Free with a sign-in.
The best 14 min of 54
Who Wins if AI Models Commoditize? — With Mistral CEO Arthur Mensch
Arthur Mensch argues that as AI models converge, durable value will come from customized enterprise systems, not a lasting lead in general models.
Big Technology Podcast · Alex Kantrowitz
- Arthur Mensch is CEO and co-founder of Mistral.
The best minutes
- 1:39Model commoditization shifts value downstream4 min
- 15:55Open models give enterprises control over critical AI4 min
- 39:57AI can automate complex shipping and factory workflows6 min
Also worth watching
- 11:12Enterprise AI can unlock growth beyond labor savings5 min
- 25:39Open-model gaps are shrinking as AI shifts toward specialists7 min
- 6:48Enterprise AI needs fixed workflows alongside agents4 min
- 45:51AI systems improve through feedback, not just code fixes3 min
- 51:42Enterprise AI adoption may take decades despite infrastructure spend2 min
Skip the sponsor reads, the Mistral services overview, the China open-source and geopolitics discussion, and the robotics segment.
Full analysis · 6 min read
Real Alpha · Nov 8
Arthur Mensch says the 2025 open-model gap was around three months
- Mensch said he thought the 2025 open-model gap was around three months, after maybe six months in 2024.
- His convergence case supports customization, but he offers no comparative cost or return figures.
- Buyers should test workflow gains, model control, and upkeep before scaling budgets.
Free with an account
Read the rest of the analysis
6 sections, the 6 other voices it tests Mensch against, and what it means for you. Every episode we cut gets one.
Check your inbox
We sent a link to your address. Click it to finish signing up. Not there in a minute? Look in spam.
or
Signed up before? The same buttons sign you in. The episode, its moments and the transcript stay free.
Training knowledge spreads and model leads fade
At 2:23, Mistral CEO Arthur Mensch says leading models will commoditize. He says around 10 labs share similar data and training recipes. The know-how is fairly short, he says, so it circulates.
That supports a convergence thesis, not a claim that every AI asset is interchangeable. Training recipes can spread while deployment, domain data, and process redesign still vary by buyer. Indeed, Mensch later says wider compute access helped open models close their gap.
Mensch says model assets depreciate quickly, and downstream cash generation must fund investment. But he is not neutral. Mistral sells models, customization tools, and enterprise services, the layers he says should capture value.
That interest does not disprove his thesis, but it makes financial comparison important. He gives no comparative Mistral or rival costs, margins, or returns.
Our read
Model convergence is the episode’s premise, while services are Mensch’s commercial answer. Buyers should treat both as hypotheses until outcomes and ownership costs are measured.
Enterprise returns depend on implementation
At 4:49, Mensch says enterprise value must justify model investment. He says companies generally report they have not made money from AI yet. He attributes that to too little customization and solution-first thinking.
For example, he describes a supply-chain workflow run by 20 people that could, he says, be handled by two. The episode gives no baseline, duration, or accounting method for that illustration. So it signals a possible labor reduction, not a verified saving.
That gap matters when a buyer writes a business case. Kong CTO and co-founder Marco Palladino asked on MTS in October 2026 how much business outcome each token dollar produces. His yardstick forces a company to link spending with results, rather than count model activity.
A pilot needs a before-and-after measure tied to the process itself. Otherwise, reported efficiency can conceal integration effort or displaced work.
Our read
Mensch identifies plausible savings, but the example does not establish their size. Measure business outcomes and delivery costs, not model activity alone.
Agents need fixed workflows and specialized models
At 9:39, Mensch divides enterprise systems into human-defined workflows and agents that choose tool calls dynamically. He says the dynamic part is growing, but static rules and guardrails remain important. That is a case for bounded autonomy, not agents running every process alone.
He expects specialists to drive progress in the next two years. For instance, he names physics, chemistry, biology, and aircraft design as fields where experts can supply training signals. Mensch says broad reasoning will improve, but does not expect one lab to create a strong lead that way.
