0:00cost is by far the largest pain point that open models can jump in and solve. But you know, every business has a vision of getting better control over AI and customizing it for their business. And that's really their north star. You know, cost is something they can solve in the short term, but it then enables them to then go and and customize these models for their unique use case.
0:20Early 2024, there was lots of interest in fine-tuning your own custom models. Then it sort of went away and all of that it will just be wasted effort. it'll get stomped by the next model release. Seems like it's coming back now. You have a front seat to all of it.
0:33Do you think we're going through like another cycle or is it here to stay this time?
0:43Welcome back to another episode of the Lite Cone. Today we're talking to Jeffrey Morgan, co-founder and CEO of Olama, the easiest way to run open source AI models locally and in the cloud. Alama is used by 9 million developers, has 178,000 GitHub stars, and is used by 85% of the Fortune 500, which means Jeff knows a lot about the state-of-the-art of AI, what models score highest on benchmarks, and what developers actually download and keep
1:12using. Jeff, welcome to the Lite Cone.
1:16Thank you for having me.
1:17We're down to here. Like, what is the state-of-the-art? What are you seeing out there? Well, I think the biggest thing we're seeing is a shift to open models, especially in enterprise and that's from a mix of US and and Chinese origin models and it's predominantly driven by coding agents and also AI assistants more co-work cases like openclaw and Hermes and because you sit in the token flow of like so many tokens you have really good data on what models people are actually using and how it's changing. What are the trends that you're seeing? Yeah, you
1:46know, started as a way to run open models on your MacBook or other hardware, Nvidia, AMD, Intel, and um earlier this year, we launched Alama's cloud. And what we're seeing there is that it's predominantly Chinese models right now of Chinese origin. Um but they're being accessed by businesses all over all over the world. Um especially US and and Germany is actually a big source of where open model tokens are being accessed.
2:13Is it all about cost? Is it are enterprises coming because they just want to get the cost down or is there anything more to it?
2:18Cost is by far the largest pain point that open models can jump in and solve but you know every business has a vision of getting better control over AI and customizing it for their business and that's really their north star. You know cost is something they can solve in the short term but it then enables them to then go and and customize these models for their unique use case.
2:39Is there a particular uh large enterprise that you can name that has done this? I think there was a great article in the information yesterday uh from AT&T and it ends up they've already shifted 40% of their token consumption to open models and that's right now predominantly through US uh and and Europe models but they're also evaluating the Chinese models.
3:01What kind of workflows do they run?
3:03Predominantly coding agents. I think what we've seen just from the extreme growth and you know per developer or per user token usage has predominantly been from coding agents and then earlier in March and April we saw open claw take off and subsequently the Hermes project the Hermes agent project take off which has then opened up that ability to automate a huge chunk of work over a long span of time to non-developers too whether it's like finance or support or marketing or sales. you had this
3:32actually very cool graph on the takeoff exponential for for OpenClaw.
3:37Yeah. So earlier this year, you know, this is a graph of token usage by developer on Olama's cloud. You know, the average amount of tokens they use per week and we kind of this is per developer. So like it looks this graph looks like it should be an aggregate of Lama's growth, but this is actually the per user.
3:56This is on an individual user user basis. how many tokens are they using a week? And so there's kind of like two big inflection points. One is that initial runup at the start of the year which was driven by coding agents. So we saw Kimmy, the GLM models, Miniaax launch. Finally, we had open models that could power coding agents. And then in April, we saw this incredible growth from OpenClaw really, which was then not just developers, but the rest of the world could take a hard problem, give it
4:24to an open model, and let it go complete the task, which obviously consumes a ton of tokens as it's figuring out what tools to use, what data to go fetch. We went from a context window of 128K to a million plus with open models. And so all that enabled this explosive growth.
4:39So it went from uh roughly 5x so under somewhere around I don't know 15 million tokens before before all these core work type of use cases.
4:49I think that's about right for that open claw jump we saw in April. Um obviously in aggregate it's you know in the 10 to 20x if not more as a whole through cloud we saw 150x since the start of the year and so it's just the big you know interesting thing here is this huge surge in demand for open models right whereas open models I'd say in 2024 2025 from the large models being served were mostly being served as custom models so you'd take an off-the-shelf model like
5:17deepseek or kimmy and you'd fine-tune it for your use case um like for example you know cursor had famously done and from there, you know, you could serve that at scale. But seeing out of the box open models being served, that really only took off at the start of this year.
5:31Even since we started this podcast, these things sort of come in cycles maybe. Like I feel like early 2024 there was lots of interest in fine-tuning your own custom models. Then it sort of went away and it's like all that it will just be wasted effort. It'll get stomped by the next model release. Seems like it's coming back now. You have a front seat to all of it. Like what's do you think we're going through like another cycle or is it here to stay this time?
5:55I think the release of these models the cadence is only speeding up which makes it ever more harder to you know stay on top of that and have a a post you mean the the the latest closed frontier models are releasing faster than ever. I think on the open source side it's getting faster and faster. You know this summer we've already seen three iterations of the Deepseek Flash model as an example. um what used to be more of a six-month cycle.
6:20Yeah. So the gap is like closing essentially closing and I think that makes it even harder to custom train models. On the flip side, I think the tooling is getting better and so it allows teams that want to fine-tune their models to stay on top of it.
6:32I mean, we're also entering this moment where AI safety is becoming more and more of an issue at uh the Frontier Labs. So the frontier uh may well slow down to figure out its alignment and containment issues. And then meanwhile the open source models and openw weight models are continuing to grow and get better.
6:51Yeah. And you know we saw the announcement and the release of the GLM53 model and its capabilities from a cyber security standpoint you know being extremely impressive. it creates a big opportunity for whether it's startups or existing businesses in the security and governance space to really jump in and help because I think if you look at you know that AT&T article I was speaking about it largely the the blocker to adopting open models is largely around security and safety um but from our experience talking to customers whether it's in Europe whether it's here in the US if you can solve the
7:20safety problems by and large adopting the Chinese origin model labs is completely on the table and it's really exciting for these businesses there's sort of this interesting moment right now where a hugging face had to use openweight models to actually even detect the hack from the frontier.
