# Open-Weight AI: Control, Infrastructure, and Licensing Shifts

**Podcast:** AI + a16z
**Published:** 2026-08-06

## Transcript

The fun thought experiment is if GPUs dropped in price by 99%, then do we get back to a real open source world?
If moderation is never solved, in the future people will go to open way by default.
Because that is where you know for sure you can control your guardrail for trusted use cases.
And can you talk about where VLOM sits in that stack?
VLOM is an inference engine.
It is kind of like databases and...
operating system and other critical software to power AGI, NVIDIA, AMD, Google, their newest chip will make sure VLOM can run on them.
And in a lot of cases, they use VLOM as a benchmark.
We're bridging almost a 10x gap.
For proprietary models, there is a regular mode and fast mode.
But for open-weight, every provider can offer potentially even 10 different levels of speed.
Five years from now, open source AI models, have they closed the gap with frontier models?
Capability-wise, I don't really see a big gap, not even today, because...
Open source AI has become one of the most important forces shaping the industry.
But making frontier models available to the world requires far more than releasing model weights.
It requires an entirely new infrastructure layer.
In this episode, Elena Berger and Matt Bornstein...
are joined by Simon Mo, co-founder and CEO of Infraact, to discuss how open source inference became critical infrastructure for AI, why enterprises are increasingly turning to open-weight models, and what the next generation of AI systems will require.
They also explore model licensing, inference economics, and the future of open AI.
Today, we're here with Simon Mo, co-founder of Infraact and a lead maintainer of VLLM, the open source inference engine now running on half a million GPUs at any moment.
We're also joined by Matt Borenstein, an A16C general partner.
Simon, Matt, thank you so much for joining us.
I think first we should start with open source AI and kind of the more recent history of open source AI.
So VLLM actually has its origins kind of back in 2022 pre-ChatGPT.
And your team set out to make a slow open source demo faster and instead just found this.
pile of unsolved problems.
So can you talk about what made serving an LLM so fundamentally different from the ML workloads everyone already knew how to run?
Yep, good to be here.
So serving a large language model is a fundamentally different problem because serving it requires to run it on accelerators like GPUs or TPUs.
And it is...
a computationally intensive process that will require a lot of engineering and ensuring that for each request, user can see the LOM's response quickly and efficiently.
So this typically means we need to handle differences in input distribution, how long each request is, output distribution, which is non-deterministic, and batching and scheduling a lot more at the core of the inference engine.
So the team, the project has been around for about four years now, but the company is a little bit more recent.
So I want to throw this out to either of you.
Matt, I know you've known the team and observed the team for a very long time.
So at what point did you see this sort of transitioning from being a much beloved open source project to critical infrastructure and then a company?
Yeah, I mean, I think you have to go back a little bit, right?
Open source.
was the norm for AI models early on, right?
I mean, we literally had this company called OpenAI, which, you know, it's become a little bit of a joke.
It's not as open as it once was or not.
nearly as open as it once was.
But early on, all the frontier AI work was being open source or at least released into open weights, which is a little bit different than true open source.
And people could mostly run these models sort of on hardware or computers that they already had.
I'm curious, Sided, like maybe you remember, like what was like the first model that like you actually needed to go out and get special software and like a special set of computers to be able to run?
Probably BERT.
And before that, it was like ResNet for computation.
images, computer vision, classification.
So ResNet already needs to run on NVIDIA K80, which is kind of one of the first SEU on AWS and other places.
But even at this point, ResNet, you can still kind of run on a commodity, even CPU devices.
It's just very slow.
But for BERT, where running at it is like, wow, you have to run it on GPU to make it anything faster and efficient for anything, translation, any task.
So that was like before 2020 even.
It's so funny thinking about this.
I'm like, all the memories are flooding.
Hugging Face had a thousand BERT variants on it.
You have to go find like the right BERT variant for your particular task.
And yeah, and you're right, you had to.
Some people, I guess, had their own GPUs and could run it.
But yeah, a lot of people had to go provision cloud.
Exactly.
Stuff like that.
That's really funny.
And so, yeah, so look, I mean, BERT was an early language model that newer models are much bigger, much more sophisticated, take up a lot more memory, a lot more compute.
And so BLLM, really from the early days, right, was about running these more powerful models that you can just sort of do it, figure it out on your own.
Yeah.
Yeah.
