536. Weekly Pulse Check of Automation & AI News - 10-30-2026

Speaker A: Foreign. Welcome to Transform now brought to you by ssnc, where we explore the future of work through the lens of AI across industries. Leaders are rethinking strategy, culture and operations to harness AI's full potential. This podcast gives you the insights and ideas to transform your organization, stay competitive and and thrive in a world driven by change. Let's get into the show.

Speaker B: Hello everyone. Welcome to the Transform now podcast. I'm your host, Michael Marchuk. And I am Brad Hairston. And this is our weekly pulse check of AI and automation news. As you know, every week we go through a bunch of articles. This week we got uh, a doozy of an article, uh, that covers an industry moving kind of thing related to an ipo. We'll start off with that one because I suspect it will be a lot of discussion. This is from NBC News. The article is entitled Anthropic reveals $80 billion loss. Oh my goodness. Anthropic reveals 80 billion dollar loss. Huge spending spree ahead of IPO. Um, this article covers some of the information that was posted in the Anthropics SEC release pre, uh, ipo. And uh. Brad, have you read the IPO statement?

Speaker C: I have not. I must admit I have not.

Speaker B: It, I went through it. It's very interesting but I'd like to get your thoughts on the materials that were covered as part of this article and then I'll go through some of the stuff that I found inside of the ah, the IPO statement.

Speaker C: Yeah. And Michael, I know you're just ready to make a huge investment and buy some stock as soon as it's.

Speaker B: Especially after, after reading this here. Um, I have rethought all of my, my investment concepts. And you'll know why after I give you that my, my view on this. But go ahead.

Speaker C: Yeah, the headline numbers are wild for this one. Not many companies prepare for an IPO by admitting to an 81 gr billion loss. That's, that's not typically what the marketing people tell you, you know, to bring forward and to try to drive interest in your stock, but.

Speaker B: Very true.

Speaker C: But this one, this one, and I know you're thinking this way too. This, this has connotations around is enterprise AI and, and you know, these journey, the, these uh, uh, these incredibly large, large language models are, is it paying off? Right? Is the, is the uh, their space in the market going to continue to deliver, um, you know, incredible revenue growth, um, but also edge toward profitability? That's what everybody's wondering because that's a very large loss that they're admitting to so let me, let me read some of the numbers to talk about the rev. This is highlighted in the article. But the revenue growth has been staggering for anthropic. So in 2023 they were at 41 million. Then in 2024 they escalated to 386 million. 2025 they went to 4.5 billion. And now in the first half of 2026, not even the full year, they're at 16.3 billion. So this, that alone, I don't know that we've seen a company that's had that kind of revenue growth at that pace. That alone is unique. I do think some of the numbers behind their customer base are pretty interesting as well and which you could say are good. They're showing that they're rooted in the right kind of companies. They have more than 11,900 enterprise customers with 100,000 or more in revenue. They have 1900 customers that are generating more than a million dollars a year for them and then 250 that are generating more than $10 million a year in revenue for them. So that, you know, if you just stop right there, everything looks really good. But then you get back to that $80 billion loss, uh, which is in the headline. And you can't, you can't just write that off.

Speaker B: No.

