OpenAI’s Secret ‘Garlic’ Model, C3 AI’s Turnaround Plan, Robotics and D.C. | Dec 4, 2025

0:13Welcome everyone to the information's TITB. My name is Akash Pashrich. It is Thursday, December 4th. We have got a great lineup for you today. First up, we are diving deep into new reporting on web publishers' newfound confusion around whether or not heightened traffic on their websites is coming from bots or from humans. We'll then turn to some earnings news, including a conversation with the new CEO of C3AI as the company faces falling revenue and a stock that has taken a steep hit this year. We'll also check in with our AI reporter who is tracking OpenAI's efforts to counter Google's new Gemini 3 model with its own in-development system codenamed GARLIC.

0:54I don't know why it's called GARLIC. That's going to be the first question we ask her. And last but not least, we will wrap up with Robotics Roundup, breaking down the biggest headlines and news and what they mean for the sector's rapid growth. It's going to be a big show, so let's get right on into things. Chatbots are wreaking havoc for companies operating websites because the companies behind those sites can't figure out whether or not this new traffic they're seeing is in some cases coming from real people or from bots. Joining me now to discuss this is Catherine Perloff. Catherine, welcome to the show.

1:27It's great to see you. Hi. So what's going on here? How did you even come across this issue of there being heightened traffic and people not knowing where it's coming from? Yeah, well, it actually started, and we'll get into this, because I was talking to a publisher who was saying that they were having trouble blocking Grok because they couldn't tell, you know, Grok being the XAI chatbot, because they couldn't tell whether it was a bot or not, and that made it hard to block. And then when I dug more into the issue, I saw that this is a pretty widespread problem. You know, scraping has been going on for a while, crawlers, bots to websites.

2:06But it's taken on sort of a new significance for websites in the age of AI when they feel, you know, that the AI companies are taking their content without paying for it or, you know, replacing the visitors that advertisers are paying them to reach. So blocking bots has become a bigger priority for websites as these AI chatbots become more popular and the sort of threat they pose to websites, especially web publishers. But in some cases, retailers has become more important. But the actual blocking of the bots is sort of a challenge. So the end goal is to block the bots. But to do that, we have to figure out if it's a human or a bot in the first place.

2:50Why is it so hard to detect which is which? Yeah, and one caveat is, you know, not all websites want to block all bots, especially with the rise of like agents. You know, you're an e-commerce site, you might want the agent to come to your site and let people make a purchase. But yeah, it's so hard because, you know, a lot of this kind of depends on what the company sending the bot says and how they present their bot. You know, OpenAI, for example, people told me was pretty good about being pretty transparent about their bots and, you know, telling websites, hey, I'm coming to your site to scrape.

3:30I'm coming to your site because a user asked a question in search. I'm coming to your site with my agent. You know, they're pretty transparent. But not all AI companies are. I think the whole agentic issue makes it complicated because, like, you know, if it's a human asking an AI to do something, does that make it a human or a bot? Is it a human visitor or a bot visitor? But, yeah, I think also a lot of AI companies also know that websites don't want them there. So if they look more like a human visitor, then they will not be detected. They won't be blocked. and then they can get the content they want, which can inform their models.

4:10So who are the companies now that are making this an opportunity for them? You outlined a couple of the startups and a couple of the bigger tech companies. Is this a big business now helping websites figure this out? Yeah, you know, it's interesting. So I mean, I kind of focus on two companies in my story, though there is many companies that work on this. One is Datadome, which is kind of a specialist just in the blocking of bots. One is Cloudflare, which is a company more known as a CBN, or so it helps websites sort of just kind of connect with servers and deliver the content. But it also offers this bot blocking software.

4:48I actually also highlighted this company, Dark Visitors, which was started by a meta engineer, and then he quit six months ago to sort of work on this full time. And it's sort of more analytics for finding out what all the bots are. Who's the best company? Who do people turn to most? You know, some publishers kind of did flag Datadome. Datadome and other company, Human, was flagged in a Forrester report as, like, the leader in this bot management category because they're more specialist. But, you know, other publishers like Cloudflare, too, and Cloudflare's kind of, you know, been very public about this.

