Curated by THEOUTPOST
On Thu, 31 Oct, 8:07 AM UTC
2 Sources
[1]
Meta says open sourcing Llama models will be a money-saver
It's not like Zuck needs the coin despite increased infrastructure spend, headcount, losses on VR Meta boss Mark Zuckerberg has told investors that open sourcing its Llama AI models is not entirely altruistic - he thinks it will also save his social media conglomerate money. Speaking on Meta's Q3 2024 earnings call, Zuck was asked how the social media conglomerate plans to build the massive infrastructure required to deploy AI across its products. Zuckerberg replied that making Llama available sees "a lot of researchers and independent developers ... work on Llama and they make improvements and then they publish it, and it becomes, it's very easy for us to then incorporate that both back into Llama and into our Meta products like Meta AI or AI Studio or Business AI." Meta has seen this before, he remarked, when it sparked the Open Compute Project. "Part of why we are leaning so much into open source here in the first place is that we found counterintuitively with Open Compute that by publishing and sharing the architectures and designs that we had for our compute, the industry standardized around it a bit more." "We got some suggestions also that helped us save costs, and that just ended up being really valuable for us," he added. Zuck hopes the same thing will happen with AI. "As Llama gets adopted more, you're seeing folks like Nvidia and AMD optimize their chips more to run Llama specifically well, which clearly benefits us," he told investors. "So it benefits everyone who's using Llama, but it makes our products better rather than if we were just on an island building a model that no one was kind of standardizing around in the industry." "I think it's good business for us to do this in an open way." Open source contributors may be thrilled to know their work helped Meta to quarterly revenue of $40.6 billion and net income of $$15.7 billion - numbers that rose 19 percent and 37 percent respectively year over year. They may also think again after CFO Susan Li revealed that Meta hopes Llama can one day be used "against a lot of our content moderation efforts to help make the big body of content moderation work that we undertake to help it make it more efficient and effective for us to do so." Meta mostly outsources moderation to contractors, who presumably will hear Li's remarks and start worrying. Meta's own headcount has rebounded since the layoffs of 2022 and 2023 to reach 72,404 as of September 30, 2024 - an increase of nine percent year-over-year. An average of 3.29 billion people used Meta's products each day of September - five percent growth. The web Kraken revealed plans for massive new infrastructure spending: $9.2 billion went out the door, "driven by investments in servers, datacenters, and network infrastructure." That number would have been higher but for late delivery of some servers that won't be counted until Q4. For the full year, Meta expects to spend between $38 billion and $40 billion - an increase from its previously predicted minimum spend of $37 billion. It didn't put a number on its 2025 spending plan, but told investors to expect "significant capital expenditure growth." Investors were also warned that Reality Labs - Meta's business unit dedicated to the metaverse and mixed reality hardware - made $270 million in the quarter but lost $4.4 billion. "We continue to expect 2024 operating losses to increase meaningfully year-over-year due to our ongoing product," explained CFO Susan Li. Meta has burned billions on its metaverse adventures, with little to show for it. But Zuck told investors "We're not too far off from being able to deliver great looking glasses that let you seamlessly blend the physical and digital worlds so you can feel present with anyone no matter where they are. And we're starting to see the next computing platform come together and it's pretty exciting." AI will be a part of that experience - and many others across Meta's product portfolio. Another thing on Zuck's mind is customer service. "We do want to continue making progress on this vision of making it so that any small business or any business over time can, with a few clicks, stand up an AI agent that can help do customer service and sell things to all of their customers around the world," he said. If Meta can do that it will scare plenty of software vendors and e-commerce players, and perhaps propel its growth beyond even the $45 to $48 billion of revenue predicted for Q4. The biz decided not to discuss 2025 revenue until next quarter. Investors weren't excited by these results, which fell a little short on the earnings front. Meta shares dropped three to four percent in after-hours trading. ®
[2]
Mark Zuckerberg Says Meta's Open-Source Gambit Is Paying Off As Nvidia And AMD Optimize For Llama: 'A Big Deal On Several Fronts' - Meta Platforms (NASDAQ:META), Advanced Micro Devices (NASDAQ:AMD)
Meta Platforms Inc META CEO Mark Zuckerberg revealed during the company's third quarter 2024 earnings call that Meta is developing its next-generation AI model, Llama 4, using what he described as an unprecedented computing infrastructure of more than 100,000 NVIDIA Corp NVDA H100 GPUs. What Happened: "We're training the Llama 4 models on a cluster that is bigger than 100,000 H100s or bigger than anything that I've seen reported for what others are doing," Zuckerberg said during the earnings call. The new models are expected to be ready in early 2025, with smaller versions launching first. Zuckerberg also highlighted how Meta's open-source approach to Llama is yielding benefits through chip optimization from major semiconductor manufacturers. "What we're seeing is that as Llama gets adopted more, you're seeing folks like NVIDIA and Advanced Micro Devices, Inc. AMD optimize their chips more to run Llama specifically well, which clearly benefits us," he