Hey Folks 😀

We took advantage of the last few days off to improve the content of this newsletter… or at least we hope so!

1. First resolution (yes, we’re already preparing for 2025): no more Sunday news—after all, it’s the day of rest! 😉 

2. Second resolution - easy - The Top AI news section will stay the same. So many of you find it useful, so that’s great!

3. Third resolution (and we’ll stop here): Explore one topic in depth each week from different angles.

   • Monday: Define the concept.

   • Tuesday: Share a video showcasing practical applications.

   • Wednesday: Introduce an “AI-Venger” working on the topic.

   • Thursday: Simplify and explain a scientific article.

   • Friday: Spotlight a startup offering a related solution.

We hope you’ll enjoy it! This week, we’re starting with ‘Meeting Productivity‘.

Top AI news

1. DeepSeek-V3: Open-Source Model Rivals Leading AI Systems
DeepSeek-V3, created by China's DeepSeek, is setting new standards in AI with its frontier-grade open-source language model. Featuring 671 billion parameters, with 37 billion activated per token, the model leverages a Mixture-of-Experts architecture for exceptional efficiency and performance. Trained on 14.8 trillion diverse, high-quality data tokens—including synthetic tokens crafted by other top-tier language models like GPT-4—DeepSeek-V3 significantly outperforms other models in the open-source arena, such as LLaMA 3.1 (405 billion parameters), demonstrating clear superiority in benchmarks and capabilities. It even rivals proprietary models like GPT-4. Remarkably, DeepSeek-V3 achieved this with a training budget of just $6 million, using 2048 GPUs over two months—an unprecedented feat in cost efficiency.

2. OpenAI Restructures as For-Profit Public Benefit Corporation
OpenAI, initially a non-profit AI research organization, intends to restructure as a for-profit Public Benefit Corporation (PBC) to attract substantial investment required for AI development. This change aims to balance shareholder interests, stakeholder concerns, and public benefits in decision-making. The non-profit arm will focus on charitable initiatives, such as healthcare. This move follows internal upheavals, including the dismissal and reinstatement of CEO Sam Altman in 2023. Investors have pushed for a more conventional corporate structure, though legal challenges, including opposition from Elon Musk, may arise.

3. AI Renders CAPTCHA Obsolete, Challenging Online Human Verification
CAPTCHA tests, designed to differentiate humans from bots, are increasingly ineffective due to AI advancements. Modern AI can solve these challenges in milliseconds, outperforming human capabilities. This development complicates bot detection and raises concerns about online security. The rise of autonomous AI agents, capable of performing complex tasks without human intervention, further blurs the line between human and automated interactions, rendering traditional verification methods obsolete.

Bonus. 2024: AI Mishaps and Misadventures Captivate the World
The year 2024 witnessed several AI-related incidents underscoring the technology's unpredictability. A malfunction in ChatGPT led to incoherent responses, raising concerns about AI reliability. In Scotland, "Willy's Chocolate Experience" used AI-generated images for promotion, resulting in attendee disappointment due to unmet expectations. Additionally, search engine errors produced absurd recommendations, such as suggesting rock consumption for a balanced diet. These events emphasize the importance of cautious and ethical AI development to prevent unforeseen and potentially harmful outcomes.

Riddle of the Day

— I can read and write all day,
-— But I’m not human in any way.
—— I learn from data, that’s my might,
-—- Guess who I am, you’re probably right?

Who am I ?

Theme of the week

Meeting Productivity - The definition
Discover how AI enhances meeting productivity by automating scheduling, generating agendas, transcribing discussions, and managing follow-ups, transforming collaboration efficiency.

Stay Connected

Feel free to contact us with any feedback or suggestions—we’d love to hear from you !