Meta, IBM Form AI Alliance to Counter OpenAI with More Open AI

0:01Today on the AI Breakdown, we're looking at the new AI alliance led by Meta and IBM and IBM, stands as a direct contrast to open AI. Before that on the brief, Google delays Gemini again, but may be doing virtual previews as soon as this week. The AI Breakdown is a daily podcast and video about the most important news and discussions in AI. Go to breakdown.network for more information about our YouTube, our Discord, and our newsletter. Welcome back to the AI Breakdown Brief, all the AI headline news you need in around five minutes. Another day, another delay of Google Gemini. For the last few weeks, the only news we've gotten about Google Gemini are, in fact, delays.

0:41The latest comes from the information Google preps public preview of Gemini AI after postponing in-person launch events. Now, this is actually a two-part story. The first part was about Google scrapping a series of in-person events that were meant to launch Gemini, and then the second part, which was updated just yesterday, was the report from inside the company that instead they were planning a virtual preview that could come as soon as this week. All right, so let's go to the first part of the story first. The TLDR is that Google had been planning to debut Gemini, which has just an absolute ton of pressure around it.

1:13This can't just be another interesting, even highly capable AI model. It has to, at the very least, meet GPT-4 capabilities, and frankly probably needs to exceed them at least in some fairly meaningful ways. Writes the information, Google CEO Sundar Pichai recently decided to scrap a series of Gemini events originally scheduled for next week in California, New York, and Washington after the company found the AI didn't reliably handle some non-English queries. The planned events, which hadn't been publicized, would have marked Google's most important product launch of the year after it strained its computing resources and merged large teams in an urgent pursuit of open AI.

1:48The publication writes, It's rare for Google to launch a major product between Thanksgiving and the end of the year, but Google intended to make an exception for Gemini because it's arguably the company's most important initiative in a decade. The Gemini event in Washington was intended to showcase the technology to policymakers and politicians, which have increasingly discussed potential regulations involving AI. Pichai and other executives have been concerned that OpenAI's ChatGPT has become a household name, and that Microsoft's co-pilot features for productivity software, including for software developers, are turning into a significant business too.

2:17Now, here was the most interesting paragraph to me. A key challenge for the Gemini team is making sure the primary model is as good as or better than GPT-4. It has met that standard in some respects, said one of the people familiar with it, but the company is still making improvements because it wants the technology to work well globally in numerous languages. Now to me, the whole thrust of this is very clearly how much pressure Google feels on this release. As of the time of recording, we haven't got any more information about these virtual previews, but apparently Google's team has been for weeks giving private demonstrations of Gemini to business partners, even though they say that people won't actually be able to get their hands on it until January.

2:52Anyways, yet another interesting moment in the saga that is Google Gemini, and I will certainly be keeping a lookout for any more information about these virtual previews should it come out. Now, one interesting aside, given how generally Google's other AI model that is available in BARD has been treated, Neil Patel from NPD shared some really interesting information today. He wrote, Do people prefer content written by ChatGPT or BARD? We had ChatGPT create 1 ,000 articles on various topics. We then gave BARD the same instructions on the same topics. In total, we created 2 ,000 pieces of content with AI.

3:251 ,000 from ChatGPT, 1 ,000 from BARD. We then paid people to read the articles. We didn't tell them that they were written by AI based on their interests. In total, 249 articles were read from ChatGPT and the 249 counterpart versions from BARD. We then asked which one they preferred. Interestingly, BARD absolutely crushed ChatGPT in this one study. 188 of the articles written by Bard were preferred to just 61 written by ChatGPT. In other words, 75.51 % of the time, people preferred the Bard articles. Now, the one thing that people are asking is whether this was GPT 3.5 or whether it was GPT 4, and obviously without that, it's incomplete information, but still an interesting piece of counterintuitive data.

4:05Next up on the brief, two interesting stories surrounding Getty images. The first is that their lawsuit against Stability AI is heading to trial in the United Kingdom. Basically, a UK court has ruled that there are merits to Getty's arguments that its copyrighted material was used to train AI models. And in spite of Stability's argument that the case should not be heard in the UK, because as The Verge put it, it said no one involved in the training or development of Stable Diffusion was based in the UK, a UK justice, Joanna Smith, didn't agree. Instead, she said that evidence from the company and its CEO, quote, raised the specter that evidence is either inaccurate or incomplete, at the very least suggest a conflict of evidence.

4:40So, on the one hand, we have Getty Images suing a company, Instability AI, that is using their images they claim in a way that violates copyright to train AI models. But then we also have the announcement of a partnership between Getty Images and Runway, which is of course a text-to-video leader. A Runway blog post reads, Runway is partnering with Getty Images to launch a new video model for enterprise customers, addressing companies' ever-growing need for high-quality customized content. This model will combine the power of Runway with Getty Images' world-class, fully licensed creative content library, providing a new way to bring ideas and stories to life through video in enterprise-ready and safe ways.