Speaker C: However, there, there is some interesting angles to this. So of the $81 billion first half net loss, 72.4 billion of it was a, a fair value change on uh, convertible notes that they had taken. And, and that those notes, you know, that debt got more expensive as their valuation climbed. So it wasn't that, you know, they took out that much debt, it just, it just increased over time. And then they also had a, uh, huge, um, booking of 974 million at operating income in the second quarter. So now if you boil all that down, and this is, this to me was one of the more interesting facts. They now currently generate roughly $2 in revenue for every $1 spent on compute. And that's a pretty big jump from this, you know, early 2025. So you can, you know, you can dissect the numbers million different ways. And certainly Dario Amade is doing that to the market and to the potential investor community in anthropic. Um, but bottom line is this is a consumption based business. 83% of their revenue is consumption based. And the net revenue retention was Approximately as of September 30th was approximately 900%. Meaning everybody that came on board with Anthropic and started using it, uh, whenever that was today they're basically paying about nine times as much as they were when they first started. So everyone's either their token bill has just, you know, increased their activity or they're using more features you, they've upgraded to more powerful model. Whatever it is, their base customers now are giving them 9x the amount of revenue that they normally did. So I think this is a, uh, fascinating story and there's going to be so many Harvard Business School case studies built on this, and this may be something that determines the success of future, um, you know, companies like Anthropic, but they're so unique and they've got so much attention, they've got so much, they've got such a strong customer base. People that love Anthropic love Anthropic. I'm one of them. I admit. I'm a big Anthropic user. I really, I use it more than any other model. Bar.

Speaker B: Are you one of those top 1% that uses more than a million dollars a year?

Speaker C: Uh, no, slightly below that. Yeah, slightly below that. But, yeah, I think, uh, you know, this, this IPO is setting the stage for a very interesting time. And, and I'm sure they're going to get a lot of investment despite this huge loss that they're still, uh, encountering. And it gets bigger and bigger. But I think people buy into the pitch, they buy into the value and they say, yeah, this is something that's got some merit and it's going to be, it's going to edge toward profitability, hopefully faster than, uh, than they think. And uh, yeah, I, I, I really don't feel like it's going to be a failed ipo. I just think it's going to be one of the more interesting ones in the history of the stock market. And I, I just find it fascinating the way that, uh, Anthropic has gotten to this point and now they're, they're unabashedly talking about an $80 billion loss because they've got a lot of other positives and they've got a lot of customers and a lot of people that are so crazy about them that they're ready to take the market by, by storm.

Speaker B: And they are sitting on, um, 112 billion in cash. So they right now have, they've got money in the, in their pockets.

Speaker C: Yep.

Speaker B: But let me tell you what I found. Scouring through their actual filing, the sec.

Speaker C: Yeah.

Speaker B: So there's so many different things. Um, first one is, uh, so based on the public offering, public shareholders won't, um, have any uh, meaningful control over this because the Long Term Benefit Trust holds a special class of shares with 50.1% of the company's voting power. So there we go. Then there's class A shareholders which will get one vote per share. Kind of normal, but obviously even if one other person bought all of the rest of shares, that voting block is still not enough to overcome the voting of the shares from the Long Term Benefits Trust. Okay. Kind of different, um, but understood, you know, people say, okay, maybe we can go in with that. Second one is there are multiple classes of shares that are being offered and this is kind of normal that there'll be class A shares and various other types. There is a class of share, uh, that is being offered that um, will have, quote, uh, the ability to be redeemable by Anthropic for no consideration upon occurrence of certain events. Meaning those shares can be redeemed or pulled back for no compensation. So how do you value something? Would you buy something that in fact if whatever the event occurs means, um, your shares are worthless? I don't know. That's a risk. Um, I'll get to this. I'll keep going on this part of it. So part of their viewpoint on this is um, Anthropic is really banking on cheap compute. That's their, that is like their biggest thing right now because that's their, one of their largest, um, expenditures. In fact they said they cost of renting servers and ships jumped over 300% over the period of time since they started in 2023-2026. So it, that's a huge number and it's, it's obviously going to continue to grow because they're continuing to consume more and more as we all consume more and more.

Speaker C: Right.

Speaker B: Um, they have mentioned, as you said, while they have strong revenue numbers, they have a few number of customers that are really contributing to those numbers. So as you know, the fewer customers you have, the more at risk you have. For one of those customers, leaving one of the articles we have coming up a little bit here, we'll be talking a little bit more about this. But that is a risk for them. Another one is, which I thought was, I don't know, let's call it shocking, but it's pretty darn, um, questionable is that they're telling investors openly that shareholder value is not their only objective. So if you compare that in place with the fact that I may have a share that is of a particular class that could be redeemed for no money, that makes me kind of concerned about my Investment in the organization. Um, the other part that was extremely interesting and I'll combine it with another comment they made and I'll quote it directly. They obviously openly warned that um, its AI, as most of them have, could act in its own harmful way, unintended ways. And we've seen this. How many stories have we covered now about various models breaking out of their labs and going and doing something bad? Right.

