Stripe buys OpenRouter, Ramp’s AI Index & IBM’s OpenAI deal

Introduction

00:01I mean, this this market is ebbing and flowing. And, you know, we just have to trace where the heartbeat is coming from. Right. And it seems like that right now the heartbeat is around routing. All that and more on this week's Mixture of Experts.

00:20I'm Tim Hwang and welcome to Mixture of Experts. Each week Emily brings together a panel of technologists working at the frontiers of artificial intelligence to lead you through the week's news. On this week's episode, we have Kaoutar El Maghraoui, principal research scientist, AI Native Systems, Aaron Baughman, IBM fellow and Mihai Criveti CTO watsonx Orchestrate.

00:39Welcome to U3. We've got a bunch of big stories to follow on. Today we're going to talk a little about Stripe buying OpenRouter. We'll talk about some really interesting business data out of Ramp. And then finally we'll talk a little bit about AI for legislation in Congress.

00:53But first I really want to talk about one big announcement coming out of IBM itself.

IBM OpenAI partnership

01:01Just this last week, there was a big announcement that IBM was going to launch a big partnership with OpenAI. And this is sort of interesting. It caught my eye just because not too long ago, IBM had also closed a very large partnership with Anthropic as well.

01:12And specifically, the partnership pertains to essentially training up large numbers of people to facilitate basically an open AI practice within IBM consulting. And I guess maybe I'll throw it to you first. You know, I'm kind of interested in how this kind of world of consulting is evolving around AI, because normally I think we've thought about like, okay, it's going to be OpenAI versus Anthropic, and it's just going to be, you know, the sharks and the Jets.

01:44But this kind of partnership really almost suggested a different world is going to merge where it's not going to be so black and white, and in fact, that there will be a lot of kind of professionals that work to integrate these systems that are kind of much more ecumenical with these models.

01:58And so curious about your take about this partnership and where you think it all might go. I'm very excited about it. I think it's a potent combination. You know, to have OpenAI's AI capabilities, plus IBM's and our ability to put them within a Fortune 500 company and beyond.

02:11You know, within this new practice that we're creating, we're going to have forward deployed units and forward deployed engineers are going to be equipped with these kind of capabilities that can go out and meet clients where they are, right.

02:23You know, and this is a big signal that we IBM we're not trying to pick a model winner. You know, we're actually working and we're going to become the owner of the enterprise AI control plane here. You know, we're betting that the enterprise AI value, it's shifting upwards from owning the model to now orchestrating, governing and and integrating.

02:49And even you could argue operationalizing, you know, these types of models. And what what I also think is fascinating, right. Is that, you know, both Anthropic and OpenAI, they they somewhat own the market. Right. And and because we're we have a partnership with both.

03:05Both in different ways, it's quite important that we're putting together Anthropic. Right. Where we have Anthropic is really geared towards IBM software engineering. You know, you know where we have IBM Bob built around that ecosystem while OpenAI is really for IBM consulting.

03:18Right. And going out and having these forward deployed units and engineers, you know, so now we're really this neutral enterprise AI integrator where we have the options right, of both of the leaders Anthropic and OpenAI. But yet we still maintain our own, you know, Granite type series models, which I think would become even more important.

03:41Right, because this gives clients the freedom of choice. And the Granite models are the much smaller ones, right, where we can route simple requests to those. So they're smaller, they are cheaper. Whereas you could argue that the OpenAI ones, they would handle the more complex and type of task planning.

04:00Right. And so and so by having, you know, these model gateways, I know Mihai works works on a lot, gives us the ability to go and any direction that our clients would want to go in. Yeah. And I think, you know, I think one way of looking at it is almost like everybody's competing with everybody across every single layer at the moment.

04:25You know, this partnership makes me think a little bit about OpenAI. And not too long ago both launched this own deployment company, which is kind of a consulting arm that would help people integrate OpenAI. They also signed a partnership with McKinsey, which is like another consulting org at the same time.