5:13Now, basically what's going on here is that Runway is trying to get out ahead of more business use cases of their text-to-video tools by partnering with the content creators so that they can get around these potential legal issues. Again, from their blog post, they write, this new Runway Getty Images model, RGM, will provide a baseline model upon which companies can build their own custom models for the generation of video content. Runway Enterprise customers will be able to fine-tune RGM using their own proprietary datasets. This enables companies in all fields, Hollywood studios, advertising, media broadcasting, and more, to enhance their creative capabilities and provide new channels for video creation by powering entirely new content workflows and making it easy to craft delightful experiences tailored to Enterprise's styles and brand identities and to their unique audiences.

5:56Now, get outside of all the announcement language here, and what this is really saying is that a lot of different companies and a lot of different industries are going to use these text-to-video tools, and here's a version where you can be confident that you're not going to get sued for copyright violations. Now, of course, what's interesting to me is that we're seeing Getty Images play both sides of the AI space. On the one hand, they're suing people that they say are violating their copyright, and on the other hand, they're partnering with companies to provide copyright-protected versions of those models.

6:21It's frankly a pretty smart strategy from a business perspective, and kind of gives them two bites at the apple if court decisions don't go their way. That's not to say that the questions that are playing out in court aren't significant and very, very complex. Billboard recently wrote a piece called As AI Grows, Artisan Labels Consider Who Owns a Voice? Some record deals they write may cover voice rights that could be used for artificial intelligence. Artists are negotiating for clarity and exceptions. Rights are an extremely complicated thing. In music, for example, there is already an established difference between rights that relate to the recordings of songs and rights that relate to the underlying melodies and words.

6:55This is the reason that Taylor Swift can re-record her own music. As per the terms of her old contracts, while she doesn't own the masters to her recordings, she does own the publishing rights. So basically, she granted herself the right to cover her own songs, producing a version where she now owns not only the publishing rights, but also the actual recording rights as well. Effectively, this Billboard article is about a new vector for rights, which is, of course, who owns voices and who can use them in what ways when it comes to AI. I think this is a question that is going to get immensely more complicated before it gets solved.

7:25Over in the world of big AI fundraising, Fidelity has just led a new investment round in Vast Data. Vast is an AI software and storage startup that has been backed by other companies, including NVIDIA, and the latest round values the company at more than$9 billion. That's up from$3.7 billion back in 2021. From the information, Vast Data sells software to companies like Pixar, Verizon, and high-frequency traders to store data for AI applications. The 600-person company is one of a number of startups building digital or physical infrastructure for AI-focused apps that has gotten a lift from the rise of ChatGPT.

7:57Fast expects to pull in more than$100 million in revenue this year, more than triple last year's sales, one of the people said. Speaking of big fundraisers, much buzzed about Mistral, whose open-source 7B model, LLM, has gotten a ton of buzz among AI developers, has raised a new€450 million or$487 million round that values the company around$2 billion. The investment is being led by Andreessen Horowitz, but also includes participation from NVIDIA and Salesforce. force. Writes Bloomberg, the$2 billion valuation for a company less than a year old underscores the tech world's unbridled optimism about the future promise and profit of artificial intelligence companies.

8:33Now, what's interesting to me is that between their first round, which came pre-product and was once again a nine-figure round, and this one, the company has actually released a large language model that people really like. And if you are interested in the open-source competition with more closed models from the big AI labs, well, then stick around for the main part of the episode, which is coming up next. Quickly, a brief word from today's sponsor. As a listener of this show, I suspect you like to stay up to date on all things AI and tech, which is why you have to check out the chart-topping podcast Web3 with A16Z Crypto.

9:07Produced by venture firm Andreessen Horowitz, Web3 with A16Z is the perfect companion podcast to the AI breakdown. Web3 with A16Z Crypto is your definitive resource for the future of the internet, whether you're interested in the convergence of AI and crypto or simply curious about what's next. If you need a place to start, they recently released an excellent episode with Stanford cryptography professor Dan Bonet and former Google X engineer Ali Yahya in conversation with host Sonal Choksi about the intersection of AI and crypto. From fighting deepfakes and proving humanity to large language models like ChatGPT, they cover it all.

