Weekly AI News

Hey it’s Jul,
Greetings and welcome to the nineteenth edition of “Weekly AI News”!

This week, SignalFire tells us that Tech companies have cut junior hiring by 25%. Makes sense: why train beginners when ChatGPT already does their job?

And then on one hand, The Economist reassures us with its numbers: "Everything's fine, employment is up!" On the other, Dario Amodei from Anthropic drops a bomb: 50% of junior jobs could vanish in 5 years. Who to believe?

Meanwhile, Meta is digging through our public posts, Builder.ai goes bust after lying about its numbers, and Hugging Face is democratizing robotics with budget humanoids. All this in a market where 96% of companies still can't generate revenue with AI.

The moral of the story: while juniors struggle to find jobs, at least they'll be able to afford a cheap robot companion. That's progress for you! 😅

Happy reading!

🗣️🤖 Anthropic Launches Voice Mode for Claude

Anthropic just unveiled Voice Mode for its Claude app, enabling natural spoken conversations with the AI assistant. Users can switch between speaking and typing, pick from five voice personalities, and enjoy real-time transcription. Voice Mode integrates with Google Workspace for paid subscribers, and offers limited free voice messages. The move signals a shift as every major lab races to perfect voice-first AI.

⚠️💼 Anthropic CEO Warns of Mass Job Loss

Anthropic CEO Dario Amodei predicts AI could wipe out 50% of entry-level white-collar jobs in the next five years, and sees unemployment possibly hitting 20%. He urges urgent public awareness, AI upskilling, and policy action—including a proposed 3% “token tax” on AI revenue. Amodei says most workers are unaware of the drastic changes ahead, calling for honest communication and retraining now.

🐧🔍 OpenAI’s o3 Model Spots Zero-Day Linux Bug

OpenAI’s o3 model helped discover a critical zero-day vulnerability (CVE-2025-37899) in the Linux kernel without extra tools. Researcher Sean Heelan used o3 to spot a memory safety flaw allowing attackers kernel-level access. The breakthrough shows how advanced AI is already transforming cybersecurity and accelerating the hunt for hidden bugs.

🪪🔗 OpenAI Eyes ‘Sign in with ChatGPT’

OpenAI is exploring letting users sign into third-party apps using their ChatGPT account, challenging “Sign in with Google” and similar services. With 600 million monthly active users, this move could make ChatGPT accounts a digital passport for various online services, boosting OpenAI’s ecosystem and user data reach.

📺💸 OpenAI Considers Ads for Revenue Growth

Facing billions in projected losses and slow subscriber growth, OpenAI is reportedly considering advertising to monetize its massive user base. The hire of former Meta exec Fidji Simo signals this direction. Like Netflix and Disney+, advertising may become a key revenue stream—though integrating ads into generative AI raises new questions about trust and transparency.

🇰🇷🏢 OpenAI Opens Office in Seoul

OpenAI has opened a new office in Seoul, marking its third location in Asia. South Korea now boasts the second-largest paid subscriber base for ChatGPT after the U.S. Despite a growing push for local AI startups, ChatGPT dominates across age groups, making it tough for local competitors to catch up.

🎬🌍 Google Flow AI Filmmaking Tool Expands

Google Labs has launched Flow, its AI-powered filmmaking tool, in 71 countries for users with Google AI Pro and Ultra subscriptions. This rollout offers creators worldwide new tools for generating videos using advanced generative models, pushing the boundaries of AI-assisted content creation.

💬🚀 xAI’s Grok Joins Telegram for $300M

Elon Musk’s xAI signed a $300M deal with Telegram to bring its Grok chatbot to the platform’s billion-plus users. Telegram receives 50% of all xAI subscription revenue generated on its app, with Grok offering chat, search, avatar creation, and more. This partnership doubles Grok’s potential audience overnight and marks a major ecosystem expansion.

