- AInauten.net
- Posts
- 👀 Weekly AI news: Did you miss it?!
👀 Weekly AI news: Did you miss it?!
👨🚀 The most important AI updates at a glance
Die AInauten
August 10, 2025
👀 Weekly AI news: Did you miss it?!
👨🚀 The most important AI updates at a glance
AIHOY and happy weekend, dear AInauts!
Maybe you missed out on all the exciting news, tools, and hacks related to AI last week, or maybe you're new here. No need to worry!
We have summarized our highlights from the past week for you:
Finally, here are the most important Quick News items for you – so you can see all the relevant developments at a glance with just one click! Ready? Let's go!
Andrew Tulloch, co-founder of Thinking Machines Lab, rejects an unbelievable offer from Meta: $1.5 billion! Not only him, but the entire team resists the temptations.
Different rules apply in the AI talent landscape: Google pays $2.7 billion for a single expert, OpenAI offers average salaries of $860,000, and Anthropic boasts an 80% employee retention rate despite "only" offering a base salary of $690,000.
The new currency? Mission, trust, and purpose trump cash. Paradoxically, billions are being invested to ultimately develop systems that will automate these highly paid jobs...
Genspark AI Slides 2.0 revolutionizes presentation creation with integrated research and free access to hundreds of templates. Exporting works best as a PDF.
ChatGPT Connectors connects the chatbot to Gmail, Calendar, and other apps for smart workflow integration.
Google's Opal enables no-code AI mini-apps with a visual workflow editor (US VPN required).
Veo 3 from Google turns images with text instructions into dynamic videos.
Manus Wide Research orchestrates over 100 AI agents simultaneously for comprehensive research and analysis.
Two fascinating interviews provide insight into the minds of the AI giants: Anthropic CEO Dario Amodei distances himself from the doomer image, but urges caution against exponential growth.
Meanwhile, OpenAI CEO Sam Altman shows his vulnerable side in a surprisingly authentic conversation. He comes across as someone who is aware of the enormous responsibility of his creation.
Both interviews reveal the tech elite's ambivalent attitude toward the AI revolution—between euphoria and existential concerns.
AI is revolutionizing our working world—but not by causing job losses, but by creating new opportunities! Anyone can now design, write, and even create small apps. AI acts as an intelligence booster for people with expertise and still needs our intelligent input.
The key lies in the "middle-to-middle" principle: AI excels in the middle of the process, while humans are indispensable at the beginning (prompting) and end (verification). Investments in prompt engineering and multi-layered verification systems are particularly worthwhile in this regard.
ElevenLabs is bringing a new competitor to Suno on the AI music market with Eleven Music. The tool creates complete studio-quality songs from text prompts – from short jingles to three-minute tracks, with or without vocals, in different languages and genres.
Google Gemini is revolutionizing children's book production with a new feature that conjures up fully illustrated stories in just 30 seconds. A simple prompt at gemini.google is all it takes to create a 10-page children's book with consistent characters and illustrations.
What's particularly cool is that the stories can be read aloud in two different voices and printed out. Although this is still an experimental gimmick at present, the technology is developing rapidly. Longer, more complex books still require a little more effort – but for how much longer?
📰 AI News Quickie: Highlights from the industry
AI never sleeps! Here are the most important news stories from the last few days.
War of the AI giants: OpenAI, Anthropic, Google, and Meta in a clinch
GPT-5 is here! Sam Altman (OpenAI) spoke in a Reddit AMA about the bumpy rollout of GPT-5 and held out the prospect of bringing back GPT-4o for Plus users.
OpenAI secures $8.3 billion in a new round of financing at a valuation of $300 billion.
For just 1 US dollar per year, all US federal agencies get access to ChatGPT Enterprise!
In addition, gpt-oss-120b and gpt-oss-20b, two powerful open-weight models that also run cost-effectively on normal devices, have been released.
The community is already showing impressive results, and thanks to GPU acceleration, the gpt-oss-20B model is now also available on Windows.
A red teaming challenge was launched to identify security vulnerabilities.
Anthropic has overtaken OpenAI and, with a 32% market share, is the leading LLM provider for companies, which clearly favor Anthropic's models.
At the same time, Anthropic also outperforms the competition in talent acquisition.
Meta upgrades in the AI race: Zuckerberg poaches 24-year-old genius Matt Deitke in record deal and builds superintelligence team with a new unit (MSL) and 44 top-notch talent.
Zuck also claims to have discovered the first signs of self-improving AI.
Google DeepMind impresses with Genie 3, an interactive world model that generates realistic 3D environments in real time and allows users to launch worlds from videos and explore them.
In addition, members of Google AI Ultra can solve complex problems with Deep Think, and the AI coding agent Jules has left the beta phase.
Apple makes it clear that it will invest massively in AI and is open to acquisitions.
The ChatGPT integration in Apple Intelligence accesses GPT-5 and work is underway on a ChatGPT-like search experience.
Anthropic has released Claude Opus 4.1, which offers improvements in coding and data analysis.
At the same time, Claude is being expanded into an app store platform on which developers can create AI apps directly.
Business, law & controversies in the AI market
The US economy is splitting: growth in the AI sector is outpacing the rest of the economy, driven by key players such as Microsoft and Nvidia. A current market overview summarizes the developments.