9:40I highly recommend checking it out, especially if you'd like to learn more about how AI and crypto will impact our everyday lives. Beyond crypto and AI, this show is for creators seeking more ways to truly own their work, for business leaders trying to prepare for the future today, and for innovators exploring trending tech topics. Don't miss out. Follow Web3 with A16Z Crypto on Apple Podcasts, Spotify, or your favorite listening app. Welcome back to the AI Breakdown. One of the big themes for the show over the last week or so, which is pretty inevitable and always going to happen is where the industry is at the one-year anniversary of ChatGPT.

10:17Whether you like ChatGPT or loathe it or like OpenAI or dislike them, there's no denying that the launch of ChatGPT set off this wave of generative AI and will be seen as the starting point for this whole movement and industry that we're all a part of. Now, of course, the drama surrounding the OpenAI leadership has been the dominant theme. However, it's not the only one. Today, we are discussing a new alliance led by Meta and IBM, which is a fairly strong statement of disagreement with the way in which OpenAI and some of their peers are behaving as relates to the artificial intelligence industry.

10:50So let's talk about what the news is, and then we'll do a little bit of analysis and interpretation. The Wall Street Journal writes, Meta and IBM launch AI Alliance. One year after the debut of ChatGPT, companies including AMD and ServiceNow are teaming up with academia to promote open alternatives to open AI. So, a couple things to note about this alliance. First of all, it is not just Meta and IBM. There are more than 50 AI companies and research institutes that are part of this alliance. As I mentioned, it is called, perhaps not all that creatively and certainly with difficulty to Google, the AI Alliance.

11:23Other members include Intel, Oracle, Cornell University, and the National Science Foundation. The journal describes its members as, quote, largely supporting open source, an approach in which technology is shared free and draws on a history of collaboration among big tech, academics, and a fervent movement of independent programmers. And to be clear, this is absolutely a response to the way that the industry has developed and specifically the way that narratives around the industry have developed. Said Dario Gil, Senior Vice President at IBM and Director of IBM Research, Frankly, we've been a little bit unsatisfied with the overall debate and the discussions on AI over the last year.

11:57We did not feel that it reflected the diversity of the ecosystem that is making this AI moment possible. Said Meta's president of global affairs, Nick Clegg, we believe it's better when AI is developed openly. More people can access the benefits, build innovative products, and work on safety. Now, in addition to any sort of statement of principle, there is also a clear business objective here. Again, from the Wall Street Journal, many of the Alliance's members are companies that have their own AI products but are struggling to catch up with the rush of attention that OpenAI and its investment partner Microsoft are drawing.

12:26The article also points out that although they have been working on this alliance since all the way back in August, the timing is opportune. Quote, since the upheaval at OpenAI in late November, businesses want to have more providers of AI products to diminish the risk of working with a single vendor and are exploring other AI systems as viable alternatives. The timing of the AI alliance's launch underscores that message IBM's Gill said. This other way, it's a much more distributed approach but much more resilient because no given institution can derail the success of the open engine. It definitely feels a little bit like an alliance of companies who just a little behind.

12:58For example, quote, Advanced Micro Devices, the chipmaker aiming to take a piece of NVIDIA's dominance in AI chips, said that it will support an open AI ecosystem with its hardware and that it, along with other alliance members, will build the software that enables businesses to use the chips, said Forrest Narad, AMD's executive vice president and general manager of its data center group. The company is set to spotlight AI accelerator chips this week that Narad says will be a strong alternative to NVIDIA's offerings. Said IBM's Gill, If you think the future of AI is going to be determined by two, three, or five institutions, you're mistaken.

13:29I hope that it gives more clarity and confidence that the world of open innovation is a world to bet in. Now, interestingly, there is far more coverage given to those competitive dynamics of the Alliance than there are to what they're actually doing. For example, the only real mention of what they're doing in that Wall Street Journal piece are the last couple sentences which say, The Alliance is focusing on six areas, including regulation and safety, as near-term initiatives. Gill said it would soon release a benchmarking tool for AI safety and model validation. Now, what's really interesting about this to me is that on the one hand, it's totally easy to be cynical about.

14:01To see this as something that is forced largely by competitive pressures. Basically, the open AIs and anthropics of the world, along with their big tech partners, have sucked all the oxygen out of the room, dominating not only the media coverage, but by extension, the enterprise business. And this is a way to try to reclaim some of that narrative space. However, even if that's true, that's not to say that this message, that the future of AI is not going to be determined by two, three, or five institutions, isn't an important and resonant one. TechCrunch's coverage of the alliance gets at some of the skepticism.

14:30In fact, the title of their piece is Meta and IBM Form an AI Alliance, But To What End? Now, part of TechCrunch's questions come from the fact that there are already similar alliances to this in the space. They write, So what will the AI Alliance do exactly, and how will its work differ from the quite similar, at least in terms of its overarching mission, members and tenants partnership on AI? The partnership on AI years ago promised to publish research using open source licenses in minutes from its meetings to, as the AI Alliance purportedly seeks to do, educate the public on pressing AI issues of the day.