And I think to get us to this present day, I think it would be good to talk about kind of at what point it really became critical infrastructure for these even larger open source models.
And when did we even start to see these larger open source models kind of come into the field?
Well, we really see the criticality of the software stack is always about bringing open.
Frontier Intelligence.
This is Open Frontier Intelligence is a catchphrase for the K3 model release this time.
And it's also about VLM's mission about inference infrastructure.
So if we look at it, when do they start being critical is when people are relying on it for their application, for their day-to-day productivity.
So maybe 2023, when GitHub Copilot and ChatGPT are something that people just cannot live out with anymore.
And at that point, open-weight model is already backing and become a cornerstone of how people are living every day.
And at that point, we will need the accelerator running open-weight model and open-weight software to open-source software to make sure it works well and have all flexibility and control over it.
It's sort of interesting from a startup standpoint.
Like you mentioned sort of GPT-3 or like early chat GPT.
those closed-source solutions were starting to become critical to like a small group of people around that time.
And open-source existed, but it was a little bit of a curiosity or sort of an enthusiast thing.
As the frontier has expanded, particularly with closed-source models, more and more open-source has been dragged in as kind of like critical behind it, like if that makes sense.
Like at any given point in time, including now, I think models from OpenAI and Anthropic are kind of...
more widely used and more critical kind of in general than open source models.
But I do think we passed a threshold, and I want to say about a year ago, where a bunch of smaller companies or like new application companies, as they were trying to figure out, how do I really build an AI without just being a wrapper on top of open AI?
The answer to that question turned out to be open source.
I mean, this is what Cursor did.
This is what sort of Decagon and Harvey are in the process of doing now.
And a bunch of other like really, really strong application level startups sort of made the determination we can't build just on closed source.
We need to do our own mid-training, our own post-training, our own sort of inference and deployment tricks.
And all of that means it must be built on top of open source.
The closed source vendors won't give you the access to do this.
So my read is like kind of a year-ish ago.
Open source became really central in a way that's not always visible because it's deeply embedded in some of these products.
Some of the most innovative products and applications now really depend on this very deeply.
Yeah.
Yeah.
And can you talk about where VLLM sits in that stack where we do have these larger enterprise companies that are choosing to use open source models?
Like where does VLLM sit in the stack for them?
Yeah.
I mean, you should like just about everybody uses VLLM.
You should describe it.
Just about everybody uses VLLM.
VLLM is a inference engine.
That means its job is to turn available GPUs into a running endpoint for intelligence.
So that means it is kind of like databases and operating system and other critical software to power this economy or power the AGI that everybody really uses today to ensure they can have cost effectiveness, efficiency, reliability, and also always staying on the frontier.
Because for VLM, we support more than a thousand model architecture up to today.
And a lot of those are proprietary, but also a lot of those are open-weight, right?
And a lot of those model architecture, when they're becoming transitioning from a research prototype to world-accessible open-weight model architecture, they are live on VLM immediately.
So there's a process we call day zero model release.
And additionally, VLM also work closely with all the hardware vendors.
So that means across like NVIDIA, AMD, Google, and Amazon, Intel, and a lot more, their newest chip will make sure VLM can run on them.
And then a lot of cases, they use VLM as a benchmark to make sure it runs well on them.
So this kind of fusion of where models run and where it gets to meet the hardware is where the magic happens.
And this is where VLM is.
And you've told me some of the behind-the-scenes stories.
Like, it's actually not easy.
These days are a model releases.
It's like a lot of human drama in addition to, like, technical work.
I guess, are there any stories there that you think are okay to share?
It's actually a very fun co-design process because from Model Labs point of view, these are brilliant researchers who have built this model.
Now their biggest question becomes, how do we get this out of the world and make sure everybody's able to use it and run it well?
And we have worked with Model Labs that are very, just because they just use VL.
already in production or in their research process, they will just done everything for you.
Because this is a moment when we go to them, it's like, hi, we're the VM team, we would like to support your open source model, we would like to offer in a way this kind of open source, but why gloss service to get your model running well on architecture?
And then you return, we get the model labs like, oh, we're going to working already because we're running it for the RO process.
Here you go, just review our code and merge our pull request.
And on the other end, we really have model labs that just don't know how this can work.
Because systems is not like their core.
Yeah, because systems is not their core and they have been training or maybe they have their internal inference engine that just don't know how it will adapt to the open-way ecosystem.