Jean-Denis Greze, Town co-founder, discussed the adoption gap on the a16z Podcast in October 2026. He said companies need someone who understands AI and will tinker with a solution. On Unsupervised Learning in November 2026, Nebius CTO Danila Shtan prioritized reliable agent outcomes over model size. That supports the systems focus, but neither moment proves customization earns a return.
Jean-Denis Greze · a16z Podcast · Oct 19AI adoption depends on showing people what to automate- DSDanila Shtan · Unsupervised Learning · Nov 2Agent reliability may matter more than model size
Our read
The model is one part of a system shaped by rules, tools, and domain feedback. Workflow ownership and reliable outcomes beat general rankings as procurement tests.
Open weights trade convenience for control
At 16:38, Mensch says Mistral chose open-model implementation, customization, and services. He says this lets firms run systems on their own terms, avoid single-vendor dependence, and train around private knowledge.
Open deployments also carry costs. Mensch says they require tools, expertise, and deployment maintenance. A hosted closed model may be easier to use, while open deployment offers more control over hosting and customization. His case is strongest for critical or proprietary workflows, not every use.
Bernardo Aceituno, Stack AI co-founder and president, emphasized hosting, editing, and version control on MTS in October 2026. That supports control as one reason to choose open weights, beyond price. On 20VC in November 2026, Modal CEO and co-founder Erik Bernhardsson treated ownership as one factor. He put latency, cost, and performance alongside it.
Mistral research scientist Sagar Vaze made a similar proprietary-knowledge case on MTS in October 2026. Because he works for Mistral, that is supporting context, not independent validation.
Bernardo Aceituno · MTS · Oct 17Open-weight models give enterprises control over AI deployments
Erik Bernhardsson · 20VC · Nov 2Model ownership matters for some workflows, not all
Sagar Vaze · MTS · Oct 29Open weights let firms turn proprietary expertise into a moat
Our read
Open deployment is a trade, not a default winner. Buyers should compare control and customization against their internal skills and upkeep.
Industrial examples show potential, not proven returns
At 40:55, Mensch gives two industrial examples. Take CMA CGM, a shipping company, where he says AI coordinates container dispatch after ships reach port. Dispatchers must contact hundreds of people, including harbor officials and regulators, and act across 20 software systems.
He thinks a few hundred people currently handle this work. Separating contacted people from the worker estimate matters. He says automation could save 80%, but never specifies whether that means labor, time, or total cost.
He also says the system is live in certain agencies. The episode supplies no baseline or independent measurement. At ASML, Mensch describes models analyzing scanner images for chip-engraving errors.
He says the work will increase factory throughput, but gives no measured change. The example shows why proprietary inputs may matter, not the value already realized. Both cases make the implementation thesis concrete. But they leave buyers to verify reliability, safety, integration effort, and actual operational gain.
Our read
The industrial cases make a credible use-case argument, but do not establish ROI. Treat them as places to test, not proof to budget against.
Adoption may trail infrastructure by years
At 52:04, Mensch says enterprise adoption is slow because companies must build custom software and reorganize teams. He says buying off-the-shelf tools alone will not deliver immense productivity gains. That is a material counterweight to the industrial examples. It also means infrastructure spending can precede customer value by years.
He thinks the industry may be over-investing a little and over-committing a little. He excludes Mistral from that assessment and points to some others. That is his judgment, not industry data.
He expects the economy eventually to run on AI systems, but says the shift could take 20 years. That is a forecast, not a schedule enterprises can plan around.
His customer advice offers nearer markers. He says build a prototype working 80% of the time, then gather feedback and retrain toward 99%. Those percentages are his examples, not a universal production standard.
Budget owners should ask whether a pilot can prove value before deeper reorganization. Otherwise, infrastructure growth can be mistaken for current demand.
Our read
Mensch pairs confidence in eventual diffusion with a long timeline and some acknowledged overbuilding. Near-term budgets need evidence from staged deployment, not assumed economy-wide demand.
What it means for you
- Set an outcome baseline. Before a pilot, record the process’s current cycle time, exception rate, and labor effort. Mensch’s examples focus on workflow change, not model scores.