7:36A common question we get is like well where where can I use open models that are leaps and bounds of an advantage over uh using a frontier closed model and one of the key use cases is security testing and making sure that you know your software is secure.
7:51Yeah. Because like if you try to get claw to like pentest your product it will just refuse to do that.
7:56Correct. Yeah. By and large. Whereas there are literally obliterated uh security researcher models that you can find on hugging face that allow you to do it.
8:05That is true. There are ones that are are you know customtrained to be you know even more you know liberal to to go and attack these problems. But even the out of the box models um they they do come with safety training but they're a little better understanding if you're doing this for you know a good you know use case versus versus one that's more of a negative or you know malicious use case. A cool thing about is that because you guys are such a key distribution channel for these models, my understanding is that typically the
8:34model developers are contacting you before the general release to like coordinate launches and stuff like that and you often get sort of like previews of what's about to about to happen and you get these incredible growth spikes when a new model drops. Like I wonder if you could like tell us a bit about like what it's like to operate this thing at scale.
8:52Yeah, absolutely. And like I said before, the models are coming out faster and faster and faster. And so we've developed a playbook to what does a successful day zero model launch look like? And there are a lot of things to get right. There's making sure that it's supported in your favorite inference engine, which is generally sometimes a multi-week process to make sure it's fast, to make sure it's accurate, to make sure it's up to spec with the reference. There's also finding the right use cases and harnesses for developers and users to make use of this model. Generally, these models have new capabilities. This morning, you know,
9:21there's announced Deepseek launched their first multimmoal model um from a large model LLM standpoint. They had previously had some smaller OCR models, which unlocks a whole bunch of use cases. But what they also changed was the Deep Seek harness, which launched recently so that it could support this capability. So step two is then to find the harnesses, make sure they're prepared to actually run this model and to do that effectively. But you know, every model's different. They all have different challenges and architecture changes and tool calling mechanics and
9:50getting all this right is really hard. I think the key thing to do at the end of the day is to run benchmarks against the final product ahead of release and make sure that you know it's it's running as as it as the research team at the lab specified. So I think one very important part that you play in the whole ecosystem is you kind of create a very legible standard to be able to make sure you have the best way to use the hardness for each new model each new tool call and all of it is consistent across all the different models which is
10:20pretty hard to do. Yeah, I think there are three things really that you know we try to package together. One is a harness for example, you know, whether it's an off-the-shelf harness or an SDK to help use the model. Some existing harness that is designed for this and and what's great now there's so many great open source harnesses. Um the codeex harness is open source. Uh open code's a great one and one of the most popular harnesses from OAMI users. Step one's getting that right, but then you got to package it with the model, making sure that it's, you know, available,
10:49it's reliable, if it's in the cloud, that there's enough capacity for it because day zero tends to be the largest growth day obviously. And then importantly, there's the hardware and the providers. And that's actually where there's a lot of collaboration to be had whether it's like an inference provider optimizing the model or you know with some of our partners whether it's Nvidia or or you know for example working with the Apple silicon stack making sure that it actually runs the model really fast because if the model's capable but it's really slow that's not a great experience. So getting those three
11:18things packaged into a box and generally you get the model you're lucky if there's a you know the ability to access the model a few weeks in advance. A lot of this stuff comes together in the last 24 hours before the model gets released and so it's generally a fire drill.
11:31So you're kind of almost a like an operating system in the old world where you needed to really integrate very tightly with all the drivers, all the hardware and then at the application level to make sure that all the apps were really tuned up well and you are the glue for all of it. Right? I do think the OS, which is generally a cliche analogy to use, is a good one because you've got the drivers for the hardware and the providers and the inference layer, but you also have the application runtime and making sure that the harness works. Gluing that together,
12:00it's a very combinatorically large problem to solve. And so doing that well is really difficult. And so, but you know, over time you develop pieces that, you know, allow you to quickly develop that, test it, release it, um, and kind of have this common runtime that can match any harness to any model. And that's kind of the role we're really playing for developers.
12:22You guys are very hardcore engineers and you fine-tune things all the way to from the origins with Apple silicon all the way now to DJX. How do you build such a deep technical bench with that? I think a lot of our team, you know, we aren't AI researchers by background. We're from VMware and Docker, um, and from, you know, other networking companies. And so the classic compute problems are kind of reinventing themselves in the inference land, whether and and so largely, you
12:50know, that's where we like to focus our time. But I think at the end of the day, you know, it all comes down to the developer experience. What is it like when the developer makes a call to the API and gets tokens back? What happens?
13:01And it there's more and more happening in that layer right now. And getting that right needs all the layers of the stack to work well together. And I think you know one of the challenges with open models has been that hasn't been happening at the rate of what a frontier model lab puts out where they have you know the classic five layer cake right that Jensen mentioned which is like the apps you have you know the model you have the infrastructure and inference you've got the chips and you got the energy and they've got all that ready to go for developers on day zero. And
13:30that's really the thing that we're trying to reproduce for open models. Of course, we're not going to do every layer of the stack, but we can help orchestrate that.
13:37You layer the lay layers.
13:39Yeah. And maybe in open models, there's more than five layers. Like that model layer actually has a lot to it, right?
13:44There's the model weights, but there's also a lot of the orchestration components that each model is uniquely good at. There's this developer API layer. There's so many opportunities to build between the model and the application layer that are kind of hidden in today's, you know, five layer cake stack.
13:58That's interesting. Do you do you think any of those like hidden layers might get unbundled and become their own companies or providers?