And by the way, this is also a very much a multi-party kind of involvement process.
Every model race typically involves...
The model lab involves primary or secondary hardware vendors, involves us, involves Hug and Face, who are the model format and model hub vendors.
And then depending on the appetite, the model lab involves 10 or 20 different kinds of release partners.
These could be inference clouds, these could be public hyperscalers, whoever is going to run this model and you want them to ensure the model are running successfully.
So even up to today, if you look at the K3 model release, it's a whole partnership and a drive to make sure that once the model is released, because it's just a few terabytes of files sitting on the internet, that people are actually going to use it really, really well.
Even from the beginning of 2023, 2024, if you remember, when Mistraw dropped their first model, they just dropped a torrent.
linked for PTP and then everybody's like struggling and trying to get it up and running and then we're working behind the scene with Mr.
O'Team trying to get the inference engine support working in VOR and this one is the most probably early on exciting weekend that we are able to spend on this and then after the weekend when everybody's trying to really analyze what's going on and Monday, Tuesday, Mr.
and us just announced, here, you can run it on VLM successfully here.
And everybody will be able to easily reuse a lot of the work and start building on top of it.
That was sort of a fun time where enthusiasts like me could just scramble to download the model and they're running somewhere.
I'm glad the professionals took it over because it never worked very well.
It was like a fun moment in time.
So to bring things, you know, forward to the present, I think open source models and also.
you know, Distillation have been in the news recently.
Infraact signed the NVIDIA OpenWeights and American AI Leadership Letter that was signed by also A16Z, Meta, Amazon, dozens of other companies.
Can you just talk about, you know, your decision to sign that and sort of what you were really kind of responding to in the market and kind of in the news?
Yeah, so for us, what I really want to stand behind is open way absolutely matters in the ecosystem.
The world cannot just be controlled by proprietary APIs and where open way, open development and research of these models are blocked or banned.
Right.
The pledge that InfraGrade sent out for is we want to help and foster this ecosystem where we are typically in a little bit downstream of this ecosystem.
Right.
Influence engine are not part of the pre-training process nor the RL process, but we're where the model actually meets the world.
And from what we're seeing, people are just.
really using their imagination and ability to materialize this imagination of open-way model.
They're able to leverage this open-way model so much effectively.
There's almost two pieces to this, right?
There's like the cost thing where it's like the closed models are too expensive.
And then there's sort of the control thing where I want to sort of be in control of my infrastructure and control the model, right?
If I need to extend it or put on my own guardrails or anything.
I'm just curious, have you heard from, from customers?
Like, are both those things important to them?
Or like, are they kind of willing to pay as long as they have the control?
Or maybe there are different use cases.
I think like, you fluctuate over time.
So control matters a lot over the last few years and then costs just start to matter over the last few months.
So cost really matters starting from people trying to migrate off their expensive coding plan and like every skyrocketing token maxing spend.
But control has always been.
in the backbone of this is they want to even in a way to control the cost, right?
But also it's about controlling the system performance against what they're paying for.
So for example, for a voice agent company, they want to control their own model so that they can make sure the model actually respond by the required time.
So the customer, when they're on the phone, they can ensure the agent is responding according to a SLA.
And this sometimes is only you can do with your controlled intelligence because you know the whole hardware you're running and the whole system you're monitoring versus signing up for relying on your critical infrastructure with a proprietary API where they might go down anytime or have violation of the contract anytime.
Simon, you also, to go back to the cause point, you actually make the point in an essay you recently wrote about the release of Kimi K3 that actually the economics is besides the point.
And it's actually, you know, in the case of these, you know, just really, really great open weight models that are on the frontier that were designed by really brilliant researchers.
Like, these models are in some cases.
just as expensive as, you know, the closed source model.
So in those cases, kind of what is the point of running them and kind of what do we learn architecturally in the course of running them?
Yeah, so first on cost, it's not necessarily as expensive as a proprietary model, but rather first the cost discourse has been discussed over and over again with even GLM 5.2 a few months back.
So open-air model are sometimes definitely a lot cheaper.
But for this model, there's a big sort of step change where we're bridging almost a 10x gap, but strike somewhere in the middle, where Kimi K3 is not as expensive as Claude or GPD Sol, but it is a lot more expensive than JLN 5.2.
Why is that?