04:37Aaron, as you mentioned, IBM's got its own models that are kind of playing in the space and also has its own consulting arm, which is now partnership with OpenAI. You know, it kind of feels like, again, it's sort of like everybody is going in every single direction.

04:50I guess the kind of question is like, what's OpenAI's strategy here? Because it feels like they are sort of almost like choosing every single route and sort of seeing what sticks when it comes to this world of working with clients to kind of integrate this technology.

05:02I think it's system integrators, which is boots on the ground. This is where consulting organizations come in and have the ability to really integrate specific models for a specific model providers into use cases, which are company specific, client specific.

05:16It's also important to note that there is no such thing as the best model. The best model for what? Right? Different large language models and models providers offer multiple models with different cost, different latencies in time to first token, different performances in things like tokens per second, different performance as in the outcome.

05:37And even that differs for use cases. Which is the best model for coding? Which is the best model for information retrieval or summarization? Different models model providers have different geographic availability, and in many cases you're only allowed to use as an organization as models in your geography.

05:51If you don't have OpenAI in your London data center, you're going to have to use what's available there. And even there you have limits with these models, which is what most companies tend to find out the moment they scale AI hey, can I just give AI to my 100,000 Like this is really hard.

06:14Like, no, you can't buy it. You literally can't buy it to go to these organizations. And you're like, hey, this thing only does 50, 150 to 100 requests per second, or the inference performance is not sufficient. And no matter how much money you throw at it, you realize there is not sufficient model to go around.

06:33So then you either do model routing or you go to multiple providers and you find out, hey, I need to prompt this model differently. So this is where again, that relationship and partnership with system integrators, with the forward deployed engineers and so on comes in to be able to help you consume the right model, for the right task, in the right way, in a way which is governance and secured and controlled and prompted the right way to get optimal results.

06:57yeah. I wanted to also mention there's a cost factor involved as well. It's very costly. Right? For over 100,000 employees each to begin to use these types of models. And I think to his point, you know, there's no no free lunch, right? Where you have to route the appropriate problem to the appropriate model.

07:19That's like this multi-objective issue, right, where you want to have the best capabilities that match the problem, but you also want to minimize cost, right. Because because if you think it's like ten bucks or $10 US dollars per employee and you have hundreds of thousands of employees per month, you're looking at millions of dollars a month.

07:35It costs token costs, which may not be sustainable for a business. Right. And so by having these partnerships, you're not not only get access to some of their leading edge models, but you presumably also get, you know, economies of scale pricing, right, which does help to bring down the cost of your workforce so that you can equip your forward deployed engineers and forward deployed units.

07:58Right. So I do think it's a competitive advantage for both IBM and also for for OpenAI to saturate the market with their capabilities. Yeah, I think there's another also point that's very important, which is kind of the distribution and the legacy plumbing that we have, you know, at IBM, which is very strong.

08:20And I think those be true benchmark edge. If you look at what's happening in the Silicon Valley, there's a lot of belief that maybe having a 2% higher score on the MMLU or the HumanEval benchmark wins the Fortune 500. But in the real world, 80% of enterprise value is locked inside legacy mainframe COBOL databases.

08:40Maybe SAP, ERPs Epic health care systems, HIPAA compliance vaults, etc. and IBM owns those enterprise relationships. OpenAI and Anthropic needs IBM's distribution and consulting pipes far more than IBM needs their specific model checkpoints.

08:54So I think here this is really a great partnership because, you know. IBM, I don't think we need to spend 20 billion training Frontier Foundation models to compete with maybe Microsoft, Google or Meta. Instead, IBM's winning strategy here is becoming the trusted, vendor neutral enterprise orchestrator.

09:16Through the watsonx Govern, IBM provides the audit trails, the bias detection, the prompt monitoring, and the regulatory compliance layer while letting clients kind of hot swap models based on cost, speed, and capability. I agree with Mihai.

09:31I mean, there is no model that fits everything here, so it depends. The best model depends on so many other other dimensions here. And you have to tune it to your enterprise, to your data, etc.. So I think it's a very important to have that layer, that orchestration and those pipes, as are so important with the regulatory compliance, with the audits, with the governance, which is really important because the bottleneck here in AI is not a shortage of intelligence models.

