Meta has unveiled a suite of new updates for the Facebook app, aiming to make the experience less cluttered and more relevant to users’ preferences. The changes focus on simplifying the interface, improving algorithms, and making posting and search tools more accessible.

The company stated that the goal of the updates is to make finding, creating, and sharing content easier, having observed that providing users with what they actually want encourages them to stay on the app.

These updates include a complete redesign of the post layout. Facebook will now arrange multiple photos in a unified grid, and content will be displayed in full screen when tapped. Users can also now like a photo by double-tapping, similar to what is available in other apps.

Meta also stated that search results will now appear within a comprehensive grid that supports all types of content. The company is also testing a new full-screen viewer that allows users to explore photos and videos without losing their position in search results, with plans to gradually expand this feature to include more content in the coming months.

The update also allows users to provide direct feedback on posts and short videos (Reels) to improve recommendations and make them more relevant to their interests. The company confirmed that more options for controlling the appearance of posts will be released soon.

Meta will place the most frequently used navigation tools, such as Reels, Friends, Marketplace, and Profile, in the middle of the tab bar for faster access, along with a redesigned menu and clearer, less cluttered notifications.

On the publishing side, the company says creating Stories and posts is now easier thanks to highlighting common tools like adding audio and mentioning friends. Advanced options—such as text backgrounds—will require an additional step. Meta has also redesigned how comments are displayed across feeds, groups, and Reels to make them more organized and easier to follow.

The updates also include suggestions for new friends when a user makes profile changes, based on shared interests. For example, if a user indicates an interest in crafts or travel, the app will suggest potential friends who share that interest. Meta emphasizes that users will still have control over what they post, both publicly and privately.

The company hopes these changes will improve how posts are displayed on Facebook, a platform that has faced widespread criticism in recent years for the abundance of unwanted content and posts that are irrelevant to users.

A radical internal transformation: Meta re-engineers its AI plans to remain competitive.
Mark Zuckerberg’s Meta is poised for the biggest shift in its AI strategy in its history, with the company beginning development on a new, closed-source model codenamed “Avocado,” expected to be unveiled next spring, according to Bloomberg, citing sources familiar with the company’s plans.

This shift marks a significant departure from Meta’s long-standing open-source approach since the dawn of the modern AI revolution, bringing it closer to models adopted by companies like Google and OpenAI, which are sold commercially rather than being fully made available to developers.

A radical shift after the failure of “Llama 4”: Bloomberg noted that the change of course began earlier this year following the lackluster performance of the Llama 4 model, which disappointed Zuckerberg. According to sources, Meta dismissed several members of the team responsible for the model, while Zuckerberg personally recruited prominent figures in AI research with lucrative offers reaching hundreds of millions of dollars for some, including Alexander Wang, who joined Meta as part of the acquisition of his startup, Scale AI, valued at approximately $14.3 billion.

Zuckerberg is now working directly with the new team within an internal group known as TBD Lab, where he spends considerable time overseeing daily development tasks, reflecting the project’s importance to the company.

Competing models such as Google’s Gemma, OpenAI’s gpt-oss, and Alibaba’s Qwen are undergoing a process of “distillation” within Meta to train the new model. Meta’s reliance on Chinese technology is a significant shift, especially after Zuckerberg’s statements earlier this year criticizing Chinese censorship of AI technologies.

Investment Pressure and Fears of “Overspending”

AI development has become Meta’s top priority, with Zuckerberg committing to spend $600 billion over three years on US infrastructure, mostly for AI projects. These plans include cutting investments in the Metaverse division to focus on developing smart glasses and AI-powered consumer devices.

This massive spending has worried investors, especially as the company continues to expand through 2026 without guarantees of short-term returns. Meta’s stock recently rose following reports that the company intends to reduce its Metaverse investments, signaling a market desire for a more disciplined spending policy.

Internal Exodus and a Declining Role for Open Source

Meta has experienced significant internal turmoil in recent months. Several new researchers left within weeks of joining, and the company eliminated 600 jobs in its AI units, including significant reductions in the FAIR research team. Yann LeCun, one of the world’s leading AI scientists, left after disagreements over a lack of support and a decline in the open-source approach.

Meanwhile, tools like Vibes—a platform Meta hastily launched to compete with OpenAI’s Sora 2 video generation model—received lukewarm reviews after being overshadowed by rival products.

The future of this transformation within Meta hinges heavily on Alexander Wang, who plays a central role in developing the Avocado model despite not being a purely technical researcher and being relatively young, at just 29 years old.

Those close to Wang believe his organizational skills and quick decision-making have earned him the trust of Zuckerberg, who oversees him directly. Within the team, some feel that Zuckerberg’s meticulous involvement sometimes leads to burnout and pressure, according to the report.

Organizational Challenges and Concerns Over “Superintelligence”
Meta is striving to achieve what it calls “superintelligence”—models that match or surpass human capabilities. However, the term is generating widespread concern, particularly in Europe, where scrutiny of the company’s products is intensifying.

Market research conducted by Meta with external experts has shown that the term raises concerns among policymakers about the potential for “uncontrolled power.” In the United States, some public figures are calling for a halt to the development of this level of artificial intelligence until its safety can be guaranteed.

Amid a fierce global race, enormous costs, and a profound shift in open-source strategy, Meta’s Avocado project is entering a critical phase. If Meta is to achieve a breakthrough next year, it must overcome internal, regulatory, and economic challenges while trying to convince the world that its biggest gamble ever is worth the cost and that it will not be left behind in the global AI race.