And I do believe this is the point of where pricing intelligence with the market correctly and understanding where it is.
But then the majority part of the Discord should be focusing on, wow, this model is bringing a Opus 4.8 level model to our own infrastructure that I can use, I can run, I can fine tune, I can be able to understand exactly how many tokens do I need, understand the exact performance profile.
The reason here, for example, is for proprietary model, there is regular mode and fast mode.
And that's only the two switch here.
But for open weight, when you're running it, every provider can offer potentially even 10 different levels of speed going from like the slowest mode, which can be a lot cheaper, to 400 tokens per second, almost up to 500 in many cases for some workloads.
And this is typically 2x or 3x faster than the fast mode out there today.
So this kind of level of control even in terms of performance and then let alone control over how customer interacting with the model, control over data retention.
Keeping in mind Fable doesn't have zero data retention policy and at least a lot of the data need to be staying there and let it control security and compliance a lot more.
Yeah, this is why I'm particularly excited about K3, not just from the cost perspective, but a lot more on bringing this level of intelligence to something people can own.
In terms of, you know, calibrating things like speed, calibrating other things, just sort of on the back end, what needs to happen and kind of what are you seeing your users do and like who is being really clever about this?
So we do see users are able to get the maximum benefit out of this model when they enable fast mode.
Like what I'm talking about here, of course, is VL on its own fast mode, getting up to 400 and 500 tokens per second, because it is really a big step change from like, especially when developer interacting with the model, they can see, oh, I can really just get my task done faster here.
And the model are no longer stuck in syncing.
Rather, it is just executing, executing, and interacting with the environment.
So for premium developer blocking focus task, we're seeing it's very benefiting.
But also, case three are just be able to have the ability for people to modify it and fine tune on top of it, allow them to make it better for their own workload.
And this is definitely happening today as well.
Can you just explain what the...
licensing term is for the most recent open source models compared to the past and why you think they're doing that?
Oh, yeah.
So historically, the OpenWay model are just like Apache 2, like our software, which is like take it, modify it, do it, whatever you want with it.
Here is a gift to the world.
And then recently, the model lab are trying to understand a way to economically fund their own model development.
After all, model training and research and the data are very, very expensive.
So we have been starting to see terms, even to the Llama days, for when Meta was releasing Llama, they do have a term of if you're a daily active user or like...
annual recurring revenue exceeds some threshold, please enter into a commercial agreement with Meta specifically, right?
I do remember that the numbers were like specifically chosen at that time that you could go find it was like two companies in the world that like fit the definition that they had excluded from their license.
Yeah, exactly.
But like people have taken a hint from that, especially now the labs are trying to figure out a way to economically fund it, especially when they're open source model.
Everybody can just take it and run it themselves, whereas nobody will use their API anymore in many cases, while their API currently still taking up shape, right?
And now we're seeing a very healthy ecosystem development, starting from even Minimax recently, when they're releasing the M2.7 model, they have a term specifically focusing on usage.
And Kimi initially also has, like, if you have derivative works, like this is kind of big news.
back then with Fireworks and Cursor about how they built on top of Kimi model.
And it's, if I could just expand on that a little.
I don't think it's greed, at least what I've seen from open source model labs, right?
Open source models, really what we're talking about are open weights, right?
And it's just not software, right?
Like an AI model is not software at the end of the day.
And so open source software used to be supported by people donating their time or big companies kind of authorizing their employees to donate their time.
So it was sort of like a bulk in-kind, you know, donation of people's time.
That really doesn't work at AI, right?
Like I can't just like go home at night and like train a frontier open source model with friends for fun.
Like we need millions or billions of dollars of computing resources in order to do it.
So I think it does support your point.
Like, obviously, there need to be economic incentives and there need to be funding mechanisms in place.
Frankly, I think even more so with Chinese models than with domestically produced models, right?
If there's no source of funding for Moonshot to continue.
to train models, like, we know where the funding will come from instead, and it's not like something we, right, you know, it's government and things that, like, are actually worse for us, I think.
So, like, I think you've raised sort of an interesting point that this is an important economic structure, and, like, I think this means we'll see more of this in the future.
Would you agree with that?
Yeah, it's really about sustainability in the end.
It's about how do you make sure that all this initial capbacks almost to train the model.
fail again and again and train the model again.
Like, how do you really pay it back and how do you make sure that there's enough confidence and funding from everybody involved to go to do the next one, right?
