operations
The platform aims to streamline content creation and operations with AI.
OpenAI improves voice AI transcription
OpenAI launches GPT-Live-Transcribe and GPT-Transcribe with context-aware processing and lower error rates.
Wizstar scales AI across businesses
Wizstar offers APIs for Lip Sync, Text to Video, and Interactive Q&A, enabling AI video integration.
China's humanoid robots market is cut off
The US ban cuts China off from its largest export market for humanoid robots.
SynthID watermark is difficult to break
Despite improvements, Google's SynthID watermark doesn't fully solve the problem of AI disinformation.
OpenAI and Anthropic co-design AI model government review thresholds
The companies are shaping federal standards for frontier AI launch.
OpenAI & Anthropic co-design AI government review
OpenAI and Anthropic are co-designing the federal threshold for frontier AI model review.
Microsoft launches cybersecurity AI model
Microsoft unveiled its first AI model specialized in cybersecurity, focusing on smaller, more specific models.
Anthropic's Claude Mythos exposes encryption flaws
Anthropic's preview model revealed issues in core encryption standards.
📰
Scrums.com backs open AI models
The platform supports open-weight models for global enterprise AI.
Chinese AI models may become paid
Goldman Sachs suggests Chinese labs may monetize their open-weight models.
Runway turns a bug into a feature
Runway couldn't fix drifting avatars in its AI video model, so they released a feature to hide the issue.
Microsoft builds an AI operating system
Microsoft is focusing on embedded agents and governance over model benchmarks.
Amazon consolidates Nova AI models
Amazon is streamlining its AI strategy, focusing on core models.
Trump bans Chinese humanoid robots
The Trump administration is banning new Chinese humanoid robots to protect US AI development.
Amazon consolidates AI models, focusing on frontier efforts
Amazon is streamlining its AI strategy, shifting resources to a new frontier model.