Hello everyone, welcome to TechCrunch's regular AI newsletter. If you'd like to have this sent to your inbox every Wednesday, sign up here.
Longtime readers of this newsletter may have noticed that we skipped a week last week. That was not our intention and we apologize.
The reason is that we have reached a tipping point in the AI news cycle. We are a small team and underpowered. It's becoming nearly impossible to cover all the announcements, controversies, academic papers, trends, open model releases, lawsuits, etc.
Let's take this month as an example. OpenAI is in the midst of what is effectively a 12-day press engagement. Like Elon Musk's company xAI, Google is gearing up to launch a major new AI product. And that's just the news from the biggest AI player.
We're making some small changes to the AI this week to better deal with deluges. In the future, the number of characters in the newsletter will be slightly shorter. It's not a huge reduction, but you might not notice it, but the idea is to make this week's AI more concise and get it to your inbox on a regular basis.
We hope you find the improved newsletter easier to understand. As always, we welcome your feedback. Please feel free to share your feedback with us at any time.
news
Image credit: Kirillm / Getty Images
Testing AGI: A famous test for artificial general intelligence (AGI) is close to being solved, but its creators say it's not a genuine breakthrough in research, but rather a flaw in the test's design. It says that it shows that.
Amazon's new lab: Amazon announced the creation of a new research and development lab in San Francisco, the Amazon AGI SF Lab, to focus on building “fundamental” capabilities for AI agents.
OpenAI's video generator launches: Most subscribers to OpenAI's ChatGPT Pro and Plus plans now have access to Sora, OpenAI's video generator, starting Monday. But Europeans were out of luck.
China investigates Nvidia: China's market regulator has reportedly launched an antitrust investigation into Nvidia's acquisition of Mellanox, an Israel-based company working to develop high-performance chips for supercomputers.
Yelp adds AI: Yelp released several new features this week, including AI-powered review insights. The platform's AI analyzes the sentiment of reviews and tries to highlight them by category (such as food quality).
Google's renewable energy expansion: Google has signed a deal to generate enough carbon-free electricity to power several gigawatt data centers. Investments in renewable power total approximately $20 billion.
Reddit Debuts Conversational AI: Reddit's newest AI-powered feature, Reddit Answers, lets users ask questions and receive a curated summary of relevant answers and threads from across the platform. .
New image generator for X: X (formerly Twitter) has a new image generator courtesy of xAI, Elon Musk's AI startup. It's called Aurora and is tuned for “photorealistic rendering.” Found in X's Grok Assistant.
This week's research paper
Image credit: Bryce Durbin / TechCrunch
Ai2 and a team of computer scientists from the University of California, San Diego announced they have created an AI model that can predict climate patterns for 100 years in 25 hours.
The model, called Spherical Diffusion, starts with basic climate science knowledge and then applies a series of transformations to predict future patterns. Unlike many state-of-the-art climate prediction models, Spherical Dyffusion can be run on relatively small-scale hardware, the researchers claim.
The model has limitations. However, researchers plan to continue making improvements. The next version will simulate how the atmosphere reacts to carbon dioxide.
Separately, Ai2 released the second generation of climate modeling AI, Climate Emulator.
this week's model
Image credit: Yang et al.
Sora may be getting all the attention, but a new video generation model developed by MIT CSAIL and Adobe Research could be even more exciting.
This model, called CausVid, can start playing a video the moment you start generating it, providing a kind of preview of the finished clip. This is in contrast to models like Sora, which cannot display clips in progress.
The researchers plan to release an open source implementation soon.
grab bag
The group of artists whose access to Sora was leaked last November has published a series of essays explaining why.
The essay is worth reading, but the point is that the group wanted to denounce what it saw as the exploitation of creatives for research and development and public relations.
“We called on artists to think beyond their own systems,” the group wrote in the post, “and the limits of encouraging models through big technology.”