Speaker C: Too many.

Speaker B: Too many. Right. But um, one of the quotes they have in there, and I will read it verbatim because I think it's very interesting, they say, quote, in addition, with the scale of harm that could result from a catastrophic AI related incident, no single insurer, and potentially the insurance industry, uh, in aggregate may have sufficient capacity or expertise to underwrite insurance commensurate with such exposure. So the same, it's not that if our insurance company went under, but like every insurance company, because there wouldn't be enough insurance to cover the damage that could be potentially done by their AI. I thought that was pretty bold for them saying that if their filing that not only we can, we have dangerous AI, we may not give your money back. We actually don't care about profits specifically. And in our, we may bankrupt an entire industry that's outside of our control in case something goes wrong. All those things are pretty shocking, honestly. But interesting enough that they, they've brought them up, up front. So you have to, for that

Speaker C: they're clearly not betting on people like you that actually read the IPO release. So. Yeah, I'm

Speaker B: sure.

Speaker C: I think that's fascinating. Um, yeah, those are, those are some, uh, some red flags, I think at um, least just at face value. Um, are they really, you know, gonna experience that kind of, you know, devastating, uh, you know, AI outage or what have you? I mean I, I would, I think most people would say anything's possible given what we, what we've seen in just in 2026. So yeah, um, we'll see, we'll see if this scares off a lot of people.

Speaker B: See, we'll see. But anyway, I, again I will leave links not only to this article, but also to the anthropic SEC filing. So if you are interested in going through it, um, you can take a look for yourself because some of the quotes in there are kind of eye opening. So our next article is from Investment News and it's titled AI Governance Gaps Leave RIAs Exposed to a New Wave of Cyber Threats. Now we've talked about cyber threats and AI for a little bit here now and this one is Very, is much more specifically related to financial crimes that could be experienced by RIAs who are left unprepared. Tell us more about that.

Speaker C: Yeah, this uh, this one follows the same topic or plot that we've, we've been going over, uh, ad nauseam, which is that adoption is outrunning governance in, you know, most sectors. So those that delayed it or said oh, uh, we'll go back and we'll apply governance later, they're finding out that that was a really bad idea. And in, in the wealth management space, as you said, these uh, RIAs are in smaller firms or maybe they're one man, one woman operations. And so you could, you could easily see how they could get in over their head or leverage AI in an unsafe way, or not have the same level of governance and security and what have you at a large financial firm would have. And so, um, that is an area of concern. And so this article talks about a recent study of 68 RIAs that found the governance structure, not the assets under management or the actual technology spin, but the governance structure is the strongest predictor of AI maturity. So those firms that, you know that, that centralized or they, they had, you know, kind of a hybrid governance model, they scored two and a half times higher on average than those who had an undefined approach to, to governance. So governance, uh, is certainly a differentiator and certainly a competitive advantage if you do it the right way. And the article does highlight some pretty scary scenarios. They describe one where the firm's entire AI operations are either shut down or poisoned by cyber attacks to the point where the AI is no longer viable and AI just simply cannot be trusted anymore. Um, so that, that's less about leakage, it's just more about malicious or attacks or compromised AI tools that kind of poison your, your whole environment. And again that,

Speaker B: and you know, the bad, the bad guys are not going to stop. Right. They will

Speaker C: find every

Speaker B: possible way, especially when there's money involved, to be able to get that money away.