14:06Absolutely. There was a really good talk from some of the anthropic team, the platform team, and they talked about three big things. One was knowledge. How do you connect your company's data and context to the model? One is coordination. So, as as you know, when you make a request to in your cloud app or codecs, it goes off and spins out a bunch of sub aents, some of those in the cloud, some locally, there's a coordination problem. Um and then lastly there's an execution problem which you know we talk a lot about as sandboxes but these agents more and more of them
14:34are moving to the cloud. There's a huge compute problem to be solved there. And if you look at what's happened classically and in the cloud business open source has meant that there are best of breed companies for each of those problems. Whereas you might have had kind of the if you look back to the original generation cloud products you have like the Herokus of the world you have Google app engine all those things were bundled together. But what developers ended up preferring is best of breed products for each one.
15:00It's the nalism which is uh all things are just bundling or unbundling like the frontier model frontier model labs want you to be totally in their walled garden of manage agents and their context their memory layer. And then meanwhile little tech and all the founders out there and all the open source developers uh don't want to be caged in so we're going to make all this other stuff and it'll it'll be an interesting moment to figure out like what ends up winning. I mean it'll probably be some mix of both.
15:27I think so. And you know you've got this abundance of open model tokens that's being created. There are just dozens of open model providers that are able to serve these tokens. The new scarcity the problems now are what's above the tokens right? You know how do you orchestrate an agent from you know A to B. These are problems that are have tons of new you know systems and and engineering problems that are just really hard to solve for an individual dev. Like there's no way they're going to build all those layers. Well, the interesting thing now is um because the coding
15:56agents themselves are getting a lot better and ostensibly this is the worst the models will ever be. The classic reason why there was a moat here was it was just too hard to have really well-maintained software that was properly tested that actually satisfied user need. And what if that goes away?
16:14Like we're literally at this moment where actually maybe the a you know you'll just have a curron and it runs a markdown file in some TypeScript and it'll just you know there's no lock in anymore right like you could be using OpenAI's memory system one day and then actually you know you could have an agent be uh constantly syncing that against your own memory system and actually it just works it's fine like it works over MCP like there's all you know
16:42there's plenty of like runtime testing and then there's no lock in.
16:46Yeah, I think a lot of the, you know, what we know of as harnesses today, a lot of those pieces will go down into the model. But, you know, as you kind of push a lot of the core loop of the model and hooks down to the model itself, there are these pieces that kind of come out like memory is a great one. And the the general guidance we're using is if it's a stateful problem like there's storage involved that's something that you know in the end can't go into the model because the model's trained and
17:14it's you know as we know training runs are now happening on like a monthly basis but it's still not up to date with the latest data. So generally storing data is like a huge problem space that I don't think will ever make its way down into the model layer like managing credentials and that kind of stuff isn't security credentials safety um open models don't have all the safety tooling that closed model providers give you out of the box but that's super important especially for businesses to adopt them.
17:40I'm curious where you see sort of the the end or future state for enterprise on the balance between sort of frontier closed models and open source model um especially on spend like my it feels like you know initially it was just the everyone's just like allocating all of their budget to anthropic or um open AI uh my sense now is yes open a open open source is clearly growing but like so are so is like anthropic spread so like the things seem to be growing together
18:08like does that continue or Do you think there's sort of like a steady state where it's like I don't know like half the budget's going to be on the closed source frontier model and half's going to be in open source or or something different?
18:19The super majority of tokens and this is our take it will be open models within a business. Call it 80 90%. That doesn't mean 89% of the the budget will go to open models. In fact, I think what the open model community is doing incredibly well together is lowering the cost to make it more accessible. And so maybe you'll only pay 10 to 20% uh of of the cost towards open models, but your token most of your tokens will be going through the open models, right? Which will enable a whole bunch of use cases on top because you have this abundance of tokens. You're not thinking about taking away token access
18:49from your team. You're giving more and more access. I think for the hardest tasks that's reserved for these frontier labs where a lot of the best researchers are. And then from there, there's a whole bunch of problems in the middle, right? Where maybe it's a combination of open and closed models working together.
19:02I think the steady state is that most of the software is most of the models are open. It's kind of an interesting idea because as the models get more powerful, ideally you just want to delegate to like your smartest model to figure out when to go to like an open model, but the labs who own the models presume you don't want that.
19:20And I think look, I think that all the labs are aligned in many ways to one thing, which is how do you serve the customer? And I think it will be up to the customer to decide if I have a router where some of the scheduling and harder you know orchestration happens through a frontier model so be it. But a lot of the kind of line item work can happen through open models and the collaboration of the two together. I think we've seen a ton of projects whether it's from Sakana AI or open router that have combined the two and it's seen really good results.
19:47This is not too dissimilar from a human organization, right? Like you have you know like a law firm there's like a partner and then there's like a bunch of associates and like the partner farms out the work to the the associates, right? It's like the same.
19:59We saw the same thing with cloud computing where it was really a blend of proprietary software. some of them provided by the the cloud providers themselves. For example, you know, AWS had Dynamo DB which was kind of their proprietary scale out database. But then a lot of customers use that in conjunction with Postgress DB and in the end you know what we see is customers will will use a combination of the two.
20:20I think it's a very common pattern. This is also the same design for why the Apple silicon is actually more superior.
20:26the special spe special special accelerators for different kinds of workloads for let's say image processing supposed to audio that's been or even go way back in in in the PC era you had like your standalone audio card right graphics card and all that and speaking of Apple Silicon should we talk about local models because you're you're in a bit of a unique position because you have large businesses both in cloud hosted models and locally hosted models that will run on your laptop what are you seeing in those two
20:55worlds and What what do you think is going to happen?
20:57I think it's incredibly exciting because it's similar to the closed versus open model question. It'll be a mix in our mind and and that's what we hear from customers as well where for easier tasks you could run them locally and with lower latency and of course lower cost when it comes to the per token cost.
21:15Ultimately, you're buying hardware up front and you'll use that in conjunction with these cloud models. What's exciting about this next generation of hardware, which we've had for a few years now, is just how good they are at running the 20 billion parameter to 40 billion parameter range of models. Sometimes up to 128 billion parameters.