09:59It's really change management, system integration and also dealing with dirty data. And I think IBM consulting is kind of weaponizing OpenAI and Anthropic models to accelerate that enterprise digital transformation, kind of converting legacy code bases and automating banking workflows at scale.

10:16Yeah, absolutely. And I think the question being asked is it's almost a little bit like what. You know, how much will the model matter in the future. And I think, you know, you've seen this debate. We've talked about this a lot in MoE on the compute side where it's like, okay, well if everybody can get great open source, maybe it's really the compute that's going to matter.

10:35It feels like the consulting side is actually the other part of this discussion, which is okay, when a world where there's lots and lots of models and there's there's lots and lots of options, then ultimately the kind of winning factor is going to be the sort of trusted consultant.

10:48Aaron, you want to get a final thought in before we move to the next topic. Yeah, yeah, I really do like this constellation of partnerships. Right. So there's there's three in my mind right now that really stand out. It's IBM Plus Anthropic plus OpenAI.

11:02Right. Because the Anthropic deal and partnership was announced around October 2025. And the idea was to bring in AI to change enterprise software development and integration. Right. And that's that's how IBM Bob somewhat emerged. Right. And then this OpenAI partnership is different.

11:24You know, that's about equipping and training thousands of consultants through IBM Consultant Advantage so that we can go to clients right, and meet clients. So we're taking the full stack of the business with these two partnerships. And I'm really excited about the future and what we're going to do together within IBM.

11:40Great. Well, that's a great segue to the next topic I want to cover.

Stripe acquires OpenRouter

11:48news broke this week that Stripe, the payment processor, is finalizing an agreement to purchase a company called Open Router for more than $7 billion. And again, if any indication of how crazy the market is right now, the stat that was kind of going around the internet, which is that 7 billion is more than five times the $1.3 billion valuation that OpenRouter had just 82 days ago.

12:13And so this is like a really wild story. And I think for a lot of folks who spend their time focusing on the models and hearing about what OpenAI and Anthropic are doing, OpenRouter itself may be kind of an unfamiliar name. And so I guess for folks who are maybe less familiar with what Open Router is or does, do you want to give a quick intro like why it's such a big deal and why it could be possibly worth $7 So as somebody who's written his own model gateway, MCP gateway, Agentic gateway and gave it away for free, I'm kind of A little bit depressed?

12:51I believe the future is multi-modal, multi-model, multi-agent, multi framework, multi harness, multi-cloud multi everything. There is no one size fits all for models. There is no one model which is good and you can't even get enough of one model to fit every use case.

13:06So with platforms such as OpenRouter, you have the ability to access a large variety of models from every vendor to a single platform, which gives you access not only to the models, but you can do things like, for example, semantic caching, or you can help optimize some of that work.

13:29And some of the future that we will see in this space is integration into unified gateways, gateways that have the ability to access not only models, but MCP servers, agents, skills, evolves controls with semantic routing, the ability to route the right request to the right model, MLOps and cost management, the ability to say, yep, we're going to limit and cap your cost and spend with things like virtual API keys where you're not accessing the key of the model directly

14:02so you can lose it, or somebody can steal it, but you're actually accessing a virtual key, which is tied to role based access control. So a lot of these capabilities are must haves in enterprise. They are the capability which allows an organization to distribute these models MCP servers, tools and so on safely to their enterprise, but also get things like metrics and observability data from those interactions.

14:26I guess. So now we know what OpenRouter is. Why Stripe? This is also kind of like the other part of the transaction. We've been hearing, of course, a lot about the big acquisitions at the Frontier Labs. Do you know Stripe is you know, I use them to pay for subscriptions and stuff online.

14:40Why are they getting into this space, you think? Yeah, that's a very good question. I think the way I see it, models are cheap. The tollbooth is where the money is. Every six months, models get smarter and way cheaper. Token prices have dropped over 90%.