Speaker C: Yeah, doubt. And so this is definitely concerning the institutional investors. So they recently took proactive steps where they're requiring RIAs who manage funds for those institutions to materially update their AI policies. So they're, they're definitely concerned about this, this kind of risk and they're wanting these RIAs to uh, prove that they're taking the right steps and they have the right governance. Um, and you know, in the same study or survey that was conducted, 85% of the respondents named AI as their top compliance concern. And that's a 28% jump. 28% percentage point jump. Gosh, I just. That was so redundant. It's a 28% jump from last year. How about that? So that's a big. That's, that's big. I mean there's a, that's a, that's a bad trend. So yeah, governance, uh, is, is here. I mean it, it is, it should be a priority. And as we've talked about from other points of view that all are saying the same thing, if you do it right and do it early, it actually will help you speed up and do things more rapidly with AI versus more slowly, which is what most people think.

Speaker B: Right.

Speaker C: Um, so, uh, yeah, here's, here's to another audience, another sector calling out for governance. And in this case it makes a lot of sense given the, the people that are the main players here.

Speaker B: I agree, I agree. As we talked about before, governance is the way that organizations will ensure the success of their program. So that's a good thing. Well, our next two articles are sort of antithesis of our first article with Anthropic. So um, as you might imagine, anthropic has, and OpenAI and some of the other like Meta, larger frontier models, um, have a very big influence in how the public perceives where AI is power, API's powers at. And organizations over the past few years have been using that. However, things are changing now with these open weight models. So from Constellation magazine, uh, we've got an interesting article called Open Models Gain Traction in Enterprises. Tech CEOs say, um, this article and the next article say very similar things but in a slightly different way with some real numbers attached to them. So the idea behind it is what though? Why are they gaining so much traction?

Speaker C: Yeah, they are. They are gaining traction. And so much so that as this article highlights, You've got the CEOs of IBM and SAP and Meta all saying basically the same thing, which is the loyalty to a single model is gone. It's just no longer going to be the norm. Which is ironic timing for this article to come out right when anthropic trying to get people to invest in their stock.

Speaker B: Exactly.

Speaker C: Um, so our IBM, you know, came forward, uh, Arvind Krishna, uh, from IBM, saying that um, open models are now 90% as effective while running at anywhere from 25 to 10% of the token cost of the frontier models. Um, so that's. And you know, he dates that closing of the gap between the large and the small models to late last year. He feels like there was kind um, of a big Pivot that occurred at that point. Um, SAP, uh, they talk about how they're checking each agent use case to make sure that it's on the best model on, on both outcome and price. And so they are switching models more often now and they see this open weight, um, surge these open weight models, uh, as the next great thing. And so they see their, you know, their internal usage of those is going to be even higher uh, going forward. And then you have Constellation, uh, Research, the firm behind this publication and their uh, editor in chief, um, you know, he, he describes this funny situation which is becoming more normal where a CEO is tracking a Google sheet of dozens of models, um, you know, looking at, okay, how do I optimize the cost here? What if one LLM fades off? Can it be replaced with another? And what do I do? And uh, so, you know, there, it's a, it's a funny um, scenario to think about. Um, but yeah, you know, and then Zuckerberg, um, chimes in from Meta and says that he expects to deliver today's frontier intelligence, you know, from all the models they have today at a fraction of the cost in a smaller model a year from today. So this is uh, this is interesting uh, to think about. And it does fly in the face of the big pitch from Anthropic and you know, trying to get people to not only, you know, all their customers that love them now, but you know, stay in love with us because we are, we are here to stay. And yet you've got this other huge headwind that's facing them with the small language models and the frontier, the, the uh, open weight models and all of that. And these are reputable companies that are coming out and saying this. They're not like little guys that are, you know, the ones that open models are all they can afford. So Michael, I guess to me one giant takeaway from this, which is going to sound a little bit self serving, but it seems like if the frontier models, their loyalty is gone, uh, and people are going to be switching and experimenting and trying all kinds of things. It elevates the importance of the orchestration layer that you have that is helping you continue to automate processes end to end and use the right tools and the right technology, knowing that those tools are going to change over time. And even what you give to an agent that needs Claude to make a very robust decision tomorrow you might be giving it to a smaller model, uh, or to an agent leveraging a smaller model because you don't need the firepower of Claude for that decision. And so there's so many ramifications like that. And so having the right orchestration layer sitting on top of all your IT stack, all your applications, all your tools and technology is critically important and it's going to become even more important. So I would just highlight that shout out to SSNC Work hq. Um, but what do you think? What are your thoughts on what these big wigs, these big companies are having to say about the uh, changing market and the lack of loyalty?