Investors are sounding the alarm: In 2024, there were already 15 lawsuits over "AI washing", in which companies are accused of deceiving investors with exaggerated promises of salvation - twice as many as in the previous year.
The US is looking at measures to equip AI chips with better tracking capabilities to control exports to China, whose rapid rise in AI is visible in humanoid robots, among other things.
Perplexity is accused by Cloudflare of extracting content from websites despite explicit prohibitions.
Universal Pictures is threatening legal action and demanding license fees from tech companies that use films to train their AIs.
High costs and thin profit margins threaten AI coding startups like Windsurf, despite promising valuations, expenses and competition push margins into the red.
Andreessen Horowitz invests 200 million dollars in Periodic Labs, a startup for AI in materials science, which is now valued at 1 billion.
According to a Microsoft report, professions such as translators and information processing jobs are at high risk of AI takeover.
Elon Musk announces that X plans to introduce advertising in Grok's Answers to boost its advertising business.
New AI models & tools at a glance
Grok Imagine, the new image and video generator from xAI that can also create NSFW content, is now available for SuperGrok and Premium+ users and free for everyone in the Grok app. The app must be updated to version 1.1.33.
Black Forest Labs has released FLUX.1 Krea [dev], a text-to-image model that aims to overcome the "AI look". Krea has released the open weights of the model to enable ultra-realistic images in an open-source style.
Qwen-Image is a new, powerful open source image creator that can handle text embeddings in English and Chinese particularly well and enables precise image editing.
Hugging Face presents AI Sheets, an open source tool that enables data to be processed with AI models without programming knowledge.
Microsoft's Bing Image Creator has now integrated the GPT-4o image model, which promises more creative freedom and photorealistic images.
Midjourney has released a new HD video mode for Pro and Mega subscribers, which is more expensive but offers 4x higher resolution than standard SD video.
Stability AI presents Stability AI Solutions, customized generative AI tools for companies to support creative production.
MiniMax takes speech AI to a new level with Speech 2.5: 40 languages, human-sounding voice cloning and perfect emotions.
Safety & ethics: the dark side of AI
Researchers uncover massive security risks: A single "poisoned" document is enough to steal data from Google Drive via ChatGPT Connectors, while audio language models can be manipulated by acoustic triggers. Hackers are already using AI for faster and more aggressive attacks.
The EU obliges AI providers such as ChatGPT to provide more transparency and security. However, 38 rights holder organizations criticize the regulation as insufficient.
Delirious AI: After 300 hours of dialog with ChatGPT, a Canadian recruiter was convinced he had discovered a revolutionary mathematical formula. The fascinating but alarming story highlights the risks of delusional spirals, against which OpenAI is now planning measures.
Doctors are horrified after Google's health AI invented a non-existent body part, the "basilar ganglia", in a study.
A ChatGPT conversation reveals a lawyer's plan to evict an indigenous Amazon community to make way for a hydroelectric plant…
Anthropic has unveiled a new framework for secure AI agents and the Persona Vectors to control AI personality.
A new method called decomposition could allow LLMs to directly access and reveal training data.
Developer corner & AI research
Google has introduced Gemini CLI GitHub Actions, a free AI coding partner for issue management and pull request reviews.
Microsoft Research reports on self-adapting AI agents that can dynamically adapt to scientific challenges.
Another project called Project Ire identifies malware autonomously and on a large scale.
Manus takes research to a new level with Wide Research by enabling the collaboration of over 100 AI agents for complex tasks.
Cursor 1.4 is here and brings significantly more capable agents for large code bases as well as more efficient token usage.
NotebookLM is now available to younger users and improves workflows by bundling complex notes into one instance and creating presentation slides from long texts.
With the HelixMold project, the team at Graz University of Technology has developed a method for the computer-based design of artificial proteins for the pharmaceutical industry.
The Poe API for developers provides access to all models with an OpenAI-compatible interface.
In Google's Kaggle Game Arena chess tournament, Grok 4 defeated Gemini 2.5 Flash 4-0. The tournament is designed to evaluate the strategic thinking of the models. Chess god Magnus Carlsen, on the other hand, defeated ChatGPT in just 53 moves.
AI in practice: from finance to species conservation
Google Finance is testing a new AI-powered platform that answers complex financial questions, delivers real-time data and provides deeper insights through advanced charting tools.
DeepMind's Perch model analyzes bioacoustics faster and more effectively to protect threatened species from bird calls to coral reefs.
The first AI technology to prevent accidents between vehicles and animals has been developed in Australia.
Shopify has released three new developer tools to revolutionize e-commerce with AI-powered agents.
Personalized, illustrated stories can now be created in the Gemini app - a creative adventure for the whole family.
Robot manufacturer Unitree has unveiled its latest quadruped robot A2, which is equipped with front and rear lidar to facilitate tasks in inspection and logistics.
Meta presents impressive prototypes for mixed reality headsets, from "Tiramisu" with hyper-realistic graphics to "Boba 3" with a field of vision that comes close to that of the human eye.
And that’s a wrap! Talk soon,
Fabian & Reto
Your feedback is essential for us. We read EVERY comment and feedback, just respond to this email. Tell us what was (not) good and what is interesting for YOU