How Moderna's ChatGPT Use Shows the Future of Corporate AI

0:00Today on the AI Breakdown, why the OpenAI Moderna deal shows off the future not just of AI in healthcare, but AI in corporate settings more broadly. Before that on the brief, big ol' valuations for some very young AI startups. 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 Discord, our newsletter, and our YouTube. Welcome back to the AI Breakdown Brief, all the AI headline news you need in around five minutes. We kick off today with yet another story of a very young company raising money at a very high valuation.

0:36The company this time is Cognition Labs. You might have seen the demo for their AI coding assistant, Devin, which initially blew people away with its ability to do much more autonomously than it seemed like other coding assistants could do, although subsequent to that initial launch, there has also been a little bit of controversy around whether the company perhaps oversold its capabilities, at least from a PR standpoint. Founders Fund seems to think that the hype is real. The venture firm has led a$175 million investment into the company, valuing the company at$2 billion. This is the second round the company has raised in the last two months, with Founders Fund leading their Series A at a$350 million valuation just a month earlier.

1:15Now, of course, the question with something like this is how much it reflects the specifics of this particular company, product, or area versus just overall hype in the AI space. My perception is that many think that when it comes to the next iteration of AI focused on AI agents, coding agents are going to be one of the first place where broad adoption happens, and the transformation to how coding works is going to be extremely significant. In other words, this is an opportunity that investors see as absolutely huge. And so when a firm seems like it has something genuinely unique and differentiated in that area, it's not surprising, perhaps, that it's going to attract a significant amount of capital at very generous valuations.

1:52Still, for many observers, it's hard not to see some bubbly signs. In another piece of the information, they write, Venture FOMO hits robotics as young startups gets$1.5 billion valuation. They write, Venture investors are paying a hefty premium to back robotic startups harnessing artificial intelligence. hoping one of them can become the robotics version of OpenAI. The latest version, they say, is Skilled, which they report is nearing a deal to raise nearly$300 million at a$1.5 billion valuation. Now, they will say that the deal isn't done yet and the terms could change, but the fact of the matter is that this is a company that right now doesn't have any revenue.

2:25The information points out that when it comes to robotics, there is something of a boom going on right now. In the first quarter of 2023, robotics companies raised$1.7 billion, dollars, and that was up to$3.2 billion in funding for the same period this year. Now, we've talked previously on the show about Figure, which earlier this year raised around $675 million at a$2 billion valuation from people like Bezos, OpenAI, NVIDIA, and others. And another company called Physical Intelligence raised a$70 million pre-seed round at a$400 million valuation. Now, that company, Physical Intelligence, is trying to do something a little bit different than these companies which are building humanoid robots, focused on, as they put it building the brain or the operating system that runs underneath any given robot.

3:06Now, speaking of the big cost to compete in the AI space, Meta has reported its earnings, projecting that revenue for the current quarter would be lower than what Wall Street anticipated. What's more, the company said that in their battle to create advanced AI, they were likely going to spend billions of dollars that would cut into profits. Meta raised its spending forecast for this year from$30 to$37 billion, where it had been, to$35 to$40 billion, driven by AI infrastructure, data center, chip design, research and development, and more. Now, we could get deep into what it says about the state of current markets that Metastock plummeted after this.

3:40But I think the more interesting thing for our purposes is that Zuckerberg is laying down the gauntlet here. He is making it clear that they are going to compete in this space, and they're going to spend the money to do so. Over in the land of chips, NVIDIA is buying two Israeli startups that both make NVIDIA's AI chips easier to use. Writes the information, During the AI boom, chipmaker NVIDIA became one of the most active corporate venture capitalists, investing in over 30 startups last year alone. Now it may become one of the most active acquirers of AI startups. These acquisitions are yet to be formally announced, but they include a company called RunAI and a company called Desi, about which the information writes, Both venture-backed startups aim to lower the cost of developing or running AI models that generate text images and code, which is important to NVIDIA's hopes of maintaining demand for its specialized chips that power these models.

4:27Basically, the idea here is that NVIDIA wants to bring down the cost of running models in order to keep companies coming back to buy more AI, which in turn will create demand for its GPUs. These companies tackle that cost problem in different ways. RunAI works to make AI chips more efficient by allowing developers to run multiple AI workloads in parallel rather than sequentially, which in aggregate reduces the number of NVIDIA GPUs needed to complete a task, while DESI tweaks AI models themselves so they can run more cheaply on AI chips. One Israeli news outlet is reporting that NVIDIA is paying$700 million for Run.AI, while Desi's purchase price has not yet been reported.

5:02Finally today, when Drake released a diss track featuring AI-generated Tupac and AI-generated Snoop, the dogfather responded with an Instagram post where he basically laughed it off and said, what the heck is going on? Tupac Shakur's estate has taken a slightly different approach, threatening to sue Drake over the track. They wrote that this infringed on Tupac's legacy and that they would never have given their approval. And so the music industry continues to be a central battleground for the future of AI from a legal perspective. For now though, that is going to do it for the AI Breakdown Brief.

5:33Next up, the main AI breakdown. Attention, AI Breakdown listeners. Consensus 2024 marks the 10th gathering for all things crypto, blockchain, and Web3. However, importantly, this year's agenda will also dive deep into AI-driven transformation. And the speaker lineup includes the leading minds and innovators at the forefront of this digital renaissance. Don't miss the Consensus AI Summit to cut through the hype to find where true transformation and opportunity lie. Listeners to this show can get 15 % off registration with the code AIBreakdown. Visit consensus2024.coindesk.com to learn more. Some of the folks who will be at Consensus this year include Guillaume Verdun, aka Beth Jezos, founder and CEO of XTropic, as well as spiritual leader of the accelerationist movement.