Speaker B: Well, you know, I have no thoughts. No, you, this is, this is something that I have been long awaiting, which is the transition from singular monolithic, um, decision points saying, we are going to use this model for all of our work to we will use the right model for all of our work, the right resource. I think that's where you talk about orchestration just in general, um, when work comes through and we've seen this, in fact I've seen many memes on uh, LinkedIn and scrolling through it saying yeah, actually it's cheaper to have this done by a person than to have AI do it, um, when you just take things into account. So identifying that is going to be something that will require a smart orchestration layer. So I envision in the not too distant future this orchestration layer or some part of it, allowing us all to be able to validate that type of cost structure where perhaps even you send it to the same work to multiple models and identify which one comes up with the best answer best being subjective or potentially objective. Um, and then selecting that model, um, for the next set of large number of scale of transactions that are runs that would happen for that particular workflow and then reevaluating that periodically with the same models or new models as you come out to say, did we still make the right choice? Is that still the correct choice? Or perhaps we should look at some other choice based on um, previous runs and whatnot. So um, there's all kinds of things you can do now that perhaps you wouldn't have done with a human model. For example, I wouldn't say, okay, Brad, Michael, I want you both to work on this report and submit it to me. And then we would do the work and we would send it back in and have someone evaluate the results, say okay, well it looks like this month Brad did the, did a better job. So Michael, we're just going to throw yours away and then the next month come up to the same question again, say okay, what you two to do the things again. And this time maybe Michael's wins and Brad's gets thrown away because that's a lot of double work for us to do from a human perspective. And we would feel, I would say, discouraged to say, well you're just throwing my work away. When in fact, if you run a similar thing through M, a set of models, it may cost you $3 with one model and 8 cents with another model to do that work to be thrown away to validate, in fact, that we're still making the right choice on the other 10,000 runs that we're having that go through this particular um, model. So we have this capability now to say M. Maybe it's not just an A and B choice, maybe it's A, A, B, C, D, E and F choice where we send it out to a number of models at once and come bring back and have a third model or maybe a person evaluate those kind of um, outcomes to really say that this is the right model we're going to go for. And that can change dynamically over time, um, depending on the complexity or what for that particular set of transactions. So I'm really looking forward to seeing that kind of dynamic workflow come into play. Um, because I really think that's going to be something that organizations will have to contend with and it will make your, your overall viewpoint of workflow is a lot more um, unstructured in terms of how you think of it. Instead of saying this is our workflow, we always do A, B, C and D in this order. Perhaps you say, well, maybe we're going to mix things up because it is less expensive to go through five different other steps because those steps are so inexpensive for us to run through another model. Uh, so change is good.

Speaker C: Hm.

Speaker B: And I think we need to be expecting a lot more change with the way we work over time. So anyway, that's it. That's pretty interesting. Now this next article goes into it in a different perspective, kind of diving down into how one company specifically handled this particular problem. So on the ThoughtWorks blog they're um, they say renting the frontier, owning the core ip. And from their perspective they threw some actual numbers into what they were seeing and then what they were seeing after they rethought that whole concept of what to use in terms of a model. So can you walk us through that?

