Meta gets a new AI model

0:00This is TechCrunch.

0:30The way Scribe works is simple. It captures work as it happens, then automatically creates clear, step-by-step guides, ready for your team to review and then share in seconds. From there, anyone can launch a live walkthrough right inside the actual tool to clearly show how work gets done. So there's no confusion. Everyone can be on the same page, and your team can skip creating trainings from scratch. You'll also be happy to know that Scribe is built with privacy in mind. It can automatically redact sensitive information like account numbers, names, or emails. Your admin picks the apps, and anything outside them is never captured.

0:59Try it with your team at scribe.how slash tzdaily and mention the show for your first month of Scribe Capture free on select plans. That's S-C-R-I-B-E dot how slash tzdaily. This episode is brought to you by Indeed. You're ready to level up your business, but the pressure to find the right person to make that happen is intense. Indeed Sponsored Jobs can help carry the load. It boosts your job post to help it reach quality talent fast. Cut through the chaos with Indeed Sponsored Jobs. Get a$75 sponsored job credit at Indeed.com slash podcast. Terms and conditions apply. Hiring now? Then this is a job for Indeed Sponsored Jobs.

1:39Meta gets a new AI model. I'm Emron Sheik and your Daily Crunch featuring three big tech headlines starts right now. As part of a planned data center in Pecos County, Texas, Amazon's investing in an on-site power plant that could become the largest source of climate pollution in the United States, according to the New York Times. You see, the New York Times says the plant would burn natural gas and is permitted to release 33 million tons of carbon dioxide per year, more than any other power plant in the U.S. In a statement, an Amazon spokesperson confirmed that the data center will be powered by new on-site generation that won't raise electricity costs for Texas families.

2:16Data centers face growing political opposition for a number of reasons, including their effect on electricity costs. AI has already had a significant impact on Amazon's carbon emissions, which it reported were up 16 % last year, the wrong direction for a company that pledged to eliminate its carbon emissions by 2040. And that could get worse as Amazon and tech companies back the development of huge natural gas plants to support their power-hungry data centers. The Amazon spokesperson said, The world looks different now than when we co-founded the climate pledge, while also claiming, Our commitment hasn't changed.

2:52Well, programming with CloudCode will soon require even less human oversight, as Anthropic says it's making AutoMode the default for Pro, Max, and Team accounts starting on August 14th. The company unveiled a test version of AutoMode in March, pitching it as a way to balance speed and control. As Anthropic explained in its announcement on Friday, when CloudCode is in AutoMode, instead of presenting prompts asking for human approval at each step, it'll proceed unless an action is determined to be irreversible, destructive, or aimed outside your environment. Anthropic also said that in testing, auto mode proved safer than manual review.

3:27In a study with 1 ,053 paid testers, auto mode caught 89 % of harmful actions while human review only caught 13.6%. Perhaps that's because manual review can become habitual. Users approved 97 % of permission prompts in CloudCode. In a post on X, CloudCode head Boris Cherney said, The team and I use Auto Mode exclusively and have been for many months. I couldn't imagine going back to permission prompts. The company also said it's been adding new safety features like prompt injection screening and customizable hard deny rules to prevent things like data exfiltration. On Monday, Meta released Muse Glimmer, an open-weight model designed to power AI agents locally on consumer hardware, providing the clearest picture yet of what CEO Mark Zuckerberg's vision of personal superintelligence could look like in practice.

4:22You see, the 30 billion parameter model is essentially an open version of Meta's most powerful closed model, Muse Spark, which the company debuted in April. Glimmer's weights are available under the permissive Apache 2.0 license, so developers can download them and modify as necessary. Now, Glimmer is designed to run AI agents that can perform multi-step tasks like call tools, write and debug code, work with files and screenshots, and execute on a task over an extended workflow, locally on a Mac or PC with a single consumer GPU. It supports text and images and was trained across more than 100 languages, the company said.

4:59Meta imagines Glimmer being used for things like managing schedules, drafting messages, and organizing files, tasks that would require large amounts of access to personal data. By processing the information on a user's device instead of sending it to the cloud, Meta is laying the groundwork for a more privacy-sensitive personal agent. Glimmer is also designed to be always-on and able to operate anywhere, anytime, with or without an internet connection. That vision mirrors the future Mark Zuckerberg has previously laid out for Meta. Last year, he argued that advanced AI should empower individuals rather than stay concentrated in the hands of a few companies, while also warning that Meta would have to be careful about which of its increasingly powerful models it released openly due to safety concerns.

5:44In a new letter on Monday, Zuckerberg reiterated that vision, arguing that distributing superintelligence widely has the potential to begin a new era of personal empowerment where individuals can use this powerful new capability to reach their full potential, pursue their interests, and improve their lives and the world more than ever before. Now to the latest in startup business news with our friend, producer Dennis. Thanks, Imran. And NextSlide recently announced that it's joining OpenAI with the presentation startup team members now working on ChatGPT. The Next Slide website currently displays a note from its founder describing the startup's product as one that could turn prompts, notes, documents, or research into a polished, editable presentation.

6:33The ultimate goal, according to the founder, was to make visual communication more accessible and help more people express their ideas clearly. So by joining OpenAI, the team will continue pursuing the same mission, building AI products that help people create, communicate, and turn their ideas into meaningful work. The financial terms of the deal were not disclosed. And chips running AI workloads are too hot. That's the one reason why data centers consume so much electricity and require cooling systems. And inevitably, entrepreneurs are turning to AI to solve the problem it created, Discovered Materials is the latest, with plans to use swarms of AI agents to find new materials that can be used to build more efficient integrated circuits.

7:19The startup said recently it closed a$9 million seed round. And folks, that's your Daily Crunch. Today's stories were reported by Anthony Ha, Rebecca Belon, and more awesome TechCrunch journalists. We'll see you here tomorrow, same Tech Time, same Crunch channel, and until then, find us at TechCrunch.com.

7:41Brussels clean up nicely at Sweetgreen. Maple glazed, roasted, and edges perfectly caramelized. Sweetgreen's fall harvest is back on the menu, and the season's most overlooked little green vegetable is dressed to be devoured. You know what to do. Order on the Sweetgreen app.