15:01Building a SOTA model is a tough, expensive race to the bottom, but the router gets paid a tiny toll on every single request, no matter who wins the benchmark race. But AI billing isn't traditional software. It's kind of a payment headache.

15:10Old software kind of charged, say, like a $30 a month per user on a credit card. AI doesn't work like that. It's millions of unpredictable micro tokens flying around like 24/7. Stripe already runs online payments, so buying the router lets them handle metering and payments right where the code runs.

15:34And if you look at AI agent, they need their own wallets down the road. AI agents will do the jobs on their own. They can sign contracts or swipe physical cards, and the agent needs a digital wallet to say I'll use a cheap five cent model, for example, for simple lookups and a $0.50 model for deep thinking.

15:47Stripe wants kind of to be the Visa card for these autonomous software. And another thing, Stripe is also it's neutral. The the big clouds aren't, you know, Amazon, Microsoft and Google want you kind of stuck on their servers. Stripe doesn't care where your code runs, whether it's on AWS, on an old laptop or a GPU cluster.

16:11They just, you know, they just take their small cut of the transaction. And, you know, I think the routing strategy allows them to be very flexible here and then kind of have these fallback mechanisms switch between various models so they don't care where things run or what models you use, but they want to control that orchestration routing layer, which is becoming super important here.

16:27And so I guess I bring it back to the original topic. It feels like we're we're back in the same place we were before, right where we were talking about. Well, you know, people don't care about the model so much anymore. It's going to be the trusted consultant.

16:40You know, it seems like, well, maybe people don't care about the model so much anymore. It's going to be the router and the gateway. And I guess kind of question for you, Aaron is like, we're almost looking at a post model market now. It feels like we're a lot of the money is going is the things in and around and outside the AI to like the AI model to wit and I don't know, like how should we think about how that markets evolving, right.

17:03Like is the future OpenRouter buys a consulting firm or is it you know Nvidia gets into the console. You know I'm kind of really interested in kind of all of these weird configurations we're about to see as the core thing we've been paying attention to.

17:16AI almost seems to become kind of commodity, and we're all kind of looking for like, what else is the sort of economic niche that you can stick into? I mean, this this market is ebbing and flowing. And, you know, we just have to trace where the heartbeat is coming from.

17:35Right. And it seems like that right now the heartbeat is around routing. Right. Being able to route traffic to the specific place of which it, you know, optimizes or minimizes your criteria that you're looking for, right. But I suspect that this routing, it's going to become commoditized at some point.

17:52You know, it's a very it's going to become a very crowded capability space. You know, I quickly looked and I mean, there's many competitors to to OpenRouter. There's LiteLLM Cloudflare, Portkey, Helicone together AI fireworks AI so on. Right.

18:05So. it's incredible. Just just the amount of space, you know, that's where all of these little companies I won't say little anymore, but these companies are coming in. And to me Stripe bought or rather. Yeah Stripe bought OpenRouter based on growth.

18:28You know they saw that over the last what you know, six months or so that the token growth by five x from 5 trillion to 25 trillion. Right. So they're betting the farm, I think, on that type of growth that they want to position themselves as the infrastructure for AI agents and e-commerce just like they are today with, you know, just commerce in general.

18:50So they're trying to make that leap, right? Saying that, you know, AI agents and models are going to be the drivers of commerce. Therefore we need to get into it. But there is risk in this transaction. Right. Is this commoditization part so and so do I think that this is going to be the underpinning?

19:05I think it, you know, of the economic drivers and decision space. I think today it is. But you know, you know, after it becomes commoditized there will be something else, you know, that that will come in. So, you know, will consulting groups want to buy, you know, OpenRouter, you know, type-ish or build their own organic today.

19:16Yes. But I think there'll be other things that are going to come up right where they want to focus in their capital. I think they also have a large number of users, which was part of the acquisition. Like it's I don't know the exact numbers, but something different.

19:39Sources claim they have 8 million or 10 million users across free enterprise, whatever else, 150,000 active. I mean, I don't know the exact numbers, but this is not just an open source project somebody buys. It's more of they're buying the users, they're buying the commitment, they're buying the wallet, they're buying those paying customers as well with this transaction.