21:32Yeah. Quinn 3.8 38B is now as good as Opus 4.6 for coding. Is that right?
21:38That's what the benchmark show.
21:40It's incredibly exciting because you can run that on not the the the lowest memory MacBook, but the second lowest memory MacBook you can buy from the store. So it's incredible.
21:50And are you seeing your users do that?
21:52Like what are you seeing people use the local models versus the cloud hosted models for in practice from the side of which models they're running? We're seeing a really solid mix of US and Chinese trained models being used for local and we have the incredible models from you know the original llama models of course but also the Gemma models from DeepMine. These are great choices for local. But when it comes down to use cases coding agents by and large are most effective with the large cloud models. You're solving really hard problems. you're writing
22:19code tests, it's really difficult uh versus some of the document processing workflow use cases that run extremely well locally because they don't have as difficult of a of a task in the endto-end, you know, problem you're trying to solve. And so that's where we see this hybrid execution model where some of the easier, more straightforward tasks run locally and then you have a router that can help decide, hey, we need to go to a large cloud model for this. And I think what that means for customers is that you're you're really
22:45dropping the costs even further when you're going from open models. Not just because they're cheaper to run in the cloud, but because now you can run them effectively for free on the hardware you're buying for your business anyways.
22:57And what we're seeing ultimately from the cloud coding agent models is it is predominantly Chinese models being consumed today. And for the local models, it's a really strong blend of uh US, Europe, and and Chinese origin models. Yeah, this these two graphs are pretty stunning in comparison. Like basically for local models, the US and Chinese models are neck andneck. We're like tied. And for cloud hosted models, like the US is like recoloring the x-axis. It's like 100% Chinese models.
23:26Basically, we need more US labs to make large models. Is is is is that what this graph is showing effectively? And you know, with the launch of the Neotron Ultra model, we're seeing kind of the first wave of that and it's really exciting. Nvidia as a company is so interesting because um their moat is not like trying to start new software businesses or sell you know tokens. They seem to be quite interested in just releasing a lot of open- source and helping the ecosystem and then the
23:55fact that they do that then helps them stay ahead of the game on the hardware side.
24:00I think so. And you know ultimately Nvidia what's so incredible is there's helping power an ecosystem around open models whether that's the hardware the models you know we've seen the uh new DGX station uh computers that they're working on which have GB300 on your desk.
24:17How do we get on that list that isn't deafening loud?
24:20Yeah. Do you know the price point on that thing yet?
24:23I don't know it off the bat.
24:25I know, right? Well, I looked it up.
24:27It's like I mean you can probably run a Frontier model for like I mean very slowly for like two two $300,000. Is that is that right?
24:36I think it's even more competitive than that. Yeah.
24:38And you can run more than a Frontier model at high speeds at a price point that isn't very far off what you can buy from a classic workstation computer.
24:48You think about, you know, quite a few of the customers we talked to, some of them are banks, for example, or industrial businesses. They already have these Nvidia workstation GPUs in every single engineer's desk. Some of them tens of thousands of them. And so this is one.
25:02They've been selling them for all the CAD work.
25:05Well, this is for this is for the kind of original RTX A6000s, but this is the next generation.
25:10Yeah. Yeah. I'm going to have to email Jensen.
25:12This was at this was at GTC. So, what we see here was at GTC. And we were one of the first uh people along with Elon and a few others to receive the DJX Spark as well which sits on your desk and provides you know 128 GB of unified memory to run that kind of 20 to 120B model range but obviously that's just the beginning of a whole new range of hardware that can run the biggest models.
25:35So did you say you can buy a bunch of these and chain them and actually run a 400B model?
25:40You can absolutely. They they have this really fast network link and so you can stack them on your desk almost like a miniature uh data center rack and people do with the Mac minis.
25:51Now this is like the production version of it.
25:53I think that's what's so exciting is you're seeing both from Apple and Nvidia this incredibly incredible leap to next generation hardware that's built for these models and is effective at running them. So you'd say like this is this is the platform to get like you could make stu apple studios work but like if you want something that just can work get DGX spark from our testing both are very competitive. Got it.
26:15So I think a lot of it come down to what you can get and and then also um the tech stack you're looking for. I think there's an incredibly uh mature tech stack through the MLX project with Apple where they've done some amazing work to run LLMs on the Mac Studio but also the smaller Macs and of course the DJX Spark stacks. just incredible. We're super excited uh as partners with Nvidia for that.
26:35I think it's going to be a cool renaissance for personal desktops.
26:40I think so. And you know, it's it's funny with Olama's journey. We started local. Clearly, the coding agent demand is in the cloud, but that's going to come back locally in our minds because the hardware will catch up. When you have a GB300 on your desk and you want the fastest coding loop, that's as fast as running your tests or as fast as making code editor change. We all remember the GitHub copilot experience of having the autocomplete come up in a few millisecond at 100 milliseconds.
27:06That experience will make its way back to the desk which is which has been a journey of starting local going to the cloud and then we think that'll come back local and you'll end up using the two together.
27:15Speaking of like what you can get in order to run a Llama cloud, you need like a ton of GPUs. What are you seeing in the GPU market?
27:24I think what we're seeing is ultimately the prices are changing very quickly and the supply and demand volatility is is very high there. And so, you know, I think ultimately for if you're a startup, getting access to some of the B200, B300 GPUs you need to run these latest models is very hard.
27:42Thankfully, there's a great set of inference providers building on top of that. And so we're seeing this extreme demand on like you know and and when are you able to get all the GPUs that you need? Are you constantly like like growth limited by how many GPUs you can get your hands on? What's the what's the current state? We're lucky that we've partnered with quite a few providers to work together to pull a bunch of GPUs together, which allows us to stay on top of our our demand, but that's a lot of work and it's definitely a lot of
28:11spending time thinking through, you know, which model will get run where, how fast should it be, which region is it in, um what will the latency be for the customer? There's a lot of hard problems to solve in that stack. And I think what's really exciting about products like Open Router, Alama, the Open Code project is for an end user developer, they can sign up and get access to this without having to go negotiate prices on a B200, B300. You know, think about their 24-month forecast in order to get access to some
28:40of these these uh you know, GPUs.