- Bound one workflow. Choose a process with a clear owner and human-set rules. Test a working prototype, then track user feedback and retraining needs.
- Price the control trade. Compare hosted and open deployments on control, customization, and maintenance. Mensch says open models require tools and expertise as well as deployment upkeep.
- Stress-test the timeline. Ask whether the business case survives slow adoption and organizational change. Mensch says the wider shift could take 20 years.
What to watch
- Whether specialized models improve measurable outcomes in the next two years, as Mensch expects.
- Whether user feedback and retraining move prototypes from around 80% toward 99% reliability.
- Whether enterprise adoption catches infrastructure spending or takes the 20 years Mensch says it might.
Send this to product, finance, and operations leaders planning AI pilots.
Watch the minute where Mensch estimates the open-model gap for 2025
Written by Real Alpha from the episode's transcript and the other moments it links, then checked against the transcript. Quotes are the speaker's exact words.
Minute by minute
- 1:49Model performance is converging
Mensch says foundational models are likely to become commodities because roughly 10 labs can access similar data and use methods that spread quickly. He argues that this makes it difficult for one company to stay far ahead, so model makers must weigh investment against how quickly those assets lose their edge.
- 4:51Enterprise value lies downstream
Mensch says enterprises often have not made money from AI because they start with a solution instead of the problem and do too little customization. He argues that model companies should focus on removing workflow friction; a supply-chain workflow once run by 20 people, he says, could be operated by two.
- 7:03Enterprises need specialized systems
Mensch rejects the idea that one system will solve every task, arguing that enterprises need specialization and data suited to specific problems. He describes AI systems as a combination of static, human-defined workflows and guardrails, and dynamic models that use tools; he says both parts will remain important.
- 11:45AI can replatform enterprise software
Mensch says consumer AI could support an advertising business, but Mistral is focused on enterprise opportunities. He describes AI systems that navigate company data and generate interfaces for specific tasks, potentially changing how people use enterprise software. He also sees value in training models on proprietary industrial data to unlock technical progress, not just efficiency.
- 16:33Open models offer control
Mensch says Mistral has chosen open models because enterprises can customize and run them on their own terms. He argues that this gives companies more control over access, deployment and service continuity, and reduces dependence on a single vendor. He says open models can also help turn employees’ accumulated knowledge into company-specific systems.
- 20:17Services help enterprises deploy AI
Mensch says few teams yet know how to build reliable AI systems at scale, and enterprise deployments also require changes to how teams work. He expects technical customization to become easier to use, but says services will remain important. At Mistral, he says, model building and customer services reinforce each other: its models and tools support customization, while enterprise needs inform model design.
- 25:39Open models are closing the gap
Mensch says the performance gap between open and closed models narrowed from about six months in 2024 to around three months in 2025. He attributes this to training saturation at about 10 to the power of 26 FLOPs and wider access to compute. He says open models still require more effort to customize and maintain, but their performance is converging with closed models.
- 28:24Specialized models will differentiate
Mensch says general reasoning will improve, but he does not expect one lab to establish a lasting lead over competitors. He sees greater opportunity in training models with domain experts and task-specific data for fields such as materials, physics, chemistry and biology. He argues that smaller specialized models can be more efficient to run, and that skills in one field may not transfer to another.
- 32:22Mensch frames control as Mistral’s edge
Asked whether Mistral benefits from European regulation, Mensch says its differentiation is the ability to give customers more control over where and how AI runs. He says this matters to regulated industries, governments and defense, which may not want to depend on a foreign vendor. He says Mistral can serve customers in the US and Asia as well as Europe.
- 36:14Open source spreads research
Mensch says Chinese AI companies are strong and that open-source releases let labs build on one another’s work. He points to Mistral’s mixture-of-experts research and says DeepSeek used the same architecture. He expects major economies to seek some autonomy in AI, creating multiple centers of expertise rather than one country pulling decisively ahead.
- 40:23AI is reaching industrial workflows
Mensch says Mistral is helping CMA CGM automate cargo dispatch, a process involving hundreds of contacts and 20 software systems; he says it can save 80% and is live in some agencies. With ASML, he says, Mistral is working on vision models that inspect scanner images for chip-engraving errors and could improve fab throughput. He expects manufacturing processes to be reorganized around AI systems that perceive information and take or route actions.