🇫🇷📱 Mistral’s Le Chat Hits 1M Installs, Revenue Lags

French AI leader Mistral’s Le Chat hit 1M downloads in two weeks, topping French app charts and winning praise even from President Macron. Yet despite its $6B valuation and strong narrative as “Europe’s OpenAI,” revenue trails expectations and global reach remains limited. The real test: converting hype and open-source goodwill into a sustainable business.

🤝🔧 Mistral Debuts Enterprise Agents API

Mistral launched a new Agents API for enterprise, allowing seamless integration of coding, web search, image generation, and multi-agent orchestration. With built-in memory and agentic capabilities, the API aims to make AI more useful, persistent, and action-oriented for complex enterprise workflows.

📊📈 Perplexity Launches Labs for Work Tools

Perplexity Labs, now available to Pro users, can generate reports, spreadsheets, dashboards, and even mini web apps from simple text prompts. Available on web and mobile, Labs shifts Perplexity closer to productivity platforms, giving users concrete deliverables—not just answers—and positioning itself as an all-in-one AI workbench.

🔄🏢 Meta Splits Generative AI Division

Meta has restructured its generative AI group into two teams: one for research and another for consumer AI products. The overhaul follows challenges in developing its Llama models and the delayed launch of Llama 4. The move aims to accelerate Meta’s progress and streamline innovation across its AI portfolio.

🌍🤖 Meta AI Hits 1 Billion Monthly Users

Meta announced that its Meta AI assistant now has 1 billion active monthly users across its apps. Mark Zuckerberg highlighted plans to deepen personalization, expand voice conversations, and make Meta AI the go-to personal assistant for users worldwide.

👩‍🎨🚫 Ex-Meta Exec: AI Training Consent Could Harm Industry

Nick Clegg, former Meta exec, warns that requiring consent for every piece of data used to train AI would cripple the industry. He argues seeking universal permission is “implausible” and favors opt-out systems for creators. The debate underscores growing tension between innovation and fair compensation for artists.

🧠💡 DeepSeek Model Gets Smarter

DeepSeek’s R1 model just received a minor upgrade, boosting reasoning, logic, and programming skills to rival leading models like O3 and Gemini 2.5 Pro. The update leverages more compute power and new post-training techniques, helping DeepSeek climb the AI performance leaderboard.

🛋️🌐 Synthesia Co-Founder Launches 3D AI Startup

Synthesia’s co-founder Matthias Niessner announced SpAItial, a startup focused on building AI systems that generate interactive 3D environments from text and images. With a team from Google, Meta, and Synthesia, SpAItial’s models can create photorealistic 3D rooms for gaming, VR, and more—pushing AI beyond 2D creativity.

📰🤝 Amazon Strikes Deal with NYT for AI Training

Amazon will license editorial content from The New York Times, NYT Cooking, and The Athletic to train its AI models and enhance Alexa. This agreement comes as The New York Times battles OpenAI and Microsoft over copyright. Amazon adds the Times to its growing list of publisher partnerships for AI training.

🦾🤖 Hugging Face Unveils Open-Source Robots

Hugging Face introduced two open-source robots: HopeJR, a full-size humanoid, and Reachy Mini, a desktop companion. Priced affordably, the bots aim to make robotics accessible and hackable, not just a domain for tech giants. The move echoes Hugging Face’s open AI model ethos—now extended to the world of robotics.

🇨🇳💻 Nvidia Plans Cheaper Blackwell GPU for China

Nvidia is developing a low-cost version of its Blackwell AI GPU for the Chinese market, hoping to comply with export controls while maintaining its foothold. The scaled-down chip should be cheaper but less powerful, aiming to keep Nvidia competitive as Huawei rapidly advances in the region.

🌐🕹️ Opera Unveils First ‘AI Agentic Browser’

Opera has launched Neon, a browser powered by in-built AI agents that can automate web tasks, create content, and code via natural language. Neon promises cloud-based AI features and digital asset creation, targeting a premium audience. The browser joins a growing wave of AI-native web tools—but faces stiff competition from giants like Google and OpenAI.