And I recently heard someone, I recently heard someone making an analogy to this, to the pharmaceutical industry.
It's almost like, how do you make sure that the R&D process of new drugs are?
properly funded and this proper sustainable method to making sure that people are willing to take big risk, big bet, to go to do research for new drugs.
And then later, because they know there's a economic incentive in the end when the new drug leads to the market, a portion of those, of course, like besides just distribution channels, right, a portion of those.
revenue will flow back to continue to fund the next R&D effort.
And this is where we're kind of seeing similar to the model development now.
That's a really interesting analogy because it's like once a drug, a molecule is released, you have the strongest possible control, which is nobody else can manufacture it at all.
It's like the most closed possible source.
It's like a secret.
But, yeah, like in the case of models, especially open source models, you know, once it's out there, anybody can take it, use it, extend it, et cetera.
So having, yeah, so having some economics attached to it probably does make sense.
Well, actually, that raises a question for me, too, which, Matt, you were alluding to this earlier about how different open source models are from...
the dynamics of open source software maintenance.
When it comes to open source AI, what actually needs to be maintained?
Is it the infrastructure around it?
Do the models themselves need maintenance at all?
Just kind of what are those dynamics?
Because I think even the developer behavior around it is pretty different.
Yeah, I mean, Simon sort of said this already, but maybe I'll just expand it a bit, which is, you know, you...
you see the results of a big training run, you know, where training now means it's pre-training and then, you know, RL kind of, you know, mid-training or sort of post-training on these things, or, you know, pre-training SFT, RL, right?
Like it's sort of a complicated pipeline.
We see only the result of this at the end.
And the numbers are big.
You're like, oh, you know, this was a $100 million training run.
But what you often forget is like there may have been five failed training runs, you know, large scale.
failed training runs before he even gets enough.
Yeah, the blood slant tears.
Yeah, exactly.
One of my favorite artifacts, maybe we can even track down the link, is one of the early llama models.
They published the whole conversation log between the people who were babysitting the training clusters while the models were training.
And it's so funny.
It's just like, oh no, everything's gone wrong.
Like chaos, like panic.
And then the next comment is like, okay, we solved it.
Everything's okay.
Cluster's up, loss is going down.
There's a lot, a lot, a lot that goes in behind the scenes before these models get released.
You know, once they're out there, I guess it's a little bit more in your zone, you know, to kind of make sure it's like operationalized.
Oh, yeah.
But this is also a very interesting point.
Once it's out there, it's a whole community effort trying to opt in this model because the model is trained on a given type of hardware, on a given type of architecture.
But when it's out in the wild, everybody has different cluster topology and use cases.
It's about how to turn like a use case of one now to a use case of almost infinity.
Now you have people trying to adapt it to the edge devices and people trying to run it at largest scale ever.
Adapt it, making sure it runs for voice agent, but also for coding agent, which are entirely different kind of use cases.
So this is a whole community effort trying to further optimize, specialize.
and making sure the running of it is reliable and continue to be able to optimize against it.
So that's a whole village later throughout the open source to make sure it's improved.
And that's cool, because that really is like open source software.
I mean, and this is what you do, but this is like anybody can contribute and make these better.
The fun thought experiment is if GPUs...
you know, dropped in price by 99%, right?
Like if GPU-based compute actually became, you know, kind of cheap and widely available, like then do we get back to a kind of a real open source world where, you know, one person sitting in their basement or 100 people working their free time can like come up with something new, try many of these sort of model training paths that, you know, that like are in the queue somewhere at one of the big companies and, you know, kind of see, you know, really expand and advance the field collectively.
Yeah.
Well, I mean, this relates to what you were talking about at the beginning.
It's like, you know, at the beginning, the amount of compute you needed to be at the quote unquote frontier was negligible.
And now it's just like it's enormous.
And how do you get that, you know, back to?
consumer parity again.
I've said this on the podcast before, but I'll keep saying it.
AlexNet, first, you know, kind of like neural network to run on GPUs that we care about, ran on two GPUs.
And that's not like there are no missing decimal points or commas in there, literally two.
Now that would get you literally nowhere.
Yeah.
So I guess this relates to another question that we've had, which is inference has gotten harder over the past 18 months because of this combination of scale and diversity and the kinds of models that we have.
And of course, like agents that are doing increasingly long running tasks.