5:19They made an announcement over the summer that, like, they were going to help publishers block the bots. You know, is it a big opportunity? You also asked, like, I think for publishers, they've said that this is, like, kind of like a thing that maybe their IT guys used to only think about. Like, you know, and it wasn't that big a deal. But now, like, the salespeople are, like, thinking about it. Now it's, like, a commercial issue. And, you know, one publisher told me, this is even more important than, like, our ad tech vendors, these vendors, which is kind of a big thing because ads are how a lot of publishers make money.

5:49So it is increasingly important. But is it a big business? These companies are growing. And they've talked about, you know, the companies I talked to, they said they've been getting business from there. But, I mean, the AI is an existential threat to their customers. So it's sort of like you have customers who are using you, but can they survive long enough to become paying customers is kind of a question I had, so. Well, and so let me ask you a slightly different question, which is that one of the fundamental issues is the idea of will websites get these crawlers, these companies operating these bots, will they get them to pay for the content in the long run?

6:30and that's the hope at least. Does this reporting that you've done, does it help us answer that question at all? Or is it really just highlighting how difficult it's going to be to get those companies to pay for the data at all? I think, you know, I think part of the effort to do this blocking was to be able to quantify how much the AI companies needed the content for publishers. So if I could, you know, go to OpenAI, if I'm a publisher that doesn't have a deal with OpenAI, Or even if I am a publisher that has to deal with open AI, but I want to renegotiate it. Because a lot of these deals, I reported this a couple months ago, they weren't really, the existing deals between AI companies and publishers weren't really based on a whole lot of math and a whole lot of like real like business considerations because no one really knew what this would, what this economy would look like.

7:22So they're kind of just flat numbers they could come up with. So the goal with a lot of this blocking is, okay, if we can quantify how much the AI companies need us by showing how often their bots come to our sites, we can start using that as a term in deals and, you know, saying. We actually have a number then. We have leverage. We have leverage. And there's been kind of models suggested by Cloud for paper crawl or paper query. So I think that the problem is if publishers can't quantify it, it makes it hard to negotiate. Now, you know, I talk about Grok, like, I don't think we know of Grok having any deal with publishers.

8:03And so maybe, you know, like, OpenAI is being maybe among the best sort of corporate citizens in this whole endeavor. And they're also the ones doing the licensing deal. So maybe the companies that like, aren't showing up, you weren't going to get a deal with anyway, because they weren't going to be willing. But yeah, I think the bigger question is like, are we going to get to a world where there are standards and that's just like, you know, kind of the way business is done? And then, you know, all AI companies to be legitimate are going to have to, you know, show up and say, hey, I'm a bot and let publishers know and give permission.

8:36Or it's also very likely that it'll just become a free for all and publishers won't be able to control it. Great. Well, Catherine, I want to thank you for coming on. That's a great conversation as always. And I look forward to talking with you again soon. Okay, great. Thanks, Akash. C3AI reported quarterly results last night. The company's revenue shrank about 20 % compared to a year ago. It was the second straight quarter of contraction. This time last year, the company's top line was growing nearly 30%. It all sets the stage for the company's new CEO, Stephen Ehikian, who just took that position in September.

9:11And I want to bring him on to help us understand how he is going to re-energize the company. Stephen, welcome to TITV. It's great to have you here. Thank you, Akash. Nice to be here. So I want to segment this discussion into two parts. I want to talk about the problems that you've identified as you've studied the company over the course of the last 90 days. And then I want to turn to talking about the solutions that you plan to implement. You know, last night on the call, you were asked about why growth or why contraction had really persisted for these two straight quarters compared to this time last year.

9:42And you mentioned sales execution as one of the fundamental reasons. Unpack that for a little bit. What did you mean by sales execution? Yeah. So when you have a business that's led by Tom Siebel, who's the CEO of this company, and he had health issues earlier this year that he disclosed, that is a hard knock against any organization when you take a leader like that and sidelined. He was the heart and soul of this company. So when he was sidelined, it really did slow down the execution across sales. What does it actually mean? Our business is about landing what we call initial production deployments, IPDs.