19:58Yeah. What's what's really interesting is that OpenRouter has like 50 to 100 employees. I mean, it's pretty small, but when you look at their market valuation, I mean, you're looking at maybe 70 million per employee, right? That they contribute towards their valuation, of which they are 70 million per person.

20:17Right. That's that's big. And so, yeah, I mean, to his point, this is a, you know, a bet on growth. But they do have the foundation to make that growth. And if that trend line continues right. Right. You know, I think Stripe might have done a good job, you know.

20:29But again, commoditization is always the enemy of these types of deals. Yeah, I also like maybe to bring here a nice analogy, like the tollbooth analogy. So I think Stripe didn't buy this for just the AI hype, especially the OpenRouter. I think it's like, think of it as they bought the tollbooths, like while model makers kind of bleed cash competing on benchmarks.

20:55I think Stripe sits at the bridge and collects $0.02 on every car driving across. So and I think it's a very interesting economy here because they want to get margin of all of those transactions happening, you know, all of these tokens being consumed.

21:08And it's a very good place where they're sitting. You know, the routing right now is becoming very important place. Well, maybe we had just maybe a final question I had. I know it's a sore topic because you released one of these gateways for free to the public, but what is the open side of this?

21:28The open source side of it. Like, do we feel like, you know, that there will be more kind of open gateways that kind of are operating outside of the businesses? Like, what's that ecosystem look like from your perspective as someone who's built something like Yeah, I think there's already more than 100 of these AI gateways in some shape or form out on the market today in the open source space.

21:43You know, I think you mentioned like 5 or 6 of them, but definitely there is a need, right? Enterprises and users and so on need to have a way to manage a combination of agents, MCP servers and models and so on with things like semantic routing and caching permissions.

22:05And I think a lot of what this space is evolving towards is that of agent management platforms or agent control planes, where AI gateway or model gateway is just a small component of the capabilities enterprises need, whether they're using open source or commercial software to provide governance, trust, observability, optimization, cost control, FinOps, run and manage capabilities for their entire AI fleet.

22:36Really interesting data coming out of Ramp.

Ramp’s August AI Index

22:44Which, as you may know, is one of the most widely adopted, you know, sort of business credit card payment management sort of platforms. And so as a result, they have a really good bit of traction on understanding how businesses down to the very micro level are using things like AI.

23:01And so they released the post, which is basically they're kind of August 2026 review of what's happening in the AI space and some really interesting stats here. I mean, I think the first one that I'm really kind of interested in to build on some of the themes that we've been talking about today, is that there really does seem to be a lot of evidence that open source AI adoption is is continuing to rise over time.

23:21I guess maybe Kaoutar I'll kick it over to you. You know, this is still, you know, a jump from 4.5% to 6.1%. So not a huge amount of the overall market, but they are seeing this kind of really measurable rise in people using model serving inference platforms and not really depending on the kind of proprietary models.

23:38How far do you think that's going to go? Yeah I do enjoy reading the Ramp and looking at the 2026 Ramp AI index because it gives you a nice, you know, sort of pulse right on on what people are doing and what the businesses are doing because it's, you know, it analyzes about 70,000 businesses and what I found.

23:57And just by looking at it, it looks as though AI adoption, it's it's really uneven, right. Because if you look at like the top 1% right of these AI adopters, they spend about, let's say $7,400 per month on AI, you know, per employee, whereas the top 10%, it drops down to $650, right.

24:21And then the median or the number of that site right in the middle of all this is around, what, $12? And this is, you know, whenever I quote monetary values, it's USD. But but this is per month per employee on AI. So that spread really shows you this huge spread right across companies.

24:41But even so there is growth. But there's a ceiling to the growth, right? It's as if companies are saying, hey, let's give everyone access to AI. But we we have to meter the use, right, right, right. We need to cap the use in some way if it's getting out of control, especially as we get these digital workers right online.