Speaker C: Absolutely, yeah. So ThoughtWorks is a global technology, uh, consultancy and you know, this is uh, definitely um, you know, speaks to the value of open weight models. Um, but even more than that, this article is interesting because, you know, here's a consulting firm that's doing A lot of AI work on behalf of customers and, and using AI themselves. And you know, they, they, they, they did the work on their own AI bill with their own data. And then they, they, in this, you know, what this article features or this blog features is they published the numbers. So, uh, there's a lot, there's a lot of interesting insights here. So By September of 26, ThoughtWorks was spending about $450,000 a month on Frontier model, uh, APIs. And that was for internal use alone. And the bill was growing about 27% a month. So doing nothing, they forecasted that they would be at 16 million in 2028, uh, and even up beyond that if the growth doesn't slow. So this was a problem that was already pretty substantial, but it was only going to get worse. And they saw that. On top of that, they felt the same way that many people feel like Satya Nadella from Microsoft, that when you're using a Frontier model, especially the big ones, um, you're, you're helping them by putting your prompts in. You're, you know, the models are learning from how you're using their models. And they weren't crazy about that either because that felt like, you know, hey, we're, our bill's going up, we're spending more and they're just benefiting from us. They're benefiting from how, um, you know, we're prompting the model and using it and everything. So with all that context, here's what they did. They put every workload that they were, um, using with AI, they put it on a grid, uh, that differentiated things. So they found that utilization, not list price, decides the cost. And then they put users on, uh, you know, in different tiers. So they had high, high, medium and light tiers rather than specific models. And then they charge back to the business units through that. So they found that, you know, open weight models. Now as a result of these tiers that they're using, they now carry 34% of the enterprise AI token usage, which is up from 23% a year earlier. Um, so by pairing the wider AI access with the, the chargebacks and then also using kind of a, they have kind of a token literacy training that they put their employees through. They felt like that, um, set them up for success in the future. And so I wonder, you know, if it's this cost transparency situation that is finally going to make people use AI responsibly no matter what business they're in. Or, or is it, is it the fact like ThoughtWorks did do you put a price tag on every prompt. But is that going to kill experimentation and innovation and the way that people across your company are using AI? So those are. Those are the kind of the questions in my head that come up when I think about, uh, this article and what it has to say. So what do you think, Michael?

Speaker B: So I think there's going to be. They can actually lead to more experimentation. And it's not because of using the frontier models, but I think maybe because you'll be using less expensive models to start with and evaluating your results. So perhaps we start using the quote, unquote cheap model. I'll try with that one first and see how it does. If, uh, it does give me, you know, the kind of results I'm expecting, I may tune my prompt to see if it's going to give me better results or if the agent is unable to work through with the, um, the capabilities of that particular LLM, that perhaps maybe I would switch models and go up. So maybe I would have more experimentation that would start at the lowest levels and work their way up. That's a possibility we could see, which would then be less expensive to do experimentation than we have currently. So that's how I might look at this.

Speaker C: Okay.

Speaker B: So anyway, well, it's lots of changes, obviously, with. When we start talking about, uh, things that are happening within the model frontiers and whatnot. But addition to the model frontiers, um, we talk about ip. So this article talks about, um, holding some of your intellectual property close to home and maybe running on local models and whatnot. Um, well, this intellectual property argument, which we've talked about from a copyright standpoint, applies to much more than just that, as we'll see in our next article. Um, so in this case here, the article is called the Molecule came from AI. The inventor still has to be a human. So as we've talked about before, AI is doing a lot more things and creating or discovering things, leveraging the powers and the capabilities that it has. But are they really the inventor? What does the article say?

Speaker C: Well, yeah, the. The author here is making a pretty big case. The human reign supreme in. In these, uh, scenarios, which you and I would both agree with as well. Um, I just want to say before I dive in, I'd love the fact that we're, for the third week in a row, we're covering a science, you know, something that's science. Right.

Speaker B: Hilly's gonna get upset that we're not talking about her. I'm just saying.

