AGI: Too Close

"This Week in AI" is now Superintelligence.

In partnership with

🚧🚧🚧

Dear readers,

Our newsletter is constantly evolving and we are delighted to announce some exciting new features: From April 7, you can expect varied reports on a selected Special Topic every day, which we hope will enrich your day and helps you to learn even more.

But that's not all: a few weeks later, we will also be completely redesigning our Saturday edition. Every Saturday, you can look forward to an inspiring DeepDive on a different topic, bringing you in-depth researched content and valuable perspectives straight to your inbox. We will offer you even more knowledge and exciting insights - stay tuned!

Until then, you can help us a lot by telling your friends about our newsletter and recommending it to them. And of course we are always happy to receive your feedback and suggestions, because your input helps us to become even better!

Thank you for your support - we look forward to taking the next steps with you!

AGI Is Closer Than We Think: Google And OpenAI Are Preparing!

The TLDR
Google DeepMind announces AGI could arrive within years, establishing a Safety Council led by Shane Legg. OpenAI introduces PaperBench where current models achieve only 21% of human performance. Major companies prepare for AGI's imminent arrival.

The future is coming faster than expected! Google DeepMind announced on April 2, 2025 that AGI - AI systems with human-like cognitive abilities - could become a reality “within the next few years”.

In its new safety concept, DeepMind addresses four key risk areas: Misuse, misalignment, accidents and structural risks. With the newly established AGI Safety Council, led by co-founder Shane Legg, and partnerships with organizations such as Apollo and Redwood Research, responsible development is being driven forward.

At the same time, OpenAI has introduced “PaperBench”, a milestone for measuring AGI capabilities. This benchmark tests whether AI agents can independently replicate research - even the most advanced models so far only achieve 21% of human performance.

However, one thing is very clear: AGI is getting closer and closer and all major AI companies are preparing for it. Not only users but society as a whole should feel safe - because the transformative power of AGI is unmeasured.

Question of the Day

Open Source or closed source: who will win the race for AGI?

Login or Subscribe to participate in polls.

You + AI = Unbeatable Presentations

Prezi AI turns your ideas into unique, dynamic presentations in seconds.

Just enter your prompt and you’ll get a stunning, fully-custom presentation—no templates or generic layouts. It’s just like having a professional designer by your side. Prezi AI even helps you refine your text, find the perfect image, and layout your content.

Get polished presentations that look like they took hours—done in seconds, every time.

Chart of the Day

Gemini 2.5 Pro reaches new heights in the USAMO 2025 Math-benchmark

In The News

Midjourney V7 Alpha Released

Midjourney releases V7 Alpha with personalization, improved quality, and new Draft Mode. The conversational interface enables rapid iteration at lower costs.

Genspark - The New Manus Competitor

Genspark introduces Super Agent, a reliable general AI using Mixture-of-Agents technology with 80+ toolsets. It makes calls, converts videos, and creates reports.

MoCha Transforms Text Into Speech

Meta launches MoCha, creating hyperrealistic talking characters from text or voice input. The technology's output appears indistinguishable from authentic human video.

Quote of the Day

Hi All,

Thank you for reading. We would be delighted if you shared the newsletter with your friends! We look forward to expanding the newsletter in the future with even more specialized topics. Until then, follow us on social media to stay up to date.

Cheers,
Dan