explained. The Meta CEO drew parallels to the company's previous success with its Open Compute Project, noting that sharing technological designs led to industry standardization and cost savings. This strategy appears to be repeating with Llama, as chip manufacturers optimize their hardware specifically for the AI model. The upcoming Llama 4 is expected to feature significant improvements over its predecessor, with Zuckerberg indicating it will be "a big deal on several fronts, new modalities, capabilities, stronger reasoning, and much faster." See Also: Riot Platforms Stock Falls After Worse-Than-Expected Q3 Results Why It Matters: The development of Llama 4 is a significant step for Meta Platforms, especially given the company's recent strong financial performance. Meta Platforms reported third-quarter revenue of $40.59 billion, beating analyst estimates of $40.29 billion. The company also reported third-quarter adjusted earnings of $6.03 per share, surpassing analyst estimates of $5.25 per share. Despite this, Meta Platforms has faced some internal challenges, including the recent firing of approximately two dozen employees, including one earning $400,000 per year, for violating the company's meal credit rules. Price Action: Meta stock closed at $591.80, down 0.25% on Wednesday. In after-hours trading, the stock declined further, dropping 3.18%. Year to date, Meta has shown substantial growth, surging by 70.90%, according to data from Benzinga Pro. Read Next: Microsoft Q1 Earnings Highlights: 'AI-Driven Transformation Is Changing Work,' Company Beats Revenue, EPS Estimates Image Via Shutterstock Disclaimer: This content was partially produced with the help of AI tools and was reviewed and published by Benzinga editors. Market News and Data brought to you by Benzinga APIs
Share
Share
Copy Link
Meta's CEO Mark Zuckerberg reveals the company's strategy behind open-sourcing Llama AI models, highlighting cost savings and industry-wide benefits. The development of Llama 4 and its implications for Meta's future in AI are discussed.
Meta Platforms, under CEO Mark Zuckerberg's leadership, is leveraging an open-source approach for its Llama AI models, a strategy that is proving to be both cost-effective and industry-shaping. During Meta's Q3 2024 earnings call, Zuckerberg revealed that this approach is not purely altruistic but also a savvy business move 1.
Zuckerberg announced that Meta is developing its next-generation AI model, Llama 4, using an unprecedented computing infrastructure of over 100,000 NVIDIA H100 GPUs. This massive cluster surpasses any reported by competitors, positioning Meta at the forefront of AI model training 2.
The open-sourcing of Llama has led to unexpected benefits. Major chip manufacturers like NVIDIA and AMD are now optimizing their hardware specifically for Llama, which in turn benefits Meta's products. This industry-wide adoption and optimization echo the success of Meta's previous Open Compute Project 1.
Meta's financial performance remains strong, with Q3 2024 revenue reaching $40.6 billion, a 19% year-over-year increase. The company's net income rose by 37% to $15.7 billion 1. However, Meta's Reality Labs division, focused on metaverse and mixed reality, continues to operate at a loss, with a $4.4 billion loss in the quarter 1.
Zuckerberg emphasized the integration of AI across Meta's product portfolio, including plans for AI-powered customer service for businesses. This move could potentially disrupt software vendors and e-commerce players 1. Meta is also exploring AI applications in content moderation, which could impact its current moderation workforce 1.
Meta's infrastructure spending is on the rise, with $9.2 billion invested in servers, datacenters, and network infrastructure in Q3. The company expects to spend between $38 billion and $40 billion for the full year, with plans for "significant capital expenditure growth" in 2025 1.
Llama 4 is expected to launch in early 2025, with smaller versions releasing first. Zuckerberg promises significant improvements in capabilities, reasoning, and speed 2. Despite the positive outlook, Meta's shares experienced a slight drop in after-hours trading following the earnings report 1.
As Meta continues to push the boundaries of AI development and application, its open-source strategy appears to be paying dividends, both in terms of technological advancement and industry influence. The coming years will likely see further integration of AI across Meta's platforms, potentially reshaping the digital landscape.
Meta Platforms unveils Llama 3, a powerful open-source AI model, potentially disrupting the AI industry. The move aims to enhance developer freedom, privacy standards, and Meta's competitive position against rivals like OpenAI and Anthropic.
4 Sources
4 Sources
Mark Zuckerberg announces Meta's Llama 3 model, claiming superior cost-performance over ChatGPT. He also reveals plans for an AI assistant to compete with Siri, Alexa, and ChatGPT.
2 Sources
2 Sources
Meta's decision to open-source LLaMA 3.1 marks a significant shift in AI development strategy. This move is seen as a way to accelerate AI innovation while potentially saving Meta's Metaverse vision.
6 Sources
6 Sources
Meta Platforms experiences a 16-day stock rally, driven by successful AI investments and strategies. The company's focus on AI-powered advertising tools and open-source AI models positions it as a leader in the AI space, rivaling Nvidia's success.
9 Sources
9 Sources
Meta CEO Mark Zuckerberg announces the development of Llama 4, utilizing over 100,000 NVIDIA H100 GPUs in a record-breaking AI training cluster, with plans for release in early 2025.
10 Sources
10 Sources
The Outpost is a comprehensive collection of curated artificial intelligence software tools that cater to the needs of small business owners, bloggers, artists, musicians, entrepreneurs, marketers, writers, and researchers.
© 2025 TheOutpost.AI All rights reserved