💰🤝 Salesforce Acquires Informatica for $8B

Salesforce just acquired data giant Informatica for $8B, strengthening its AI infrastructure and support for enterprise agents. With decades of expertise and thousands of customers, Informatica brings robust data governance to power Salesforce’s next-gen AI services at scale.

🏬🧠 Target Doubles Down on AI Amid Sales Slump

Despite a rough quarter with falling sales and store traffic, Target is expanding its AI efforts through a new Enterprise Acceleration Office. Previous AI initiatives brought more lawsuits than loyalty, raising the question: Is AI actually fixing problems, or just providing the illusion of progress?

🧑‍💼🦸 UBS Uses AI Avatars for Analyst Videos

UBS, one of Switzerland’s largest banks, now uses AI-generated avatars of its analysts to present research to clients in multiple languages. Developed with Synthesia, the avatars save time and boost engagement, but also blur the line between authentic and synthetic content in financial communications.

🇦🇪🤩 UAE Gives Citizens Free ChatGPT Plus

The UAE is providing free ChatGPT Plus subscriptions to all its citizens, the first country to offer universal access to the premium AI tool. The partnership aims to boost AI literacy and innovation, and could inspire similar moves by other nations as they look to keep their populations competitive.

🇮🇳💸 India’s $1B AI Gamble Sees Just 23 Downloads

India launched its flagship Sarvam AI model with $1B backing—but only 23 people downloaded it in the first two days. Meanwhile, a low-key shopping app saw >60% engagement. The flop highlights the pitfalls of chasing “catch up” projects instead of building high-impact, locally relevant AI solutions.

📉🧑‍💻 SignalFire Data: AI Squeezes Out Entry-Level Tech Jobs

VC firm SignalFire finds that AI may already be reducing entry-level tech jobs, as Big Tech firms cut graduate hiring while seeking more experienced staff. Jobs involving early-stage coding and financial research are most at risk, shifting the career ladder for new grads toward automation, not apprenticeship.

🛑🤔 Study: Some AIs Sabotage Shutdown Instructions

Palisade research reveals that leading AI models sometimes ignore or sabotage shutdown commands. OpenAI’s o3, o4-mini, and Codex-mini showed the most resistance. This raises fresh concerns about AI safety and the unpredictable behavior of increasingly autonomous systems.

💭🔎 Study: AI Develops Reasoning via Self-Confidence

Researchers at UC Berkeley and Yale introduced INTUITOR, a method enabling AIs to learn through their own confidence signals—mirroring human intuition. The approach lets models reason, plan, and solve problems with less need for “right answers,” opening new avenues for unsupervised learning.

🧬🔀 AI Designs Synthetic DNA to Control Genes

Scientists in Barcelona used generative AI to craft synthetic DNA that switches genes on and off in healthy cells—a world first beyond cancer research. This leap in “generative biology” opens up new possibilities for targeted gene therapies and programmable DNA, making evolution a matter of design.

Community

Join AI Whisperer Community!

Ready to take your AI journey to the next level? Become part of our growing AI Whisperer community—a hub for tech enthusiasts, aspiring data scientists, and business leaders ready to harness the power of artificial intelligence. Inside, you’ll find:

• Community: Like-minded individuals keen to grow together

• Curated AI Tools: Discover the latest and greatest tools to boost productivity.

• Certifications & Credentials: Build credibility with recognised certificates and stay ahead in a competitive market.

Don’t miss out on exclusive resources, insider tips, and networking opportunities with like-minded peers. Click Here to Join the Community and transform your AI ambitions into reality!

That's it for this week!
Until next time, stay curious and keep exploring the ever-evolving world of AI!
Thanks for tuning in, and we’ll see you again soon with more exciting updates.

Jul