So what makes open source in this world and in this scenario not just like nice, a nice thing to have, but absolutely necessary?
Yeah, so scale comes from a few points.
It comes from...
whether or not you can run this gigantic model on a data center rack.
But also it comes from, can you optimize it to the extreme point, to the speed of light, so that you are getting the most value out of it?
But also you need a whole community and collaboration and partners of effort to validate it and making sure that there is no that last remaining bug that appear like 0.0001% of the time.
So this is where scaling...
up the deployment and making sure more people are running it as like largest footprint possible will ensure everybody's experience of running this model is better.
And this is where it kind of, this is why open source inference is the current leading way right now instead of closed source inference engine.
And frankly, right, all the, a lot of the open, a lot of the open, sorry.
A lot of the inference cloud and API as a service today do use and leverage open source inference engine under the hood.
And the reason to do that is there's just so much battle, tested and learned recipes and things that they can build on top of.
And this is where open source is the current leading way to run models.
Something that I find kind of funny and like looking back at.
Not just the history of VLLM and Infract, but also a company like Open Router or even Olama.
All of these different teams kind of got started around 2022 and 2023.
Some of them even before ChatGPT, in your case, built around open models.
And what kind of special thing do you think was happening at that time, just like in the world of research and in the world of AI, where we see these companies now and we look at them and you think of VLLM as mission critical or Open Router as mission critical?
And what do you think was going on at that time where even before we had, you know, big consumer use case.
We had, you know, teams like yours that were building out these things and kind of how did you guys know?
I guess this is the question.
I think there's two parts to this.
Our team always have an open source kind of angle.
We're from UC Berkeley, a long tradition of open source software and our system research principles.
But also just curiosity overall.
Everybody was so curious about how are these AI models benefit the humanity and how can we use it better?
And this is where open source is where you meet the most mission aligned people together.
Open Router, Olama, as you mentioned, they're all great partners and friends in the ecosystem because we're there to understand how everybody else can leverage AI model better and making sure they have a good use of it.
Speaking of, you know, leveraging open-source models really well, so I think another thing that was in the news really recently was hugging face using a Chinese open-source model to help contain a cyber attack carried out by a rogue, unsandboxed open AI model that was being tested.
So, like...
What can we take away from something like that?
And can you just first, I mean, walk us through your impressions and understanding of what happened and then kind of what you think the takeaways there are?
Yeah, like this kind of goes back to our previous point about control.
So for the Huggy Face incident, they break it down pretty well on their website and blog.
So really thanks for their transparency there.
In the end, it's about...
all the closed proprietary model APIs, their guardrails are a little bit arbitrary, but also very difficult to enforce.
That means they have so much false positive in the guardrails that will have legitimate use cases just being blocked.
And this is like an evergreen problem, even in the social media days.
How do you design content filters correctly and be able to do the moderation correctly?
So if moderation is never solved, which is going to be very, very hard, then there's always a place where you have a model where you know and trust that you are publishing to and be able to use from.
So Hugging Face, they have to use OpenWay model.
But in the future, we'll also see for the trusted use case, people will go to OpenWay by default because that is where you know for sure that the guardrail is lessened or you can control your guardrail for trusted use cases.
Even for us, like this is actually applied to us today, where a lot of the anthropic models are banning frontier AI user research.
And then when we're studying GPU kernels, even as an invalid memory access error, we are triggering the red line.
And so a lot of our developers within Infrax and for VLL are like retreating from using Fable 5 because you have a two-hour job and you trigger the red line, which is false positive, and then you have to lose all of your work.
And so a lot of our developers are using like Kimi K3 today even, just making sure, because it's similar quality and it has a guardrail that makes sense to us, there it goes while using it.
I think your social media analogy is a really apt one.
Yeah.
Because like in both cases, what's kind of happened is you've taken like distributed human activity and kind of centralized it in one place.
Like before social media, people talked to each other, right?
And published articles that like messaged each other on the message boards.
But then you centralize it all under one kind of like profit-seeking enterprise.
And all of a sudden, the incentives will change like a lot, right?
We have a specific carve out for social media, which is you're not responsible for what people say on your platform.
And that allowed the moderation problem to be tractable.
Yeah.
Right.
It's like, OK, we're going to do our best to eliminate obviously illegal things, you know, like things that are going to degrade experience on the site, you know, you know, in really bad ways.