- 46:31AI systems need feedback to improve
Mensch says stronger models help, but companies must also iterate rather than expect systems to work out of the box. He advises customers to build a prototype that works 80% of the time, then use user feedback and retraining to move toward 99%. He expects text-heavy applications such as legal work to reach production sooner than applications in the physical world.
- 49:10Robotics needs tailored control systems
Mensch says robotics depends on both capable hardware and control systems customized to the platform and mission. He expects deployments first in places where avoiding human risk matters, including firefighting and manufacturing. He says household humanoid robots may take much longer to arrive at scale.
- 51:47AI investment may be running ahead
Mensch says enterprise adoption is slow because companies must build systems around their own knowledge and reorganize teams. He thinks some companies may be over-investing and over-committing, though he says the economy will eventually run on AI systems. That shift, he cautions, could take 20 years.
Published Jan 16, 2026.
Transcript · 43 min read
Big Technology Podcast, hosted by Alex Kantrowitz, with Arthur Mensch.
Transcribed by machine (Whisper) from the episode’s audio, so a word may be off here and there.
What does the AI business look if all the leading models perform the same, which they kind of are? We'll find out with the CEO of Mistral right after this. for a free consultation. Welcome to Big Technology Podcast, a show for cool-headed and nuanced conversation of the tech world and beyond we have a great show for you today we're going to talk all about what's happening to the ai business and technology race as some of the leading foundational models start to look the same and how that changes the balance of power in the industry we're joined by the perfect guest to do it arthur mench is here with us he is the ceo and co-founder of mistral arthur welcome i'm happy to be here uh and uh thank you for Yes. No, it's great to have you. So Mistral is a name that those who are deep in the AI world know very well, but might be new to some of our listeners and viewers. So for folks who are new to Mistral, let me give you a couple of stats. Mistral is an AI model builder, does some other things which we're going to get to. It's based in France. Company is valued at $14 billion after They're starting in April 2023, so a little under three years or two and a half years to make a $14 billion business.
Not bad. There's 500 people at the company. And Arthur, you are leading it after spending some time in the academy and two and a half years at DeepMind. Exactly. We're headquartered in Paris, but we have around the 444 workforce, which is actually in the US. And a lot of our activity is actually here. So that's why I'm spending a lot of time as well. and that's why we are here in New York. All right, well great to have you in studio. Let's just go right to what I think is the most pressing issue for AI today. There's been so much
“where is the value accruing? And what kind of business model should you pursue to actually make sure that in the end, you're turning profitable?”
Read the whole transcript
That is very interesting because for a long time you would hear these companies focus model, model, model. The next, what's GPT-5 was, let's say, when you think about OpenAI, the biggest news. Now, they're starting to talk more about how do you take the intelligence that you have and build the applications that work. Just one bit of reporting that I can share a couple weeks ago. I had this story basically inside a lunch with Sam Altman and a bunch of newsleaders in New York City. and Altman told him the companies, one of their biggest priorities was building applications for enterprises. Basically, it's gonna be a major priority in 2026.
And it's a little bit of a shift in rhetoric from we wanna build AGI till we wanna build applications for business. So talk about why is that happening? Is that an offshoot of this commoditization?
“It's never going to exist. I mean, you have a wealth of problems, just like you don't have any human that is able to solve every task in the world.”
I'm thinking through like what the business is. Let's say the model has commoditized.
“These are things where we are starting to run into physical constraints and artificial intelligence can actually help to lift those physical constraints.”
“if you assume that the entire economy is going to run on ai systems well enterprises will just want to make sure that nobody can turn off their systems”
especially since we were interviewing the people that thought this stuff was sentient before ChatGPT came out. But that's a conversation for another time. But what we're basically saying today, I'm going to sum up two of the main points that you've made. One is that today's AI models can't do it all themselves. They need orchestration. and the second big point that you made is uh to do that sort of orchestration or implementation with the current intelligence you need a service like a managed service so i it is interesting to me that like we've gone from like this perspective of you know maybe working towards a god model that could do it all to the fact that you know this this may be the most powerful technology that we've seen come through in our lifetimes however when you actually want to use it you kind of need it becomes a managed service in a way.
