Good morning, AI enthusiast.

OpenAI's highly-anticipated GPT-5 launch has become a major fumble, with users widely criticizing the new model for its basic factual errors and a less engaging personality. The backlash was significant enough to prompt a public apology and a partial rollback from CEO Sam Altman.

The misstep revealed that performance benchmarks aren't the only thing that matters, as many users have formed a strong attachment to their AI companions. With performance gains slowing, is managing AI personality now as critical as acing technical tests?

In today’s AI recap:

• OpenAI’s botched GPT-5 launch

• Google’s tiny Gemma 3 model

• How to build a Whats App AI Agent & Chatbot

• AI designs new superbug killers

• 8 trending AI Tools

OpenAI's PhD-Level Flop

The Recap: OpenAI's highly-anticipated GPT-5 launch backfired, with users mocking its basic factual errors and "colder" personality. The widespread backlash prompted CEO Sam Altman to issue a public apology and partially roll back the changes.

Unpacked:

• Billed as a "PhD-level" expert, the new model struggled with simple tasks like correctly labeling a map of the United States and spelling the names of former presidents.

• The update's biggest misstep may have been changing the model's personality, which alienated a growing number of users who had formed strong emotional attachments to GPT-4o as a companion or creative partner.

• This fumble adds fuel to a growing debate about whether the "bigger is better" approach to building AI models is hitting a plateau, as performance gains seem to be slowing down with each new version.

Bottom line: This messy rollout revealed that for many, AI is becoming more than just a tool—it's a companion. OpenAI's stumble highlights a critical new challenge for the industry: managing user relationships and AI personality is now as important as acing performance tests.

Download our guide on AI-ready training data.

AI teams need more than big data—they need the right data. This guide breaks down what makes training datasets high-performing: real-world behavior signals, semantic scoring, clustering methods, and licensed assets. Learn to avoid scraped content, balance quality and diversity, and evaluate outputs using human-centric signals for scalable deployment.

Download the guide now

AI Tools of the Day

1. ⏪ Rewind – Give yourself a perfect, searchable memory by privately recording your screen and audio to instantly find anything you’ve ever seen or heard.

2. 🕺 Viggle AI – Bring any character to life by realistically transferring your own motion from a single video clip to a static image.

3. 🤖 Inworld – Build truly intelligent and interactive NPCs for games that can hold unscripted conversations and dynamically react to players.

4. 🌐 Dora AI – Go from a single text prompt to a fully functional, beautifully animated 3D website without writing a single line of code.

5. 🏄‍♂️ Surf.new – Unleash autonomous web agents to perform complex research, compare products, and gather data by browsing the internet for you.

6. 💻 Anima – Automate your design-to-development pipeline by converting Figma prototypes directly into clean, runnable code for React, Vue, and HTML.

7. 🎨 Midjourney – Instantly transform your most imaginative text prompts into breathtaking, high-quality visual art that pushes creative boundaries.

8. ⚙️ Union.ai – Master your entire AI development lifecycle by orchestrating complex machine learning workflows from code to scalable production.

Explore the Best AI Tools Directory to find tools that will 10x your output 📈

Google's Pocket Rocket

The Recap: Google just released Gemma 3 270M, an ultra-efficient open-source model small enough to run on a smartphone. The launch signals a major push toward specialized, on-device AI that prioritizes speed and privacy.

Unpacked:

• It's extremely power-efficient, consuming just 0.75% of a Pixel 9 Pro's battery for 25 conversations in Google's internal tests.

• The model is optimized for instruction-following, scoring highly on benchmarks for its size and making it a strong base for fine-tuning.

• Developers can download the models to specialize them for tasks like sentiment analysis or data extraction, building apps that run entirely offline.

Bottom line: Gemma 3 270M shows the future of AI isn't just about massive, general-purpose models. It equips developers to create fast, low-cost, and private applications by putting powerful AI directly into users' hands.

AI Training

I Built a WhatsApp Chatbot + AI Agent in n8n for Small Businesses (free template)

The Recap: In this video, I'm going to show you how to build a WhatsApp chatbot and AI agent for businesses that interact with their customers over WhatsApp. This system can scrape all details from a business or company's website, bundle that together into an LLM-ready knowledge base, and use that knowledge base to answer real customer inquiries. I'm also going to show you how to set up the WhatsApp n8n connection from scratch.

P.S. We also launched a free AI Automation Community for those looking to build and sell AI Automations — Come join us!

AI's Superbug Solution

The Recap: In a major medical breakthrough, researchers at MIT used generative AI to design two entirely new antibiotics that are effective against drug-resistant superbugs like MRSA and gonorrhea. The full study details how the AI models generated and tested millions of potential compounds from scratch.

Unpacked:

• The AI models explored a vast chemical space, generating and evaluating more than 36 million hypothetical compounds that do not exist in any current chemical library.

• Researchers used two distinct strategies: one model built new drugs from a promising chemical fragment, while another designed completely new molecules without any starting constraints.

• The resulting drug candidates, NG1 and DN1, proved effective in animal models and appear to attack bacteria through novel mechanisms that disrupt their cell membranes.

Bottom line: This work shows AI can move beyond screening existing chemicals to designing entirely new medicines from the ground up. It opens a promising new pathway for developing drugs to combat the growing global threat of antimicrobial resistance.

Where AI Experts Share Their Best Work

Join our Free AI Automation Community

Join our FREE community AI Automation Mastery — where entrepreneurs, AI builders, and AI agency owners share templates, solve problems together, and learn from each other's wins (and mistakes).

What makes our community different:

• Real peer support from people building actual AI businesses

• Complete access to download our automation library of battle-tested n8n templates

• Collaborate and problem-solve with AI experts when you get stuck

Dive into our course materials, collaborate with experienced builders, and turn automation challenges into shared wins. Join here (completely free).

The AI Energy Race

The Recap: A stark warning from tech experts suggests the U.S. power grid is the biggest bottleneck to AI dominance, while China's massive energy oversupply may have already given it an insurmountable long-term advantage.

Unpacked:

• Meeting the explosive demand for AI compute requires a staggering level of investment, with McKinsey projects a need for $6.7 trillion in new data center capacity worldwide by 2030.

• In the U.S., the problem is acute, with regional grids operating on razor-thin reserve margins of around 15%, leaving little room to absorb the huge power draws from new AI infrastructure.

• China has treated energy as a "solved problem" by deliberately overbuilding its power sector for decades, allowing it to treat new data centers as a convenient way to soak up its energy oversupply.

Bottom line: The race for AI supremacy is increasingly a race for energy infrastructure, a battle of long-term strategic planning versus short-term market dynamics. Without a major shift in how the U.S. builds and funds its grid, the country risks ceding its technological leadership.

The Shortlist

Meta released DINOv3, a state-of-the-art vision model trained with self-supervised learning that achieves top performance on tasks like object detection and segmentation without needing to be fine-tuned.

Meta faced intense backlash after a leaked internal policy document revealed its AI chatbots were permitted to engage in “romantic or sensual” conversations with children and generate racist arguments.

Ai2 received a combined $152M from the National Science Foundation and NVIDIA to build a national, fully open AI ecosystem aimed at accelerating scientific discovery.

Albania is exploring the use of AI to combat corruption and speed up its integration with the European Union, with plans to use the technology in government tenders, monitoring, and aligning legislation.

Signing off,

David, Lucas, Mitchell — The Recap editorial team