But like, you know, if it's on the edge, you know, it's something that's maybe out of fashion to say, but not illegal.
Yeah.
Like we just can't police everything.
We can't be the police of like all human communications.
I think something similar is sort of happening in AI, right?
Where a lot of work, not just talking, but work is kind of being consolidated in one place.
If I need to write code or create a spreadsheet or get advice, you know, health advice or like anything, it's all happening in like one or two sites.
And they don't have that carve out, right?
Like they don't have that exemption of like, hey, we're not responsible for what actually happens on the site.
And like some of them, especially, you know, anthropic, like is kind of going further than even what would be sort of like legally required.
And they're sort of taking ethical stances on these things, which may be right or may be right.
That's sort of their decisions.
How do you actually do this, right?
Like the problem is actually bigger than just communication because it's like actual work and action.
And you don't have this exemption and you have this sort of layer of ethical stuff on top.
So, yeah, I'm just totally with you.
And it seems like some of these companies have gone, like, erred on the side of caution, which is probably a reasonable thing to do, but very frustrating as a user, right?
Like, you know, I saw online some...
translation attempts are blocked because they think the particular combination of like language and content is somehow like triggering like politically sensitive topics and you know things like this.
Simon, we're nearing the close of the conversation, and I just wanted to take a step back a bit and ask just about Infraact and, you know, running the company.
And I know that Ian Stoick of Databricks is an advisor and a co-founder of Infraact, and I'm just curious what you've learned from him in terms of taking an open source project like VLLM and building a company.
Yeah, Ian, as a co-founder, he has always...
being thinking about open source and where, how do you support open source better?
And then now with experience from Databricks and AnyScale and even Arena, which is a public platform.
And now at Infrared, his focus really stemmed from how do we build such a world where this software is being used by so, so, so many people on such a critical project.
How do you make sure they're getting the best quality and where the value comes from, right?
If the open source movement and open source inference engine is a starting point, then what are the gaps that we can fill as infract?
And then what are the gaps that we should absolutely put in?
Yang has always been open source first.
So for whatever we need to build, we would like to build them in open source.
And then for whatever value we can deliver, we want to really close the last mile and making sure that customer and partners are getting the best out of it.
Just, I guess, to close this out, five years from now, do you think open-source AI models, have they closed the gap with frontier models completely?
Are frontier models always one step ahead?
Kind of how do you see that shaking out?
Five years.
Five years is a lot.
All right, all right.
One year.
One year.
Yeah, five years, who knows?
We're all going to be like just floating around in our Wally pod on our spaceship.
Exactly.
Progress, right?
Yeah, yeah.
For me, really, at this point, there's kind of a point we haven't talked too much about is what really differentiates open-way model from closed-way model, right?
In the end, there's not much differentiation.
It's more about the distribution strategy and go-to-market strategy.
And the capability-wise, I don't really see a big gap, not even today, because for how these models are coming to being, they're really starting from the first principle, right?
You have a compute cluster, you have training data, and you have brilliant researchers that group together and really to build this.
amazing artifact that is this pre-trend model and then later our old post-trend model and that the world can use.
But if you look at the ingredients, right, one of the most important parts is just the data.
It's about who gets what data and then what are the environment you are building to let the model improve on itself and make better.
One of the very useful benchmarks that we have on Arena for case three has been front-end coding, right?
That means for Moonshot, they have built some of the best environment for front-end coding, right?
They have published amazing demo on the ability for this model to code and then see what the render is and then kind of continue looping and this iterative process.
Now, this is about their environment to improve the model.
It's not about just source data.
It's not about where to get the data from.
Rather, it's who can build the best environment and who can make the most sort of optimization and algorithmic choices to leverage all those learning from this environment.
So the next year is all going to be about that.
It's about how open-way model labs are differentiating and really getting the model to meet the real world and have this kind of...
what people are popular today, like recursive self-improvement almost to really improve the model overall.
And so really projectile in a year, there's not going to be any difference.
Yeah.
And you've used this term brilliant researchers a few times.
There are brilliant researchers everywhere in the world, clearly.
Why do you think, you know, in the U.S., all the smart researchers are working on closed models and in China, all the smart researchers are working on open models?
I mean, from my point of view, they are attracted to interesting problems, not necessarily on the open or closed stands.
But rather, but however, open way model does give people a really, really good boost on the impact of such models.