14:59Well, confusingly, the partnership on AI is in fact a member of the AI Alliance. Giving a little bit more detail on the practicals, TechCrunch writes, The AI Alliance's members will first form working groups, a governing board, and a technical oversight committee dedicated to advancing areas like AI trust and validation metrics, hardware and infrastructure that supports AI training, and open-source AI models and frameworks. They'll also establish product standards and guidelines and then partner with important existing initiatives, initiatives conspicuously not named in the press release, from government, non-profit, and civil society organizations who are, quote, doing valuable and aligned work in the AI space.

15:31If that sounds a lot like what the inaugural members of the Alliance were already doing independently, you're not wrong. But in the release, the AI Alliance stresses that its work, whatever form it ultimately takes, is intended to be complementary and additive rather than needlessly duplicative. The release reads, More collaboration and information sharing will help the community innovate faster and more inclusively, and identify specific risks and mitigate those risks before putting a product into the world. That stands in contrast to a vision that aims to regulate AI innovation and value creation to a small number of companies with a closed proprietary vision for the AI industry.

16:03Now, in a section called Key Subtext, it gets into the fact that there is clearly a dividing line sort of battle happening here that is being calcified and hardened by this announcement. On the one hand, you have companies who are not included in this AI alliance, that include Google, OpenAI, Microsoft, Anthropic, Cohere, who are skeptical, if not downright opposed, to the open source approach taken by a company like Meta. The piece points out that also notably absent are institutions like Stanford and MIT, and that NVIDIA isn't a member either. They write, I'll note that NVIDIA isn't a member of the AI Alliance either, a suspect absence given that the company is by far the dominant provider of AI chips and a maintainer of many open-source models in its own right.

16:39Perhaps the chipmaker perceived a conflict of interest in collaborating with Intel and AMD, or perhaps it decided to cast its lot with Microsoft, Google, and the rest of the tech giants opting out of the Alliance for strategic reasons. Who can say? Now, the flip side is that they did note that the groups that have come together for the AI Alliance represent a very wide swath of companies. Others they name include CERN, Yale, Imperial College of London, Stability AI, Hugging Face, and more. TechCrunch writes, the AI Alliance's initial cohort is exceptionally broad, sitting at the intersection of not just AI and enterprise, but healthcare, silicon, and software as a service as well.

17:11But they write, without the participation of so many major AI industry players and lacking deadlines or even concrete objectives, can the AI Alliance succeed? What would success even look like? Beats me. The vast number of competing interests from healthcare networks to insurance providers won't make it easy for the Alliance's members to coalesce around a single united front. And for all their talk of openness, IBM and Meta aren't exactly the poster children for the future that the Alliance's release depicts, casting doubt on their sincerity. Perhaps I'm wrong and the AI Alliance will be a smash success, or perhaps it'll crumble under mistrust in its own bureaucracy.

17:42We'll see. Time will tell. Now, overall, I just don't think it's all that complicated. There is a major dividing line in and around the artificial intelligence space, around people's attitudes towards open source and openness in general. To me, the AI Alliance simply reads like a codification of those sides, but those sides aren't necessarily any different than they were yesterday before the AI Alliance was announced. Now, is it opportunistic for certain of these members to have joined the AI Alliance versus the uncodified alliance of big tech and closed AI people? Sure, maybe. And is it opportunistic, or at least opportune, to use the chaos of OpenAI in the last few weeks as a moment to try to convince particularly enterprise customers and policymakers that maybe a more decentralized approach that is more open and less reliant on a small handful of companies might be a better one for this critical infrastructure and new technology?

18:31Sure. But as the famous saying goes, all's fair in an AI arms race. I think it's reasonable to have the sort of skepticism that that TechCrunch author does, that such a broad-based coalition could actually find much to coalesce around. But even for those who are skeptical of open source AI and who are nervous about what it might mean, having a body where proposals for standards and guardrails and things like that can at least be discussed. In other words, where there is a space for ideas to flow between these different people who share this sort of sentiment seems like a net good thing rather than a bad thing.

19:00So overall, I'm not sure. It could be that this ends up being a nothing burger, just a nice press release and a thing to discuss on podcasts for a day or two, or it could be something more. No matter what, I do think it reflects the fact that we were moving to a new stage of the conversation, one with real policy implications, where enterprises are making buying decisions. In other words, one where the stakes are higher than ever. Interesting stuff at the beginning of this week and something to keep an eye on for sure. For now, however, that is going to do it for the AI Breakdown. I appreciate you listening or watching.

19:28Until next time, peace.