28:43YC's Next Batch is now taking applications. Got a startup in you?
28:47Apply at y combinator.com/apply. It's never too early. and filling out the app will level up your idea. Okay, back to the video.
28:56Suppose you were like a a startup founder and you were just starting out now and you're building some AI company and you haven't raised a lot of money and so you like want to like use as many tokens as possible like inexpensively like what would your advice be to that person about like how they can get like huge mileage with like a limited budget?
29:16There's this new class of models like DeepSeek Flash is a great example and I think there'll be quite a few more where it's ultra low cost per token. It's also low cost per task which is a really important metric and that class of models in my mind will be the first ones that come down to this idea of like unlimited tokens. We all remember Chad GBT. You didn't really have to think about how many tokens you were using.
29:39You would just use it every day. You had unlimited. Ultimately, I think we return to that, but it's going to take a lot of work in the model, the architecture to be customtrained for high volume token usage. And if you think about the start of the year, we really wanted open models got to the frontier of intelligence. We bridge the gap. We're maybe like less than 3 months behind between the frontier closed models and the open models. But the next problem to solve is extreme efficiency. Seeing, for example, the GBT Luna model become very, very price effective for customers has
30:08been a huge boom. We talked a ton of customers where that kind of pricing enables widespread adoption within a team. I think we're going to see that with open models. We already are seeing that with open models. I think the Deepseek flash model is leading that charge. If we go back to the model breakdown on OAM's cloud, the highest growth area is definitely the Deepseek model. And this is largely powered by the Deepseek Flash uh adoption. And so this new class of flash models where they're good enough for 80% of the tasks, they're really fast and they're ultra cheap. this new class of model
30:37that I think will enable some of those use cases.
30:40Yeah. Those are going to be like the workhorse models to do like all the grunt work.
30:43Yeah. You won't have to be thinking about how many requests am I making, how many tokens. You'll be much more inclined to consume as much as you can because you know it's able to solve the hardest the not the hardest but you know difficult problems. If you think back to the coordination layer we were talking about earlier too. Being able to coordinate these flash models together to do different tasks can also yield great results that a bigger model can.
31:09So by having these cheaper models, not only are they more accessible, they can run faster and you can access them in higher volume, but you can start to chain them together and build new problems that are solved by orchestration on top. And that's a really exciting area for new startups, for existing inference providers, um for some of the larger businesses today that solve workflow problems. ultimately being able to chain these models together is going to be super helpful and you won't have to think about the underlying costs.
31:35Yeah, I guess you know when we first started talking about AGI even on this podcast um there was this sort of debate about you know and I think a lot of AI researchers would come out and say like there's just going to be a giant god model and it's going to do everything.
31:48But you know I think so far like it hasn't quite worked out that way. Like obviously you still have you know if you're if you have to literally hack the NSA maybe you need mythos or something but um for the majority of use cases like you're talking about orchestration and you're talking about um like smaller models you you you know that the task composition actually probably gives you a bunch of ways to make it more repeatable. It's more trustworthy like it actually does work at a cost that is
32:17like possible. So, you know, if it was going to be um God model versus like lots of, you know, smaller special purpose or even just like simpler models, um it's turning out to be the latter. So far, I think for most customer use cases, there's a you know level at which a model becomes good enough and then they can continue using that level of intelligence. Maybe the model will get faster, it'll have better architecture, it'll be'll have new capabilities, but they won't have to reach for the god model. But I do think there are use
32:47cases where the most powerful models unlock them and that'll continue to be a thing and it'll be really exciting, you know, and sometimes scary as well on on what they can do. But for the run-of-the-mill use cases where open models really shine, I think that's where, you know, we're we're hitting a point where, you know, you're not solving necessarily the hardest problems within the business, but they're hard enough where it's now unlocked by open models. Will there be an open model that becomes a god tier model? Um I think it's possible and we're seeing you know
33:16really exciting developments from you know Zoo AI and GLM where some of the tasks they are frontier and we all saw with the Kimmy model how for web development it became the best model and that sent this new shock wave across the market which is it's less about a gap and it's more about a head-to-head competition which I think makes all of this even much more exciting.
33:37This is a bit of a sensitive question but what do you think about this and the geopolitics around it? I think you know a lot of the geopolitical angles around this um start with you know where the where the model's from and the more we spend time with customers and users a lot of it's actually how the model's run where it's run how it's run is it run in a secure environment um and that starts to matter a lot more um but I do think you know look it it's super important
34:05that a customer in the US can use a model trained in the US and we they have two kind of classes of of customers we speak to. One is they don't really care where the model's from. They care about where it's run. But for every one of those, there's, you know, a customer that's saying, I really care about where the models from because it's data the way it it's not even just a a security issue as much as how does the model speak. You know, we all go through mo and and communicate. We all go through, you know, a lot of the models go through
34:35phases where they sound more robotic, they sound more friendly, and a lot of that matters, too. Um but I think the highest order bit is obviously making sure that you have a model that end to end you understand where the data is from which is great from the neotron models you can go and introspect what made this model because if you're putting in a mission critical task which people are absolutely using open models for mission critical tasks there's a post online about how llama powers the analytics of a power plant to detect surges in Finland to make sure that the lights stay on. That's where these
35:04models the the model origin really matters for for like critical tasks like that. How do you ensure that a Chinese model even if it's hosted in the US isn't basically like booby trapped to like cause problems?
35:19Yeah, the menuranian candidate problem.
35:22Have there any been any like known cases there? The menurian candidate yet?
35:26I think not that I can think of off the top of my head. I haven't heard. I feel like I would have heard about it, but you know what you don't see a lot on some of the press articles is how robust some of the the IT and security teams are at the businesses that we know of, the top, you know, Fortune 500 businesses, they're really used to this already because open source software, if you think the average application has thousands of dependencies. This is like isn't a new problem and all it takes is one dependency for there to be a major security issue in the entire application.