So that is like a plus.
And I think all the brilliant researchers are attracted to how to improve the model overall.
Actually, one interesting point about this, maybe fairly technical for this Kimi K3 model, is they removed a rotary positional embedding.
So Rope has always been there for a lot of the Transformers model.
And guess who removed it?
It's the inventor of Rope.
Oh, that's okay.
Yeah, like, Jenning, he wrote the first paper introducing Rope as a concept.
And then he now also wrote...
the explanation of why you don't need it as part of the technical report in this case remodel.
So like when we read it, it's like really come full circle is you have all these brilliant, humble researchers that are able to really study how this works and really study the secret of training and pre-training and share it across the whole world and recognizing and iterating on their past, right?
So really a miracle, I would say, for this model to come alive.
It's so interesting.
Like, you know, AI is in this funny zone where empirically it works incredibly well.
But then you go ask the theorists and have like no idea what's going on.
Right.
And so like you have these kind of iterative things where when you go read, you know.
primer on transformers.
You read about positional embeddings and why, or positional encodings and why it's so important because otherwise you can't sort of like understand meaning.
And then it turns out once you understand one level deeper because we've been doing this for a few years and you have all these smart people like, oh, actually you don't need positional encodings, right?
Like it's actually, you know, simpler actually is better.
We didn't talk about distillation much so far in this conversation, but I think it's very relevant to this.
I have just one question, which is like, I'm not going to ask like, is distillation happening?
I think this is kind of speculation on the part of everybody, you know, in the world.
But like, you work a lot with these Chinese labs.
Do you think distillation like is a critical component of what they do?
Or like, are they kind of just doing good work and, you know?
distillation if it's done is sort of an incidental part of it.
I will lean to the latter part specifically.
As I mentioned previously, environment matters so much today.
So these are our environments, right?
These cannot be distilled.
Like you don't have other people's.
environment to really distill a copy from.
It's about constructing it, also understanding the learning process.
You cannot distill how the model learns with the environment.
A lot of these are just not doable today.
There are things potentially you can do with rewriting the data sets, right, making better pre-training data.
But again, you can do it with any models.
Any models that are going to follow instructions are going to be useful in terms of utility there.
So I really don't think from currently what we're seeing, this is a big cornerstone of what's powering the progress today.
In the end, what's powering the progress is still just really smart people with very interesting algorithms, data environment, and they will produce, of course, compute, they will produce the models.
I think it has really interesting policy implications.
I tend to agree with you, by the way, that, you know, we have smart people.
people everywhere working on a bunch of smart things and it's not about, you know, distilling data from any one place.
It has really interesting policy implications, right?
Because it doesn't, you know, it's almost tempting if you're sort of, you know, in the White House to say, oh, sure, we'll just turn off distillation.
All our problems will be solved.
you know, I think it's more the case that they're just, you know, smart people doing interesting things.
And so it's like, how do we kind of like adapt to that?
Yeah, and creative innovations, right?
Like one part in my essay, we kind of mentioned that open source and open way really helps innovation because it's set out this racetrack where everybody can learn from each other and see what each person, like each player is in this racetrack.
And then you have to improve.
stay on the shoulder of each other kind of to improve yourself.
So that is where everybody can move forward faster.
Yeah, and one thing we're looking for a lot from an investment standpoint is people doing more open source model training all over the world, not just in China, not just in the United States, but all over.
Right, because, you know.
You get that sort of magic of collaboration when everybody's doing it and achieve global harmony and all those things.
Well, I think that's a good note to end on.
Matt, Simon, thank you so much for joining us.
Thanks for tuning in.
Thanks so much, Nicole.
Thanks, Jim.
Thanks for listening to this episode of the A16Z Podcast.
If you liked this episode, be sure to like, comment, subscribe, leave us a rating or review, and share it with your friends and family.
For more episodes, go to YouTube, Apple Podcasts, and Spotify.
Follow us on X at A16Z and subscribe to our Substack at a16z.substack.com.
Thanks again for listening and I'll see you in the next episode.
As a reminder, the content here is for informational purposes only.
It should not be taken as legal business, tax, or investment advice or be used to evaluate any investment or security and is not directed at any investors or potential investors in any A16Z fund.
Please note that A16Z and its affiliates may also maintain investments in the companies discussed in this podcast.
For more details, including a link to our investments, please see a16z.com forward slash disclosures.