6:16Neil Stephenson, co-founder of Lamina One, and Brendan Eich, the CEO of Brave Software. Again, go to consensus2024.coindesk.com to learn more and get 15 % off registration with the code AIBreakdown. Before we get back into the show, I want to tell you about the sponsor of today's episode, Plum. As a founder, your biggest fear is building something no one wants. But with the rapid pace of AI, you can't afford to spend months experimenting. That's where Plum comes in. Their no-code AI app builder lets you prototype, test, and deploy your ideas in days, not months. Validate swiftly and get your product in front of customers fast.

6:49Don't waste precious time and resources chasing the wrong ideas. Check out useplumb.com, that's Plumb with a B, P-L-U-M-B, for early access to the future of AI app development. By the way, Plumb actually built a pipeline to auto-generate ad copy that I then just put my own spin on. Pretty cool and a great use of a very cool technology. Thanks once again to Plumb for their sponsorship. Check them out at useplumb.com. That's plumb with a B dot com. Welcome back to the AI Breakdown. Today we are looking at a deal that was announced yesterday. It's a partnership between Moderna and OpenAI, and interestingly enough, I think the deal is less notable in terms of the implications of AI for health and science, and more as a model of what future corporate usage looks like.

7:34So first let's talk about what was announced, and then we'll get a little bit deeper into this analysis. The deal was first announced in the Wall Street Journal, who characterized it by saying, the deal aims to automate nearly every business process at the biotechnology company and boost the ChatGPT makers reach into the enterprise. So basically, the deal is going to give 3 ,000 Moderna employees access to ChatGPT enterprise for their teams to use it in all the various ways that their teams might use it. Moderna says that it plans to roll out 15 new products within the next five years, and they see this as integral to reaching that goal.

8:06Now, there has been a lot discussed around the impact of artificial intelligence on scientific discovery. OpenAI CEO Sam Altman has frequently said that that is what he thinks the biggest disruption from AI will be, when it truly allows us to unlock new scientific discoveries that wouldn't have been possible before or would have taken just a hugely long amount of time. At the same time, Altman has also frequently said that the GPT-4 that is available isn't really close yet to being able to do that, that it's a far-off thing. And indeed, he echoed that in this announcement saying, eventually AI will be able to do more and more science, but the best way to get there is to enhance the productivity of people and accelerate what they can do.

8:44In other words, AI isn't doing the science yet, the scientists are doing the science. So let's make the scientists work more efficiently, effectively, and better. Now, a big part of the coverage has focused on where Moderna is in its life cycle. Moderna became a household name during the COVID pandemic, as it was one of the first to produce a COVID-19 vaccine with its messenger RNA platform. However, as people's relationship to COVID-19 has changed, the company's fortunes have changed as well. The journal points out that fourth quarter revenue in 2023 was$2.8 billion, down from$5.1 billion the year before.

9:16This is, of course, why Moderna is looking to expand the products in its roster. So when it comes to what they're looking for from ChatGPT, it's very much about how they make their team work better than before. Said their CEO, how do we use it at scale to reinvent all of Moderna's business processes? in science, in legal, in manufacturing everywhere. Indeed, that same CEO said that his goal is for Moderna employees to use ChatGPT no fewer than 20 times a day. What's interesting is that this is not just some top-down imperative. It's clearly happening inside. The journal writes that Moderna employees have created over 750 unique GPTs, which are, of course, the customized version of ChatGPT that are meant to make it easier to do specific tasks or processes over and over again.

9:59Some of the examples they give include a GPT that uses years of previous research and medical knowledge to predict the optimal dose of a drug for clinical trials. Another GPT reviews research to draft answers to questions from regulators. Another one helps predict the structure of new enzymes that will enable manufacturing processes with better yield and reduced waste. So what I'm interested in here is not just the drug discovery pharmaceutical use case for ChatGPT and AI, but the fact that this company is looking to apply it across the entire organization. This is not some pilot project where 50 people on a marketing team are given access to a tool.

10:34This is the CEO of this company telling everyone to reimagine how they do their work from the ground up with these tools. I think what it shows is that AI does not belong to any one specific department. It is going to change how every department does their work. Companies that get that faster will be quicker to figure out what those most effective use cases are and be better able to share up what works across their organizations. Now, of course, one always has to be skeptical when reading press-release-based stories around whether the traction of ChatGPT inside Moderna is really what it seems to be.

11:04However, every note that I've found on Twitter and in other social media really suggests that this is something that is being adopted in a major way inside this company. Christian Hendrickson writes, Some remarkable points. 100 % adoption of ChatGPT in the legal team. Integration into core processes like dose distribution. A ton of custom GPTs for specific workflows. Scaling up necessitates AI assistance in lieu of hiring 100 ,000 employees. Further dramatizing at Christian writes, People are still being like, yeah, ChatGPT can write emails, but meh. And then meanwhile, Moderna are just using it directly in their clinical trials.

11:37What a stark difference in view of the technology. Investor Drew Volpe writes, Have been talking with friends at Moderna about their open AI usage internally for a while, and it's very real. The beginning of LLMs changing how researchers work. So when push comes to shove, I think it would be a mistake to view this as a story about how the pharmaceutical industry is using AI. Instead, I think it needs to be seen as an exemplar of what it looks like when an organization adopts an AI like ChatGPT end-to-end, top-to-bottom. In that way, I think its lessons are relevant not just for the health and pharmaceutical industries, but for any company that's reimagining processes based on AI.

12:14Pretty cool stuff, and I hope that we get more information about exactly how they end up using it, because of course, the more examples we have of how people are successfully using AI, the more others are likely to adapt it as well. However, that is going to do it for today's AI Breakdown. Until next time, peace.