35:56Clain poisoning is insane.
35:57It's a thing. It's been a thing for decades and it's not new in that sense. It's a little more opaque because you can't like dig into the model. It's it is deterministic but it's deterministic and if you screen the model properly with safety checks by and large at least what we're hearing is from customers is that can be solved.
36:14Do you want to talk about the origins of Lama? you you know you guys came up through the Docker ecosystem and uh a lot of people watching you know would love to be in the position you're in where you have this sort of enduring brand moat that looks like it will extend for you know really till the end of time. No just it's it's a very powerful situation to be in. um you basically found yourself on top of a giant oil well, right? Um for those out
36:43there wildcatting, you know, can you tell us that story? You you were actually working with Jared in uh 2021.
36:51Yeah, you know, my my co-founder and I previously built Docker Desktop while at Docker. So we really got a understanding of like what makes a great developer experience. But I have to say the first few years of Ola as a company was really in search for what's the right problem to solve with this muscle we've built of trying to design a great experience for developers.
37:12You applied to YC with a very different idea. Right.
37:14Do you remember what the like tagline was when you guys applied to YC in Winter 21?
37:18I think it wasn't well defined. I think we realized let's go back to building a really great desktop experience for containers and Kubernetes.
37:28I remember what I wrote down on the on the application. It was a kitematic for Kubernetes or a Docker desktop for Kubernetes.
37:36Yeah. Which was effectively Docker Desktop. They had a great Kubernetes. I think you know it's one of the challenges as a second time founder that you know Michael and I have told ourselves. We we tried to overengineer the idea in many ways and I think even the two to three years after like we did YC in 2021 and Olama wasn't launched until July of 2023 after we raised our series A after obviously after we had done YC that journey was one of really
38:05in search for a customer problem that could delight a developer and in some ways it was almost a good thing that we tried different ideas and pivoted until 2023 because that's when Mama Llama came out and started the open model with which is why it's called O Lama.
38:22Okay. No. Oh, really?
38:24Llama means generally from our experience whether you think of local llama as the subreddit lama really just stands for open models. You know, as we were looking through the name, it wasn't necessarily from an existing model.
38:37It was more LLM and it's like the animal plus LLM.
38:41Yeah. And I think having that character was important. We like what's a good name for a character, a face you can put to the name because models animal mascot sometimes.
38:50Docker had one GitHub.
38:52He still hasn't taken my advice to have llamas come to actual O Lama events.
38:59I'm curious what your series A pitch was because you raised from Benchmark like fantastic investor but all of this the future we're in now hadn't quite taken off in 2023. So what was like the pitch and the vision back then?
39:11Yeah. and we partnered with Benchmark in 2022. So it was Deli days but pre-C Chad GBT when it came down to the pitch I think we weighed so much on like hey we're trying to build this great developer experience we're solving this security problem and we had known u Peter the partner at benchmark from our previous lives building building a docker because he was the series A investor in docker and so a lot of it was weighted on you know the people and also why we exist I think the what I
39:40mean you know solving SSO for Kubernetes, which is a real problem, wasn't really our passion. I think we were really lucky to find a partner that could could see us for what we stood for and what we were trying to do versus the point in time, you know, problem we were solving at that point.
39:54Yeah, I see. So, you as the a sort of pre- pivot, correct? Okay. I didn't realize that actually. H I was just looking at this cloud tokens by model family graph and like basically if you just look at this graph it looks like the Olama story begins on in February 2026 and it like explodes thereafter which is like so funny because like of course it actually goes back to like 2021.
40:18What was it like to be like sort of lost in the wilderness for like many years working on stuff that was like kind of working but like not really taking off and then all of a sudden to have things like just like explode like how did how did it affect you and your co-founder psychology and the team and the employees? What what was the experience like?
40:36It was definitely scary and and and for a few reasons. You know, one is like when you're when you're trying to solve a problem for devs or for a customer and you're just getting on the phone with them over and over again and it's not totally clicking. That's, you know, it's less about the are we in the headlines or is is the project taking off the product we're building. It was just are we truly actually solving a problem for somebody? And I think being lost in the wilderness like what's your north star that customers are generally a great north star, but not seeing the north
41:06star is even scarier, right? Because often you know what problem you want to solve, you just haven't figured out what problem and I I think the you know Michael my co-founder and I we started this company because we had built a company in the past and we ended up being acquired by Docker very early. It was just the founding team and Arnorstar was saying we want to go solve a great experience for developers with something they find really hard but man in the two years where we're just finding that problem it's really scary. you know, we had a team of more than 10 people which made that really hard. And I'm so
41:36thankful to that team for staying by our side as we went through different ideas, you know, and and what what's not really obvious is we went from this security for Kubernetes to then like security for developers on the desktop, which is like the pivot that we've never spoken about.
41:50And then we kind of took that form factor when models came out. We said, well, it was a leap, but it was we knew kind of the the kind of problem and the feeling a developer wanted to have, but LMS finally made it realize like it was it was crystal clear at the point when we tried running the llama model and it was really hard and we're like, okay, this is a problem and it's really impressive when you get it working and it's kind of just a zero to one moment. I'm curious for the story of that pivot because there there actually like many pivots in the Olama
42:20story, but probably like the most critical one was like the pivot to Olama to doing like locally hosted LLMs. Like how did that come about? Were you just like tinkering with ideas on the side?
42:29And when you found the idea, was it really obvious to everyone in the company that that was the thing to do or was there like like a a big debate? And it wasn't until it took off that it became clear.
42:39We you know sat in a room together. I remember we were in Toronto because we had a team split across Toronto and Palo Alto and now we're predominantly in Palo Alto and we were saying throw everything out like if we had to start from scratch and we were just joined right now what would we do and you know we had seen two big problems because we had talked to some users and LMS we tried using open source LMS ourselves or just LMS in general one problem was could you build a gateway to access any model and host that and make that really seamless back then we were thinking of it as like the
43:08segment for you know LLMs it's a good way to think about which I think has become really this big router idea which is only at the beginning.