25:04Right. And and so there is this, this economic consequence. And what I find interesting, right, is that the top 1%, right. It's starting to approach the cost of AI per employee, started to approach the actual salary of the said engineers that are using these tools.

25:21So so now the question becomes, well, if there's if, if we're paying as much for AI tools as we are a person, right. What do we do. Right. Do we need to focus in on productivity versus revenue. Right. What's what's the most important aspect.

25:34And so businesses I think will eventually have to grapple with that, right. Assuming the economics and costs stay where they are at the moment. I feel kind of the, the end of this AI formal tax. Because what happens in 2023, 2025 CFOs approved this broad AI software budget.

25:57And they're kind of the fear of falling behind. But in 2026, the era of the unmetered experimentation is kind of officially over. And CFOs now demand kind of measurable unit economic payback. So either an AI tool demonstrably kind of reduces the headcount requirements, accelerates kind of cycle times.

26:21For example, PRs merge per engineer or, you know, drives, you know, kind of net new gross revenue. So so I think what Ramp's data is exposing here is the kind of the vulnerability of this ten wrapper startups, the which point solutions that merely summarize kind of PDFs or generate marketing copy.

26:40These are seeing kind of catastrophic catastrophic churn because foundation model updates and native OS browser features do these things for free. So I think the true cost metric here is how do we measure the cost per unit work versus the cost per seat.

26:57So evaluating AI spend on a per seat basis, I think it's kind of an obsolete legacy SaaS framework. And you know, I think the strategy has to shift here to evaluate AI on cost per resolved, maybe support tickets, for example, drop in from $18 human cost to maybe like a few cents AI cost or cost per audited contracts dropping like from say $400 to $12.

27:21So the high spend per employee is actually kind of an indicator. It's not a good indicator if it's if it doesn't correlate with the multifold productivity gains. And I think that's where I think we need to start focusing on it. What their return on investment with these spends per employee in terms of productivity gains.

27:41Otherwise it's it's not going to be useful just just to have these spends. But I also feel things are kind of segmenting initially. It's like okay let's start. Let's give everybody access to these AI tools and and LLMs etc.. But some certainties, like especially the engineering teams, the software developers, they're actually using them substantially versus maybe other teams that might be maybe using them less so.

28:07And I think that maybe we will start probably seeing segmentations for departments, you know, who really is benefiting the most and getting the ROI versus other teams that might probably don't need as much. Yeah for sure. And I think I mean, underlying that is sort of the idea that we actually may be spending too much on AI.

28:31I don't know, like I guess there's there's one way of reading these numbers, which is that the skeptics who say, hey, this might be actually kind of a bubble or maybe actually. Right, right. Because as we start to take a much harder look at, say, productivity per dollar spent, it may end up being much smaller than we think.

28:47Do you think it's like too pessimistic as a way of reading these numbers? I think the way I would look at it is that, you know, over the course of the last two years, most of the client conversations I've had have shifted from help us use AI or, you know, we want to buy your product or we want to do something with AI.

29:01We've done a POC with AI to help. We've done too much AI. We've got 60 random acts of AI. We've got agents written in LangChain, general GPT-3, we've got agents in Microsoft Copilot Studio, agents in watsonx Orchestrate. We've got agents everywhere.

29:15We've got models. We've got every model. We've got model routing. The problem we have is one. We don't even know where these agents are, where they are built, or how AI is being spent, because it's usually not that employee that's, you know, asking ChatGPT for advice on 1 or 2 things.

29:32That's really burning down the tokens. It's these hundreds of agents, the developers and everybody else who are, you know, they have 50 open terminals or they've got an event stream going into one of these agents and consuming them at scale, and they use the most powerful and the most expensive models and so on and so forth.

29:44Second comes regulatory pressure. Regulators are asking, hey, you know, there's a thing called the EU AI Act. So are we factoring into the total cost of ownership, one of those big fines, for example, or the cost of regulating those models?

30:03Because the more you have and the more AI you have spread all over the place, the higher total cost of ownership is not just the token spend of how much you're spending on an individual model provider. It's everything else to manage it the people, the resources, the infrastructure, the governance, the security.