That's interesting. This is true. I don't think it's the first time that we observe it in history. It's a new technology. It's a new platform. And so the knowledge on how to use it is actually still pretty scarce. So there aren't that many people that can build systems that are performing at scale, that can run at scale reliably, that can actually solve an actual issue. And so when working with enterprises, you always need to have some services on top because of the complexity of implementation, even with fairly well-understood technology like databases. But for artificial intelligence, it's even more necessary in that it requires to transform businesses.
So you need to also help in thinking how the team should perform around the system itself. And it does require to customize things. So you need data scientists that know how to leverage data and turn it into intelligence. And today, this is still a pretty scarce resource. I would say I do expect the part of the of the software in those deployment to increase so the the amount of the way customization occurs today with fine-tuning reinforcement learning this kind of things this is going to be abstracted away from the enterprise buyer because it's too complex and they actually should just worry about having adaptive systems that are learning from experience and from deployment with people instead of thinking about should I use fine-tuning or should I use reinforcement learning to actually put that knowledge into my models?
And the work that we are doing is to try and abstract away from lower-level routines that data scientists understand to higher-level systems that business owners can actually use. And so it's going to occur, and we're working on it, but the service part is still going to be quite important. And today, the combination of the two things is the fastest way to value, if you're an enterprise. So we've been combining the two. You know, I started our conversation by calling you a model builder and I kind of paused on it and I said some other things that we're going to get into it later. And here we are.
Basically, what I'm hearing from you is that Mistral, obviously proud model builder, but it seems like without the services, without being able to sit with the business and showing them how to use it, just would be an incomplete puzzle. So do you consider yourself, like is the most important thing you do building the models or is the most important thing you do the service or are you primarily a model builder or primarily service provider? I mean, we are there to help our customers get to value. So service. We're here to, but to get to value, they need to have great models. And to get to value, they need to have the right tools to train the models.
And so the best way to train, to create those tools is effectively to train the best models. So the two things are extremely linked together. We create models that are very easy to customize. We create models with tools that we then export to our customers so that they can use them. And we help our customers train their own models. So you can't go and sell to an enterprise that you're going to help them create very custom systems if you can't show to the world that you're effectively the leader in open source technology. And so the two parts are equally important. The first is enabling the other.
And there's effectively a flywheel there because we make our choices when it comes to the model design in a way that is enabling the various customers we have. So one example is that we've put a lot of emphasis on having models that are great at physics because we work with manufacturing companies that run into physical problems. So that's the flywheel that we have set up by having the science team and the business team actually sit together. Okay. We're here with Arthur Mensch. He is the CEO of Mistral, also co-founder. When we come back after the break, we are going to talk about the open source movement versus closed source.
Remember, DeepSeek and open source was supposed to surpass closed source. Well, has it? I'll also talk about the geopolitics and regulation and whether that's going to give this company a leg up and then maybe get into some more practical examples because we should talk about how the technology is being used on the ground. We'll be back right after this. And we're back here on Big Technology Podcast with Arthur Mensch. He's the CEO of Mistral. Arthur, I want to ask you about the progression of open source over the past year. I remember reading about DeepSeek, doing reporting on DeepSeek in January.
and the The overriding theme was it was such a leap forward for open source that soon the closed models, models like OpenAI's GPT and Anthropics Cloud and maybe Google's Gemini would be surpassed by open source because the open source community was working together and building on each other's innovations where the closed source community was kind of going at it on their own. We just had this moment, we talked in the beginning of the show about how maybe Gemini commoditized OpenAI's GPT models, but that conversation was not being had about open source living up to that expectation from the beginning of the year.
So am I missing something or am I reading it wrong? Or what do you think? If something has held back open source, what has it been?
“If you look at the trend in 2025, I think the gap is more around three months.”