Speaker C: And this comes from, uh, C and En, which most people Are going, what is C EN that is the Chemical and Engineering News. So this is one of the world's largest scientific, um, societies that is behind this. And not a publication that you and I are checking out all that often, but

Speaker B: that often.

Speaker C: But in this case, a very profound piece and it really should be looked at. So we, we spent a lot of time on this episode talking about who pays for AI and who governs it. This one is, is really asking the question, who gets credit for what it produces? And this author, uh, has a really, I think, profound perspective. She is the chair of this large scientific committee that we're talking about, and she's the chair on intellectual property. So, you know, she, she notes that, you know, really, AI is just a tool. It's, it's no different from a spectrometer or, you know, some other type of scientific equipment that you use. It's a, It's a tool that's being used by the scientist. And the, the inventorship is judged the same way whether AI was used or not. So she feels pretty strongly about that. Her biggest point in this article, I believe, is that writing a prompt is not inventing, it's applying judgment. That's where the human, the humanity comes into the scientific process and what you're doing and how you're using the AI. And she lists out several examples where the human judgment is really instrumental in using AI in a scientific way. Um, and that is examples like recognizing which output actually solves the problem. Because you'll get a lot of output from AI Human judgment, intuition, experience is going to help decide which one solves it. Um, modifying the proposed structure to fix a scientific issue, that's again, something that a human would do. Conceiving a whole family of related compounds to solve a problem. Again, the human is going to lean into that. So, and she lists several others. But these are all, I think, examples of human in the loop. But it's redefining it to something much more than just, oh, human in the loop. I got to approve something, and then it moves on. This is human in the loop, bringing decisions, bringing, um, scientific expertise to bear. Um, so, and I think her, I think her part, her uh, point here is, uh, is made. Made in a very impactful way. The most practical part of the piece is where she talks about that a good record should capture, uh, you know, the model that's being used, the version that's being used, the prompts, the outputs, which candidates were selected and why. Um, every, every change to which

Speaker B: gets back to that experience part. Right? You're not, you're not just saying solve this problem sitting back waiting for it to give you the one answer, right?

Speaker C: Yeah, exactly. And you know, the record's purpose is to separate what the machine or the AI is proposing from, you know, where, where it actually goes and what actually gets concluded and how the human's involvement has taken uh, that and, and applied it. So um, yeah, we, we, we almost always justify human in the loop as a risk control. But in this case the, the author is making a great case that it also is what makes the work, um, ownable, um, or you know, you can connect it to a human and where the human was involved. So I think this is really fascinating. I think it's a great point. I don't think it's just something that applies in the scientific realm. I think even in the traditional business setting. I think it also has um, ah, an, you know, it has implications on IP and where we consider ip. Um. But yeah, I love the article and I thought it was a good piece to end on.

Speaker B: Yeah, I agree. Um, according to US patent law, a natural person must be the inventor. So it always is going to still come back to a person. But to your point, on some of the stuff on like the experience and judgment, how many times have we found accidental things that have become huge, for example post it notes, the AI in that case would have said this is not sticky enough for the proposed application and then move on. Whereas someone actually doing it said this actually could be used for something else. It's not solving the problem I asked you to solve, but this is unique and novel enough that I think it could be used for something else. Or how many other things have we seen? The same thing was an accidental discovery that was made that wasn't intended. So I really think that this is uh, an important piece that people are still involved for a reason because we're the ones asking even to solve the problem. And when we find answers that may not necessarily solve the problem, we may find answers that solve a different problem we weren't even asking the question to. So, um, anyway, very interesting. Well, that's all we have for this week's pulse check. As you know, every week we have great guests that can tell us a lot more about what's going on than Brad or I ever could. So shoot. Tune to those episodes and until the next pulse check, we'll see you then.

Speaker A: Thanks for listening to Transform now, the podcast from ssnc. You can find all of our episodes on your favorite podcast channel as well as YouTube to stay on top of the hottest topics in the world of agentic automation. Subscribe now. The future of work is here.