43:14It's a massive opportunity. And the other problem was we were a bunch of XVMware Xdoccker folks like we know how to make things run and so like let's help make things run with open models and then we kind of systems yeah systems and so we kind of tried to really introspect our team which I wish we had done sooner because security is a very different team and sale than developer tools and just by doing that we gravitated towards saying let's just try this thing let's give ourselves two weeks to launch the first version of Llama and then Llama 2 came out and we said that was right at the end of the
43:44two weeks we said okay we're launching And we just had a bias to action. And if you think back like in 2 weeks all of that happened, going from idea to shipping it to getting to more users than we had ever had with our previous stuff. And before that was two years of just frankly overthinking the customer, the product, and just not getting something out there.
44:02The first time I actually heard about was on Reddit. I didn't realize it was you guys. I was on like that logo. I just was interested in like running local models and it was on like the I think the local LLM subreddit or whatever and everyone was just raving about OAMA and how great it was.
44:16I was like, "Oh, it's a YC company."
44:17Yeah. I found I found out later actually cuz you you were called a different company. You weren't in our internal system and Yeah. I remember catching up with Jared and saying, "Oh, hey, by the way, there's all that security stuff. We have this thing now. I think you were catching up with Jared and then I bumped into you on the stairs and I think you had your t-shirt or some swag or something and I was like like you guys are a lot like you know I meeting a rockstar or something.
44:43It would have been really hard to time this but I wish we had taken that leap much sooner. I mean the best time to do it was during YC.
44:49It wasn't possible didn't exist.
44:52Lama didn't launch yet. I guess you were also one of the first GitHub projects that very quickly got to 100,000 GitHub stars, right? Do you remember how long?
45:02It was like very quick.
45:03Yeah, I can't remember exactly how fast, but it was much faster than Docker and Kubernetes. To your point, it it things kind of just started working and started taking off and you're really as a founder just beside yourself because you can't totally explain why. I think it's the best way to explain product market fit. And there are different levels of product market fit. you know, we only started monetizing earlier this year with Llama's Cloud, but just to see people fall in love with the product.
45:30It's such a zero to one moment that um I wish we had done it during YC, but in some ways it wasn't possible. I also think it's just kind of wild to put into perspective like you sort of went from being in sort of like the cranks on Reddit like interested in running their own like rigs at home to like 85% of the Fortune 500 in like 2 years or something like that. That's like that's like a pretty That's the homebrew computer club to uh broad computer adoption like speedrun
45:58that took 10 years for the PC took like 18 months 12 months.
46:03Yeah. And that's one of the things that surprised us the most because I think look I think open models the original users very much hobbyists just tinkering. Oh my god this is even possible but very quickly because you know two things one is they were free to get started with and you could run them anywhere. that is incredibly helpful to a Fortune 500 IT developer team because they don't have to ask for permission to use it. And so it just happened what was really good for a hobbyist user translated very quickly to a developer
46:31within a business. Um it just happened to be a case where that was that was it. You know, for example, databases, we saw some of this too where a database that started for devs like MongoDB very quickly also moved to enterprise, but because LLMs are stateless, it made for such an easy transition. Now, moving to the cloud, there's a lot more in play. There's an economic question. If you're a customer, there's obviously security.
46:54Where's the model running? But what's beautiful about open models that both hobbyists and IT developers loved is you could just get started. You didn't need permission.
47:02Can we talk about the monetization angle? cuz this is this is interesting too. So like in 2023 Llama 2 takes off all of a sudden you got all these users 100,000 GitHub stars like you've clearly found something but it's basically like Reddit cranks who are using it. You're making no revenue and there's no obvious path for how you will ever make any revenue from all these like cranks on Reddit. It was 2 years before you actually figured out a business model for it which funny enough is exactly the position that Docker was in. Like how did you think about it during those two
47:32years? Were you were you worried about it? Were was the team asking like what's the business model going to be? How did you think about like figuring out how to make money from it?
47:40I think there's always two ways that we saw open models um being able to monetize in a way that's great for the company, great for the developer and great for the the customer. And one of them was a privacy focused AI product which Ola started really started with that in its open source incarnation. But we always felt that there was this moment where you know you weren't using llama with uh the llama models for example with with tool calling right
48:09away when they came out. So there were use cases where it was still reserved for the frontier models. And again with at the risk of overthinking it we kind of saw that there wasn't the product level of product market fit with open models that closed models had. And in some ways philosophically we want to align with when that happens we want to be there to capture that. I think it happened this year with coding agents running with open models because you had the largest consumption of AI being matched with finally open models being able to service that. There are a lot of opportunities along the way to do it
48:37privately securely. Again, a lot of the Fortune 500 have already adopt Lama. But we really asked ourselves what would be the most important problem we could solve for a customer. And the local piece, while an important part of that story, never felt like the whole story, which was how do you access open models for the hardest problems? And so, in some ways, waiting. We knew we had to wait a little bit for the market to mature. At the same time, what are the risks of waiting? Well, you build a
49:05culture if if not careful. And we had learned a lot of this from our Docker days where you don't think about monetization. It's not a priority. I think from our previous battle scars as a team, we we kind of had we knew about that. But I think the other component which is really important is making sure you keep in touch with your customers.
49:22Like one of the biggest risks of having an open-source project that takes off is you consider your user base and customer base like your customer just a blob on the internet which is a really risky way to think about customers because you want to meet them figure out their needs. What are they doing? What do they want to do in six months? What's their story? And I think that's the thing I wish we had done a little more in the last few years and we're doing a ton of that now.
49:43One thing I'm curious when you went through IC, you guys were second time founders. I'm curious what got you to decide to do YC. Actually, you know, we we went back and forth on this for a lot, which we shouldn't have.