I'll just tell you what people in the US say and let you address it because it's worth talking about. I think there is a feeling among some, not all, but some, that Mistral has been set up in Europe to effectively take advantage of regulatory capture because U.S. companies have a hard time competing in Europe.
And therefore, Mistral will be there to pick up all the AI business. What do you think about that argument? Well, we've built our technology so that we could... serve companies and states that wanted to have enough control. Artificial intelligence is not a technology that you want to fully delegate to a vendor, especially if it's a vendor that is from a foreign entity. And that was true before, it was true for data. It's going to be all the more true for artificial intelligence for multiple reasons. But one of them is the fact that this is, if you're depending on a external vendor, your commercial balance is effectively increasing and you're importing services. And that becomes a problem long term if you're importing too much digital services, for instance. So that's one thing. And then sovereignty and this kind of topic is also very important for defense. As if you're an independent country, you want to have independent defense systems. And if you want to have independent defense systems, you will need them to, you will need your own independent artificial intelligence because this is making it into the defense systems. So it's really working for you. This pitch being like, we are
not an American company. We're based in Europe. We'll be able to help you build, whether it's something with like important data protection or national security like defense. Well, it's a technological differentiation we've built. So because we can build on the edge, because we can deploy wherever our customers wants us to deploy, we effectively can die and the system is going to still be up which is which actually matters for many many industry and the more critical you get the more it matters and so what that also means is that we can serve the u.s uh u.s customers we can serve u.s customers that want to depend less on certain providers we can serve banks that wants to have more customization more control that are more regulated it also means we can serve we can of course serve the european industry uh where historically that's where we were based we you sell next door when you start your company.
And that's what we did. But we also serve Asian countries. And Asian countries, they have similar problems. They want to have a technology that they can rely on, even if we were to die. They want to have a technology that they can customize to their own cultural needs. And so that has been driving our business for sure. That aspect, that technological differentiation that we've built around control, open source, like a technology built on open source models, and around customization. And do you have like European governments coming to you and being like, we just don't trust Google or Anthropic and we'd prefer not to build on them?
Well, we have European governments actually coming to us because they want to build the technology and they want to serve their citizens. They want to increase the efficiency of their public sector. And we happen to have a good proposition for them, which is deployable on their premises where we can go send forward deployment people to help them get to value. And it turns out we're European as well. So it's actually pretty good for European countries to invest in European technology because the investment they're making, the revenue that they are creating for us, is a revenue that we reinvest in Europe and we're effectively creating an ecosystem around us.
So that investment of the flow of revenue from European countries to European technology provider is something that is very beneficial. And to be honest, in the US that has been working for the last 80 years, I think in Europe we haven't been doing it enough for sure. Speaking of open source companies, there are efforts that have some links to geography. What do you think about China's open source effort? Because obviously they've made a lot of noise. It seems like things are going quite well there. Yeah, I mean, China is very strong on artificial intelligence. We were the first actually to release open source models, and they realized it was a good strategy.
And they've been they've proved to be very strong, actually. And so we've been not sure if we're competing, because the good thing about open source, it's not really competition. You build on top of one other. Right, you see everything they have out there and you learn what works well. And the same is true, the reverse is true. Like we released the first sparse mixture of experts back at the beginning of 2024 and they built on top and they released DeepSeek 3. DeepSeek was built on top of that. Well, it's the same architecture. And we released everything that was needed to rebuild this kind of architecture.
And the same is true. I mean, everything that companies that are investing on open source are releasing are things that other open source companies are reusing. And actually, it's kind of the purpose. R&D is just much more efficient if you share your findings across different labs. And so it's been very effective in China. They share knowledge across the different labs. It's been pretty inefficient here in the US because there's actually no, there's like US incorporated company are not investing on open source. And we've taken the lead on just being the West open source provider. And I think it's going to be very much needed to have a Western open source provider. What do you think China's strategy is? And do you think that there's like, in the US, there's often this kind of very large conversation about the need to stay ahead of China? Do you think there's a risk if China runs away with this? Well, I think China is very strong. It's vertically integrated. They have strong engineers, they have compute, they have energy, they have everything they need to compete. Europe also has everything it needs to compete. I don't think we'll be in a setting where anyone is going to have one artificial
intelligence ahead of the others. And if you look at the world in its entirety, every large enough sovereign entity, which is a big economy, is going to want some form of autonomy in its usage of AI and its deployment of AI. So that does justify the emergence of multiple centers of excellence, I would say. One of them, which is in Europe, which is led by us. One of them, which is more in Nangzhou in China. And then you have a bunch of companies here in the West Coast. Why do you think it's in China's strategic interest to develop these open source models? Because they don't have a similar business as you do, right?