10:18These are initial use cases with customers. These are high-value use cases across core operations. When we deliver real economic value, those convert and grow. Unfortunately, in Q1, we didn't have the strong leadership to drive accountability to actually ensure the execution of those IPDs. So I'd say that was the number one reason for the mess on that. Okay. And I just want to get a little more granular here for those of us who haven't been on sales teams and don't really understand what it's like when something like that happens in an organization. I mean, is this direction on how to position the product that the teams weren't getting?

10:54Was it issues around pricing?

10:59It helped us understand really on the ground what that looked like. You know, so enterprise sales is not a quick turnkey solution. We have a little widget and you just do high, high volume. This is a highly relationship-based execution where you have a complex use case, and you really have to drive the execution with our internal teams and our customers. These are use cases from, like, supply chain optimization to asset performance to demand forecasting. I highlight the execution is so critical here because when we close an account, we have an alignment with the customer of the value we want to deliver.

11:35If we can't deliver that in a timely basis, those companies don't convert. And that requires the full scale and effort of our company to go help deliver that. So when you lose a leader, your CEO, and he's sidelined, you lose a lot. It's kind of like a rudderless organization for a while. And so for anybody who hasn't run enterprise sales before, that's a huge knock against an organization. So that's basically So basically not knowing which accounts to prioritize, what direction really you really want to take the growth, stuff like that. It's across every dimension, every part of the organization that touches.

12:09So how do you get started? Which target? How to drive velocity? It all is from the top. Right. I do want to talk about the solutions that you're planning here, but I wanted to talk about one customer segment that caught my eye. Manufacturing as a share of your business declined considerably. I wondered why that segment in particular. What I can tell you going forward is that that segment is actually a fast-growing city right now for us. And when you think of discrete manufacturing, I look at the work we're doing to deliver the nuclear-class submarine, the Columbia class. This is all about reindustrializing America, specifically the maritime industrial base.

12:48We're seeing big growth in that segment, as well as semiconductor. I kind of mentioned AMD is a customer of ours from Q2. This is, again, focusing on discrete manufacturing, and that is an area of focus for us. So what's the plan from here now for you, Stephen? How do you plan to bring this company back to growth and ultimately hopefully becoming one of the booming names in AI that we've heard so much about? Yeah, so Akash, I've spent countless meetings with customers, prospects, and partners. And I'll say this, the observation is that opportunity for C3 is bigger than I could have anticipated, primarily because the enterprise AI market is growing so fast.

13:30And I think we have a truly differentiated product here where we are not just a chatbot, we are driving core operations for large manufacturing companies, energy, healthcare, and federal and state and local. So my focus is on one execution. I'm in the weeds day to day with our sales team, driving these execution and ultimately driving economic value for our customers. If I do that, you will see the conversion. And if we focus our product, we have an amazing platform, 16 years of development, plenty of capabilities, where we focus on the right segments. I mentioned federal, state and local, energy, manufacturer, healthcare, and select others, we win.

14:13So there's two components, driving execution day to day, but also focusing our engine on the right markets and market focus. And how do you square that then with the reality that enterprises, as you said, which are a big customer group for every enterprise software company right now is getting them to buy AI? It's slow. I mean, these enterprises are saying, look, I got to find the right people. I got to find the budget. I got to find the ROI. You know, how do you square that with the fact we even saw numbers this week that AI adoption in some cases for the largest companies is starting to plateau?

14:49That's not what I'm saying. And honestly, the focus for us on these industries, like industrial industries, again, manufacturing, healthcare, energy, they're seeing strong macro tailwinds. And the problem we're solving is eliminating a lot of the drudgery from day to day of the work on the ground. Think of this, again, this is not a horizontal use case. These are very discrete applications for supply chains, for asset performance, to minimize downtime. Think of the application across the Air Force to make sure you have the maximum number of fleet in the air to Shell, to want to make sure oil production is on schedule.