49:54We should have just said, of course, we're doing YC. Um, but by and large, starting a company is a really lonely experience. Even if you have a great co-founder, and Michael, my co-founder, was the co-founder of my first company.
50:06He was my college roommate at University of Wateroo. Um, but it's still lonely and I think just having a set of peers, even though we did it during the pandemic, just talking to Jared and like five other groups of founders every week really helped you feel less lonely. And I think that's such an important part of it. And then of course when we finally moved down here and there was no more co the network was just incredible. And the fact that we could meet founders building on open models, building on any
50:36kind of AI, you know, we kind of knew that was going to happen cuz we had known so many founders from the University of Water who had done YC pre-COVID and they were like, it's really about getting together and like that was a big part of it and we knew that was there and you know, I think that's that's what made it a no-brainer.
50:52But also just I think there are a lot of mistakes you can repeat that you don't have to. And what I love about the YC community is how transparent founders are with each other about those and you know I still keep in touch with the founder of Docker uh who's an investor in our company and we're able to talk about some of these challenges we saw in the previous generation of companies that you know we don't necessarily have to repeat or things that worked and we can bring you know into the future.
51:18Yeah. If you just uh don't repeat one of those mistakes that you know sometimes is the mistake that would have killed the company potentially. Yeah. Yeah. our our famous saying, you know, a bunch of our team is from uh companies that ended up working great and Docker is doing phenomenal now, but whether it's, you know, some of our team was early early at VMware and it there are always ups and downs and I think just having a group of people around the table who have a collection of those and also what worked actually what worked is actually even more important and just being able to like
51:47have that muscle memory is a big part of it. Oh man, I was just thinking about this because we obviously hang out with and work with a lot of 18 year olds or 19 year olds and then sometimes they're always asking like, "Well, what should I do?" And then I'm starting to realize like one of the more important things is if you've never worked on a team that shipped really amazing technology to like a lot of people or like like just real clear product market fit, like do that once. Like even if it's a month, even if it's like three months, you would learn more in those three months
52:16because then you know what good looks like. And then without that it's like I mean it's not like it's impossible like people at YC do figure it out because you but it's that much harder. Like the difference between having seen something that actually works from like beginning to like some form of like this is what the bug database looks like and this is how we release and this is the quality that's necessary and here's like the bar that we hold each other to. Having seen that it just like multiplies the chance
52:45that people succeed. So it makes sense that you know starting off with a co-founding team that has seen a lot of that pretty powerful.
52:52Yeah, I think it it provides you a set of values you can work around especially when you have so much power in your hands with AI. There just parts of it that I can help you with but it won't hold you accountable to it and you know how does software work and look at for example I'm sure there's versions of running in the wild from two years ago.
53:08How will your software work when somebody falls in love with it and continues using it for two years? Is it still going to be working well?
53:14hopefully they update to the latest software or it's you know a cloud service but I think you build that muscle memory and we definitely have that from a lot of our more senior engineers on the team who are at VMware or NERA for example but at the same time I think there are a lot of lessons we learned in the previous generation of DevOps and infrastructure that aren't valid anymore in the AI world oh yeah tell us about it what have you found what is not valid anymore I think a good example that I classically used is there was this generation of companies called platform
53:44platform as a service and in the contents the Heroku of the world was a great example of this.
53:50I mean Docker started out as that Docker started out as a platform as a service and there's this concept that if you're a layer on top of something else that you're in kind of a vulnerable position as a startup which is absolutely not true in the eye world and in fact going up the stack can sometimes be even better because you're closer to the customer. in an infrastructure world that that's also the case and that that was like an analogy that we had to like so many of these muscles we actually had to break building a lama. Another one was you know these LMS are never perfect and like in the systems world you want
54:19everything to be exactly as it's designed to run. It's tested it's validated but LMS by definition are not that's a feature not a bug.
54:26Exactly. It's a feature.
54:27You want it to be a little non-deterministic I suppose. And I think from building a team too, it's that you know with AI now there are just problems that you don't need to staff as heavily whereas you you you did 10 years ago, right? If you think about what does your customer support pipeline look like?
54:46Um what does it look like to deliver a cloud service? Like it's a very different world with AI because how do you build a service where no engineer knows exactly how all the code works which is obviously the case now. And so there's just new lessons we're learning going from like a, you know, some of our team from infrastructure 1.0 in the 2000s to cloud in the 2010s to now the AI space. There are a lot of rules that break.
55:07I mean, you're probably actually doing an incredible service to um like both sides of the ecosystem and that like the end users get this like very clean thing that just works, especially like the tokens just come out and they're very clean and the API makes sense and it's rational and logical. And then on the flip side like I mean if you don't have a layer like I've directly experienced this where it's like oh yeah the underlying inference provider has a weird error for you know if you put this parameter in this way or it expects JSON
55:37and you know it's not documented. It's just like this insane minefield. Like, you know, the agents can kind of figure it out, but like you're going to like bang your head into the wall for like a couple hours before, you know, the agent figures it out. And in the meantime, you're like, "This is a terrible experience." You know, and so you're like in there probably helping the inference providers fix all these fundamental bugs, too.
55:57Yeah. And it's part of the the job we do. And I think one of the big opportunities in the open model landscape is curation and taking a fragmented universe of models and inference technology and cloud services and harnesses and like making that actually just work is a really valuable problem because the end developer to your point they just want to build their software right they just want to build stuff they want to build their next company their next application and I I think that's where we come in but it's
56:26where a ton of great services also come in and we saw you know open router obviously is a good example of that from a wide model selection so developer doesn't have to sign up for you know hundred different providers they can just go to one they can pay in one place I think we've seen with open code you know you can have one harness that integrates with any model it's a really powerful experience for developers just looking to try the next model to see if it solves their use case better so this curation and you know when there's a abundance of models and providers now
56:55there's a scarcity in bringing that together into something that works.
56:59Thank you so much for joining us. That's all we have time for.
57:01Thank you guys for having me.
