2 Sources
[1]
Meta Says It Has Scanned Every Single Public Instagram Post With its AI
During a Q2 earnings call this week, Meta CEO Mark Zuckerberg revealed that Instagram has improved its algorithm by scanning every public post and using that information to provide better recommended content. Zuckerberg says that the LLM AI model that scans public Reels and posts allows Meta to better analyze them for topics and tone, which he says is "a key building block toward greater personalization." The company is rolling out the process to Facebook as well. Zuckerberg lauded the AI technology that he says has contributed to users spending more time on the app by improving their feed and Reels recommendations. "Our recommendations are also becoming more personalized, surfacing more fresh content while giving people more direct control over what they see," Zuckerberg says. "This drove a 15-basis point increase in sessions on Instagram, with particular strength in reshares and time spent, which are both strong indicators of better content-to-user matching." But despite Zuckerberg's upbeat messaging on how Meta is deploying AI internally, Wall Street was down on the company as its share price slid by 8%. That is largely because of its earnings results that came in below expectations. Meta grew its revenue by 28% compared to a year ago to $61 billion. But profits fell 14% to $6 billion. The fall in profits is in large part because of voracious spending on AI infrastructure. The company estimates it will splurge $130 billion to $145 billion in 2026. Just three months ago, it said it would spend $125 billion. Investors are growing increasingly concerned that what Zuckerberg calls a "big bet across the industry" will not pay off as they fret about how AI firms will get a return on their huge investments. "My personal bet is that the people who invest in this will feel very good and be rewarded over time," Zuckerberg said this week. But an analyst tells the BBC that there are parallels to what's happening now and what happened a few years ago when Meta spent tens of billions of dollars on the metaverse, despite there being no proven demand for the product. "What it generated in cash this quarter almost all got eaten by AI infrastructure spending," says Mike Proulx, who works at Forrester Research. "Investors now have to decide whether Meta's growing list of AI initiatives represents company diversification or distraction." Image creditsHeader photo licensed via Depositphotos.
[2]
Mark Zuckerberg Says AI Will Create An 'Infinite' Instagram - Meta Platforms (NASDAQ:META)
Meta Wants AI To Become Instagram's Third Content Feed Today, Instagram primarily surfaces content from two places: people users follow and creators they may not. Zuckerberg said AI is about to add an entirely new category. "There are already two large sets of content to draw from -- first from your friends and the people you follow, and second from creators that you don't follow -- but now there is going to be a whole new and nearly infinite universe of personalized content," Zuckerberg said. He said Meta's newly launched Muse Image and Muse Video models will power that shift by generating highly personalized content tailored to individual users. "This is going to make our services a lot more useful and engaging for people," Zuckerberg added. AI Is Already Reshaping Instagram Meta's AI push is already producing measurable results across Instagram and Facebook. The company said large language models are improving recommendation systems by developing a deeper understanding of both content and user intent, allowing Meta to surface more relevant posts and videos. On Instagram, global time spent grew at a double-digit rate during the quarter, driven largely by improvements to Feed and Reels recommendations. Susan Li, Meta's chief financial officer, also revealed that every public Reels and Feed post on Instagram is now automatically processed through a large language model that analyzes everything from topics to tone before feeding those signals into recommendation and ranking systems. The company also rolled out its largest-ever Reels ranking upgrade, which drove a 15-basis-point increase in Instagram sessions. Meanwhile, more than half of all recommended content in Instagram Feed is now less than one day old -- more than double the level seen a year ago. Why It Matters For Investors Meta has spent years relying on creators and user-generated content to drive engagement. Zuckerberg's latest comments suggest the company now sees AI-generated content as another long-term growth engine. If AI can continuously generate personalized images, videos and other media that users find engaging, it could significantly expand the amount of content available on Instagram while keeping people on the platform longer -- creating more opportunities to serve advertisements. The strategy also extends beyond content recommendations. Zuckerberg said AI is already improving ad targeting, creative tools and business performance, reinforcing Meta's broader view that artificial intelligence will become a core driver of future engagement and monetization across its apps. Image via Shutterstock Market News and Data brought to you by Benzinga APIs To add Benzinga News as your preferred source on Google, click here.
Share
Copy Link
Mark Zuckerberg revealed that Meta AI now scans every public Instagram post using large language models to improve content recommendations. The company plans to create an 'infinite universe' of AI-generated content as a third feed category, but investors remain skeptical as Meta's $130-145 billion AI infrastructure spending pushes profits down 14% despite revenue gains.
During a Q2 earnings call this week, Mark Zuckerberg disclosed that Meta AI has scanned every single public post on Instagram using large language models to enhance its algorithm and deliver better content recommendations
1
. Susan Li, Meta's chief financial officer, confirmed that every public Reels and content feed post is now automatically processed through a large language model that analyzes everything from topics to tone before feeding those signals into recommendation and ranking systems2
. Zuckerberg described this as "a key building block toward greater personalization" that allows Meta to better understand content and user intent1
.
Source: Benzinga
The AI-driven improvements are delivering measurable results. Zuckerberg reported that the enhanced Reels ranking system drove a 15-basis-point increase in Instagram sessions, with particular strength in reshares and time spent—both strong indicators of better content-to-user matching
1
. Global time spent on Instagram grew at a double-digit rate during the quarter, driven largely by improvements to Feed and Reels recommendations2
. More than half of all recommended content in Instagram Feed is now less than one day old—more than double the level seen a year ago2
. This focus on fresh content and personalization and user engagement signals Meta's commitment to keeping users on the platform longer.
Source: PetaPixel
Mark Zuckerberg unveiled plans to introduce AI-generated content as an entirely new category on Instagram. "There are already two large sets of content to draw from—first from your friends and the people you follow, and second from creators that you don't follow—but now there is going to be a whole new and nearly infinite universe of personalized content," Zuckerberg explained
2
. The company's newly launched Muse Image and Muse Video models will power this shift by generating highly personalized images, videos, and other media tailored to individual users2
. If successful, this strategy could significantly expand available content while creating more opportunities to serve advertisements.Related Stories
Despite the positive engagement metrics, Meta's share price slid by 8% following the earnings announcement
1
. While Meta grew revenue by 28% compared to a year ago to $61 billion, profits fell 14% to $6 billion1
. The profit decline stems largely from voracious AI infrastructure spending. Meta estimates it will spend $130 billion to $145 billion in 2026—up from the $125 billion forecast just three months earlier1
. "What it generated in cash this quarter almost all got eaten by AI infrastructure spending," says Mike Proulx from Forrester Research. "Investors now have to decide whether Meta's growing list of AI initiatives represents company diversification or distraction"1
. Some analysts draw parallels to Meta's metaverse spending, where tens of billions were invested despite no proven demand1
. Zuckerberg defended the strategy, stating, "My personal bet is that the people who invest in this will feel very good and be rewarded over time" ."Summarized by
Navi
28 Sept 2024

Yesterday•Technology

30 Oct 2025•Entertainment and Society

1
Technology

2
Technology

3
Science and Research