They're not real. They're not like going out globally and becoming implementers? They have a big business in China, for sure. The companies that are building open source models in China are actually cloud providers in general. You have a bunch of startups, but you also have Alibaba, which is a cloud provider. And so they have this vertical integration that allows them to create value there internally. So in China, but also in the markets where they are operating and growing. So in Asia, for instance, which for us is a place where we tend to compete with them, not in China itself, but in the rest of Asia.
So it does make sense. for them to compete internally. And then their best way of accessing the US market is by just giving the things for free. And so it does make sense. It's a very natural thing to do, to build a business in China, which is protected, than to export the thing for zero. That's not bad. I would do the same if I were in their shoes. Right. All right. I want to talk a little bit before we leave
“I think a few hundred people to do it. And by working together around how to automate those things, suddenly you can save 80”
“the one thing that we try to convey to our customers is that they need to build a prototype that's going to work 80% of the time.”
That brings us to robotics. So let's end here. People have been talking about how we could see an explosion in robotics because of LLMs or the advancements in world models. But it still seems far off. I mean, they had this demo, what was it, the Neo, the Neo-humanoid robot. Where there's like a person controlling it, teleoperating it. Kind of weird. They might be in your house. So we haven't seen progress in robotics, you know, start to move as fast as we've seen it in the software side, in the large language model side. So where does that, when does that come, if it ever does?
I think in robotics, you have the combination of two things that needs to work. Hardware platforms that needs to be, you need to have the right actuators with the right haptic signals that needs to be built at scale with good economics. And this is starting to be true. And we're not the one working on it, but the industry has made a lot of progress on that domain. Then the other thing is that you need to be able to have control system that are sufficiently intelligent to be deployed on those on those uh robots and so that's where actually we we come in in that again you need to have custom models uh because the problem is the model needs to be customized to the platform to the whether it's a humanoid robotic or whether it's something on wheel or whether it's a flying drone and it needs to be customized to the mission uh because the mission is going to bring different kind of images the kind of actions that can be taken are going to vary across the mission maybe the guardrails are different and so that adaptation to the world and to the wealth of data that the hardware platform that is being deployed is bringing does require the right platform and the right training platform and so our bet in robotics and what we've been doing with multiple
companies in defense in particular is to build that platform that allows to train models fit to purpose that can then be deployed on the edge potentially because in robotics strategically in robotics, I believe we'll see deployment of such systems first in areas where you don't want to send humans. So firefighting, I think it's a very good example. So when the risk and benefits, the risk of deploying the system is way under the benefit of deploying the system. It's going to be the case in manufacturing as well, because there are places where you just want the factory to be dark. And I think that's where most of the, a lot of the value will be created. I would say midterm.
and then maybe long term you have things that are sitting in your house but you know it's a bit dangerous to have like some pretty strong thing out there and so the same way we've been waiting for self-driving car for the last 15 years we'll be probably waiting for like humanoid robotics in-house for meaningful time and before that what we'll see is at scale deployment in manufacturing and that will take the right software platform and that's the software platform that we're building Okay. All right. Really, the last one. We've talked a lot about AI and business. Some businesses have gotten a lot out of it.
Some have not.
“But today, I would say my view is that we're maybe over-investing a little bit and over-committing a little bit.”
Our guest has been Arthur Mensch, the CEO of Mistrawl. Arthur, thank you so much for coming in. Really appreciate being here. Thank you for hosting me. You bet. All right, everybody. Thank you for listening and watching. And we will see you next time on Big Technology Podcast.