30:16And then you've got the cost runaway. Where today organizations don't really have a good mechanism to manage things like budgeting, even tokens, like understanding how much they're spending, let alone tying that into quotas, enforcement, tying it to a return on investment, for example, and value.

30:38So this is one of the areas I'm most excited about. It's one of the areas where this is kind of what we're building with watsonx Orchestrate. So that's a product. Product I'm I'm technically responsible for to try to get alignment between return on investment, model cost.

30:55AI spend, total cost of ownership guardrails, security controls, everything else in between. So I will say AI is easy and this list shows us is true. It's getting control over AI and having predictable costs and having predictable outcome. And a good return on investment is proving extremely difficult in these early days.

31:22Yeah, yeah, yeah. I found too, that it's becoming harder and harder because as AI is getting cheaper and more capable, you would think that the cost would go down. But just in this report, you know, you can see that just over the last year token, you know, spend has increased 13x.

31:34So so it's like there's this AI splurge going on. You know companies can afford to do more. So let's do more because it costs less. But then the overruns happen because you have all of these agents that are compounding the problem, that we don't know what exactly they're doing.

31:49Right. And how to put the guardrails on. Right. So so it's quite an interesting conundrum, you know, that I think the field is And I think maybe this is kind of similar to what happened with the cloud computing when cloud computing matured between 2012 and 2015, companies stopped mindlessly spinning up idle EC2 instances, and they started introducing FinOps.

32:17So the Ramp data here shows enterprise AI is also entering this FinOps era and really pushing the companies kind of they need to start thinking about how do we spend smarter, not less so because now we're entering this maturity age and now the economics, the reality is hitting us harder.

32:37So I think a mental model here of how do we monitor these things, meter these things properly, make sure that we're spending where things need to be spent properly. So like if you look at engineering teams, it seems that they're the only department where AI spend continues to grow exponentially without pushback.

32:57Tools like Cursor, Claude Code, specialized code review models, etc. they've crossed the threshold from novelty to be an indispensable developer infrastructure. So that's been there. I think it's justifiable, but maybe in other areas not as much.

AI drafting legislation

33:16for the final topic today because we're almost out of time. This is just a fun final item that I wanted to touch on. And I guess maybe, maybe I'll give you the last word on this Politico, which is a publication that covers DC and DC politics, had a really interesting article that the Congressional Office of Legislative Council in the House of Representatives is awash in AI-driven slop, quote unquote, bills.

33:38And it reminds me a little bit of I think we've talked about it before. I think one of my favorite analyzes that anyone has ever done was observing that after ChatGPT launched, the use of certain types of phrases really exploded in the Houses of Parliament in the UK.

33:52And I guess we've all used AI occasionally to take certain shortcuts. And so maybe it's no surprise that we're finding people trying to use AI to draft legislation. Should we be worried about this? What do we do about this sort of thing? And I'll give you the last word here for the episode.

34:13to introduce a load bearing bill with and smart quotes. I think it's no surprise that's just like we're seeing an explosion in papers, publications, homework, emails, Slack messages, Teams messages written by AI. We're seeing the same thing everywhere, including legislation.

34:32Part of it is a question of education. It's educating people that AI will make mistakes, that you need to have governance, security, control. You're still responsible for the end result and giving them a way to do it. That is, I would say secure and governed, not just everyone going to, for instance, ChatGPT or online or somewhere, or, you know, just posting the information there and posting back whatever they see.

35:02So I think legislation is no different than any other field. And rather than trying to fight it, there needs to be clear guidelines and rules on what's allowed, what's not allowed, what's safe, what isn't safe, and empowering folks to use it in a smart way as opposed to having AI write a law.

35:24Words of wisdom to end on Mihai. Kaoutar, Erin, thanks for joining us on the show. And that's all the time that we have for today. And thanks to all you listeners. If you enjoyed what you heard, you can get us on Apple Podcasts, Spotify and podcast platforms everywhere, and we'll see you all next week on Mixture of Experts.