• Weekly AI News
  • Posts
  • OpenAI gpt-oss 🔓, OpenAI + AWS đŸ€, AI design agent đŸ§©

OpenAI gpt-oss 🔓, OpenAI + AWS đŸ€, AI design agent đŸ§©

Gpt-oss-120b and gpt-oss-20b push the frontier of open-weight reasoning models. Both are fully open-weight, under Apache 2.0, and free to use commercially...

An AI scheduling assistant that lives up to the hype.

Skej is an AI scheduling assistant that works just like a human. You can CC Skej on any email, and watch it book all your meetings. Skej handles scheduling, rescheduling, and event reminders. Imagine life with a 24/7 assistant who responds so naturally, you’ll forget it’s AI.

AI Jobs: Between Gloom and Opportunity

Former Google X exec Mo Gawdat made headlines by calling the idea that AI will create new jobs “100% crap,” warning that by 2027 automation will displace even high‑level roles like CEOs, collapsing the middle class unless governments intervene with universal basic income. His warning resonates with real data: Goldman Sachs reports a 3‑point rise in unemployment among U.S. tech workers aged 20–30, while Revelio Labs found a 31% drop in hiring for high‑AI‑exposure roles like IT and data engineering since 2022. These shifts hint at a labor market under real pressure.

But recent August 2025 research suggests a more nuanced trajectory. A Time report highlights that AI could ultimately generate more jobs than it displaces, with growth in areas like AI oversight, orchestration, and ethics‑driven roles. ServiceNow projects over 10 million roles in India will be reshaped by AI by 2030, not eliminated, signaling opportunities for those who reskill. AI‑related jobs now offer 12–20% higher salaries and richer benefits, while demand for critical thinking, ethics, and cross‑disciplinary problem‑solving is surging.

So what? For workers, this is a call to upskill for AI‑complementary roles, think orchestration, ethics, and creative problem‑solving. For policymakers, it’s time to scale reskilling programs and safety nets. For businesses, the winning strategy blends automation with human‑centric oversight, using AI to amplify, not replace, talent. The disruption is real, but so is the opportunity for those who adapt.

💬 OpenAI Developments

OpenAI releases a free GPT model that can run on your laptop

OpenAI has released GPT‑OSS, its first fully open‑source large language model, aiming to foster transparency, community contributions, and safer AI development.
The model includes weights, training data details, and evaluation benchmarks, marking a significant shift in OpenAI’s strategy toward openness as it faces growing competition from open‑source AI ecosystems.

For the first time, OpenAI models are available on AWS

For the first time, OpenAI has partnered with Amazon Web Services (AWS) to offer its models via Bedrock and SageMaker, marking a major competitive shift in the AI cloud landscape. This move brings OpenAI’s open-weight reasoning models—previously only downloadable—to AWS with full approval, allowing enterprise customers to integrate them directly. It signals a strategic play by OpenAI to reduce reliance on Microsoft, with whom relations are reportedly strained, while giving AWS a long-awaited win in generative AI. The partnership also positions OpenAI to undercut Meta, as it continues releasing open-source models while Meta retreats from open-sourcing its future AI.

OpenAI tells ChatGPT users to take a break — Altman asks "What have we done?" just days before GPT-5 launch

OpenAI has begun nudging ChatGPT users toward healthier usage by displaying prompts like “You’ve been chatting for a while, is this a good time for a break?” during extended sessions, as part of new mental-health–informed guardrails. Meanwhile, just days before the expected mid‑August 2025 launch of GPT‑5, CEO Sam Altman publicly expressed concern, asking “What have we done?” and comparing the model’s rapid development to the pace of the Manhattan Project.

🚀 Tech Industry Moves

Google DeepMind has unveiled Genie 3, a real‑time, general‑purpose “world model” that generates several minutes of interactive 3D environments from simple text prompts—complete with remembered scene details, dynamic events (like weather changes), and physical consistency at 720p and 24 fps.