15:25Demand forecasting, semiconductors to vaccines. These are what we're influencing. So from my perspective, and this is where we're finding partnerships with these hyperscalers and system integrators, we are able to apply these vertical technology on top of their horizontal platforms to go to market. I want to ask you a slightly more personal question, which is you've been in the role now for 90 days. I wonder in the 90 days before you took the role, and I don't know when you knew that you were going to be CEO, presumably it was at some point in that period. But how do you prepare for coming into a turnaround role like this?

16:02Who are you talking to? Are you studying playbooks of companies that have gone through this? How did you prepare? Well, two areas would probably help me. One, I was a customer of C3 first. So before C3, I was running the GSA, the General Service Administration. And I was in the process of making the government more efficient and evaluating AI solutions to help us do so. And I saw the full swath of the market, came across C3, and I saw the demo. I was like, that jumped off the page. I'm like, what is that? And so that was my introduction. I spent a lot of time evaluating the technology as a customer.

16:36So that was my first. Second is I've built AI companies in the past. two companies. I was fortunate enough for Salesforce to acquire us. So I understand where the technology is going and the capability. I'm sorry, I don't mean to interrupt you. I'm asking a little more philosophical, like turnarounds. They're really hard, right? And there's only been a couple of big turnarounds in history, at least in technology. I wonder, did you study any of these companies? Did you talk to any leaders who had sort of architected these playbooks? Yes, I've spent a lot of time. One, I got to know Tom Siebel extremely well, got to know the boards of the existing strategy, and also knew where we needed to get to.

17:16I also have, from my background, mentors in this space that have done turnarounds of large enterprise companies. So I spent a lot of time with them. And they've been helpful in my previous past building enterprise companies in Silicon Valley. So yeah, study playbooks, understood it. But most importantly, every company is unique. And the alignment with the executive chairman, Tom Siebel, and the board was where I spent the most amount of time understanding their strengths, where they have to deserve the right to win, which industries, which products. And that allowed me, at least gave me an informed opinion of where to focus my first 90 days.

17:49Right. Well, Stephen, thank you so much for coming on the show. We really appreciate it. And we look forward to seeing how you architect your playbook in the months to come. Thank you, Akash. I appreciate it. Okay. It has been a busy week of news with OpenAI. And one story that has been a little overlooked is my colleague Stephanie Palazzolo is reporting that the company is actually working on a new model that should hopefully pose a renewed challenge to Google's Gemini 3 model, which has been getting a lot of buzz. I want to bring on Stephanie to help us understand how Sam Altman is thinking about the new model.

18:23Steph, welcome back to the show. It's great to have you here. Thanks for having me. It's a bit of a busy week for you, isn't it? Gosh, you're in San Diego. You're at the conference. You got the code red. You got the new model. I mean. Yes, definitely a little sleep deprived, but enjoying it all. Okay, well, thank you so much for being here. We really appreciate it. Let's talk about the new model, which is called Garlic. Okay, and we'll get to maybe the code names later on. But what's going to be different about Garlic? Yeah, no, definitely some interesting code names from OpenAI. But essentially what Garlic is, it's a new pre-trained model that OpenAI executives have I've been saying, performs as well as Google's Gemini 3 and Anthropics Opus 4.5 models in areas like coding and reasoning, which of course are very important to all these companies.

19:11And I think the reason why this is so interesting is because pre-training is an area that OpenAI has historically struggled with. And to give some context, pre-training is basically the first step of model training, where you show a model just tons and tons of data from the web, from books, from other sources to teach it to make connections between all of them. So, you know, while OpenAI has kind of struggled in this in the past, Google has done really well on pre-training. And the reason why that's so important is because a lot of researchers believe that when you improve pre-training, you also improve a model's ability to come up with new discoveries beyond what it's been trained on.

