Dear Readers,
Google i/o starts in a few days and Google DeepMind, Google's research company, presented AlphaEvolve last week. This project has made some groundbreaking breakthroughs possible. Today we will first take a look at what AlphaEvolve is all about.
Next Saturday we will then have the first AI DeepDive, in which we will take a very in-depth look at one topic, this time AlphaEvolve.
Also: Robot Tuesday. The latest breakthroughs in AI and robotics.
Have fun reading!
In Today’s Issue:
Google’s AlphaEvolve rewrites the rules of algorithm discovery
New study questions AI explosion by 2027—compute is key bottleneck
Breakthrough genome-editing tool outperforms CRISPR
Gemma 3 dominates LMArena as top open model for single GPU
And more AI goodness…
All the best,
The TLDR
AlphaEvolve, a new AI agent from DeepMind, combines Gemini model creativity with evolutionary search to independently invent more efficient algorithms—outperforming benchmarks like the 56-year-old Strassen method for matrix multiplication. Beyond math, it optimizes real-world systems like chip design, data center efficiency, and LLM training, marking a paradigm shift in autonomous scientific discovery.
Imagine an AI that doesn't just write code, but independently invents new, more efficient algorithms - and breaks decades-old mathematical records in the process. This is exactly what DeepMind has achieved with AlphaEvolve.
AlphaEvolve is an AI agent that combines the creative capabilities of Google's Gemini language models with an evolutionary optimization process. It continuously generates, tests and improves algorithms - similar to natural selection in biology, but for code. The system has, for example, developed a new algorithm for matrix multiplication that is more efficient than the Strassen algorithm that has been established for 56 years.
The highlight: AlphaEvolve not only optimizes mathematical processes, but also practical applications such as the use of resources in data centers, the design of AI chips and the training processes of large language models - including its own architecture.
This is a milestone for the AI community. AlphaEvolve shows that AI can not only assist, but also independently generate new scientific findings. It is a step towards a future in which humans and machines work together to solve complex problems.
Why it matters: AlphaEvolve demonstrates the potential of AI to go beyond human expertise and independently develop innovative solutions. This could fundamentally change the way we do science and technology.
Stop burning budget on ads and hoping for clicks. Podcast listeners lean in, hang on every word, and buy from guests who deliver real value. But appearing on dozens of incredible podcasts overnight as a guest has been impossible to all but the most famous.
PodPitch.com is the NEW software that books you as a guest (over and over!) on the exact kind of podcasts you want to appear on – automatically.
⚡ Drop your LinkedIn URL into PodPitch.
🤖 Scans 4 Million Podcasts: PodPitch.com's engine crawls every active show to surface your perfect podcast matches in seconds.
🔄 Listens to them For You: PodPitch literally listens to podcasts for you to think about how to best get the host's attention for your targets.
📈 Writes Emails, Sends, And Follows Up Until Booked: PodPitch.com writes hyper-personalized pitches, sends them from your email address, and will keep following up until you're booked.
👉 Want to go on 7+ podcasts every month? Book a demo now and we'll show you what podcasts YOU can guest on ASAP:
A new study from researchers at the University of Chicago's X-Lab casts doubt on the likelihood of an AI-driven intelligence explosion by 2027. While rapid algorithmic advances are theoretically possible, the study argues that progress may be fundamentally limited by available compute resources. This suggests that breakthroughs in hardware, not just software, will be critical to reaching superintelligent AI.
New Genome-Editing Tool Surpasses CRISPR in Precision Gene InsertionA groundbreaking genome-editing technology has emerged that can accurately and efficiently insert entire genes into human DNA—something traditional CRISPR systems have struggled to accomplish. This advancement could unlock powerful new therapies for genetic disorders and transform the future of gene editing. | Gemma 3 Dominates LMArena as Top Single-GPU Open ModelThe Gemma 3 model family has once again topped the LMArena leaderboard, solidifying its status as the best open-source model that runs on a single GPU. With more updates on the horizon, Gemma continues to push the boundaries of efficient AI performance. |
China's technology hub in Beijing is increasingly focusing on collaboration between humans and humanoid robots, as demonstrated at the first robot half-marathon. According to a Beijing official, these robots will “increase human productivity and work in hazardous environments” without causing mass unemployment. This vision could reshape global economic policy by showing how advanced robotics can complement rather than replace human labor - a model with significant future potential for the integration of man and machine. The question remains as to how long robots will continue to support humans instead of replacing them.
Technology giant Huawei and robotics specialist UBTech have joined forces to bring humanoid robots from the lab to the real world. The agreement, signed in Shenzhen on May 12, includes the joint development of “intelligent factories with humanoid robots as well as bipedal and wheeled service robots for domestic use”. With the Chinese robotics market expected to reach 5.3 billion yuan by 2025, this cooperation strengthens China's position in the global race for practical, everyday robotic technologies - a decisive step towards the democratization of advanced automation.
Beijing is pumping billions into start-ups such as AgiBot and MagicLab to establish humanoid robots as fully-fledged factory workers. Thanks to domestic supply chains, the models only cost from 12,000 dollars, perform AI-supported assembly and quality control and are already appearing in government procurements. The government sees them as the answer to the shortage of skilled workers and ageing. If scaling is successful, China could redefine the industrial location worldwide and sell one million units by 2030.
How'd We Do?Please let us know what you think! Also feel free to just reply to this email with suggestions (we read everything you send us)! |