Anthropic has released Claude Opus 4.1, an upgraded version of its flagship model that enhances coding capabilities, agentic reasoning, and multi-step task execution while maintaining the same pricing as Opus 4. With a 74.5 % SWE‑bench Verified score—two points higher than its predecessor—Opus 4.1 is available across Claude, Claude Code, API, Amazon Bedrock, and Google Cloud Vertex AI, making it a drop‑in upgrade for developers and enterprises. Positioned strategically ahead of OpenAI’s anticipated GPT‑5 launch, the model sharpens Anthropic’s competitive edge by improving multi‑file code refactoring, deep data analysis, and agent‑driven search, signaling the company’s commitment to continuous, large‑scale model advancements.

xAI: Elon Musk announced that his startup xAI will fully open source its Grok 2 chatbot next week, making the model’s code, architecture, and weights publicly available to developers and researchers . This move underscores xAI’s sharper commitment to transparency and open‑source AI compared to OpenAI’s recent release of "open‑weight" models that omit full code or training details.

Alibaba’s newly released Qwen‑Image is a 20B parameter open‑source image generation model designed for exceptional accuracy in rendering embedded text—especially English and Chinese—handling complex layouts, multi-line text, and bilingual content with state-of-the-art fidelity . Developed as part of the Qwen series, it leverages a progressive, multi-task training framework combining Qwen2.5‑VL, VAE encoding, and diffusion-based editing to deliver outstanding general image generation, layout consistency, and editing capabilities across benchmarks.

💡 Emerging AI Technologies and Companies

Lovart is building ‘AI design agent’ that augments creative teams with single platform

Lovart, founded by ex-ByteDance exec Melvin Chen, is a newly launched AI design agent that automates the entire creative workflow—from brand identity and UI design to video content and packaging—using a single prompt. Unlike traditional design tools, Lovart leverages a multi-agent system and a contextual “infinite canvas” to coordinate creative tasks while maintaining strategic coherence. Its MCoT (Mind Chain of Thought) engine mimics how creative directors think, aligning brand strategy with design execution. After going viral, Lovart now boasts over 800,000 beta users and is positioning itself as an AI-powered alternative to traditional design agencies.

China closes gap in AI model development

China is rapidly narrowing the AI development gap with the U.S.: experts and U.S. officials now estimate China trails by just “three to six months” in model performance, driven by breakthroughs in domestic models like DeepSeek and increasing benchmark parity . Alongside aggressive state-led investments, chip-self-sufficiency projects like Huawei's “Project Spare Tire,” and widespread integration of AI across sectors, China is building its own competitive AI ecosystem that increasingly rivals American leadership.

AI Action Brief

With GPT‑5, Genie 3, Claude 4.1, GPT‑OSS, Grok 2, and Qwen‑Image arriving, the AI race is heating up fast. Action steps: start testing open‑source models (GPT‑OSS, Grok 2) for cheaper, customizable AI; upskill in multi‑modal and agentic AI to harness tools like Genie 3 and Claude; and track China’s rapid AI progress to anticipate competitive shifts. The next six months will redefine what’s possible—prototype, learn, and adapt now.

💡 AI Tool Deep Dive

Quen image

If you’re paying $$$ for Canva templates or waiting days for a designer to tweak a poster, you’ll love this: Alibaba just dropped Qwen‑Image, a free, open‑source AI image generator that can create bilingual, text‑perfect visuals (English & Chinese) and even edit text inside images—think posters, e‑commerce banners, or pitch decks in minutes.

Quick Tutorial: Try It Today

  • Run it on Hugging Face (no setup needed) or deploy locally via GitHub.

  • Use prompts like: “Create a modern poster for a coffee shop launch with English and Chinese text: ‘Grand Opening – 50% Off This Week’”.

  • Edit on the fly: change that text to “Buy 1 Get 1 Free” without re‑rendering the whole design.

Real‑World Examples

Social media campaigns for bilingual audiences.
Editable storefront signs or product listings.
Infographic slides for workshops or courses.

Monetize It

Sell custom poster/social media kits on Etsy or Fiverr.
Offer fast-turnaround design services for small businesses.
Use it to build your own branded templates for content marketing.

Why it matters: Qwen‑Image gives you pro‑grade, editable designs at zero licensing cost, perfect for lean solopreneurs who need speed + control.

If you want more updates related to AI, subscribe to our Newsletter

đŸ‘„ Connect & Feedback!

👉 Join Us:

📧 Advertise In Weekly AI News:

📧 Contact directly at [email protected]

😍 Share with your friends!


Reply

or to participate.