19:50And so that's kind of the holy grail of all these AI companies. You want the models to come up with cures for cancer or new ways to do AI research. And so that's kind of this holy grail that all the AI labs are working towards. How quickly do we think this model is going to come out? So we don't know the exact timing, but it does seem like OpenAI is moving very quickly to release its next couple models, especially given the kind of positive reaction to Google's Gemini 3. And so I think we'll definitely see some model that's based on garlic by early next year and the next couple months or so. So definitely keep a lookout for that one.

20:31I have to say, we saw, I think it was GPT-5 that had the big sort of splashy keynote and everything like that. And we obviously know the story has changed a lot since then in many ways. I almost wonder if the keynotes and the releases, everything's going to be toned down a little bit now that the model wars are so competitive, honestly. Why make a huge deal out of 6, 7, 8, 9 when a month later, somebody else could release something and totally blow you out of the water? Yeah, I think OpenAI has definitely learned its lesson from GBT6. So I don't think we should expect to see huge releases or GBT6, GBT7 anytime soon.

21:13They go maybe 5.2, 5.3, 5.5. So expect more kind of like incremental updates versus a big splashy launch. I want to talk about another topic that you wrote about this week in your column. You wrote about something called the split brain problem. Yeah, I was going to say, trying to make sure I read it right. The split brain problem, what is that? And why is OpenAI struggling with it? Yeah, so bear with me because it's going to get a little technical, but I will try to break it down. So this is essentially a problem that researchers at OpenAI and most likely the other labs have kind of noticed where basically just the way that you phrase a question to a model can lead to extremely different variations on the way that it answers.

21:58And this is basically all because of the way that models are trained. And so essentially researchers give models data to help improve the model's behavior in very specific areas like reasoning or math or science. And so, for instance, like a researcher might give a model kind of formal math proofs to teach it how to better answer math questions more accurately. And then separately, maybe the researcher gives the model just more general kind of question-answer pairs to teach it to respond like politely and friendly and in ways that humans like. So using lots of emojis and bullet points, for instance, to like break down a really long answer.

22:40And so in the future, whenever the model is, you know, powering ChatGBT, for instance, it might respond differently based on which scenario it believes that it's in. This kind of first scenario where maybe it's answering very formal math proofs, or the second scenario where it's answering more general questions and where its focus is more on being friendly and the tone and the formatting of its answer. so if somebody you know phrases a math question like in the style of a formal proof maybe the math maybe the model will answer correctly but if you phrase it more like casually the model might accidentally think it's in the scenario that it gets rewarded for like having a really friendly nice formatted answer and so it might give you like an incorrect answer to the math problem but maybe it's like formatted super nicely and it has a bunch of emojis and things like that which is obviously not what you want.

23:29So, I mean, at its core, this is sort of, it's coming back to the issue that it depends a lot what you put into the model in terms of how good a response you get out. I wonder if the granularity of this problem that you've identified, does that reflect how far the models have come? Or does it, on the flip side, reflect how much still they have to improve? Where does this problem sit in that stack? I think it's more the second. I mean, it's really interesting because, you know, on one hand, you could argue that humans get tripped up on, you know, on the way that people word questions to you. Like, maybe I give you a tricky wording to a question and you might get it wrong.

24:09But you kind of expect these models, which are, you know, supposedly going to be superhuman soon, they're going to, like, find the cure for cancer. You would kind of expect them to not have those same issues. Yeah, they should be able to interpret what you're saying, like, you know, catch my drift, man. But even like, I mean, even for humans, like you can imagine like a math professor would not answer a math question incorrectly if I gave it to them in a formal math proof or something versus if I asked them very casually in a conversation. So it's just very kind of interesting, but like very weird that models are still having these somewhat, you know, basic problems.

24:45Right. Last question for you. You are at a conference this week and there has been a lot of buzz about the Code Red story that you published. I wonder what the chatter is on the ground, not just about the Code Red that OpenAI has issued, but also on the ground more broadly at this event. Yeah. So this week I'm at NeurIPS in San Diego, and it's been super fun. Loved meeting all the researchers here. There's like a million people at this conference. I think for Code Red, yeah, definitely has been coming up in some conversations. I do think that on one hand, all the researchers talked to each other.