- The panel consists of investors from Local Globe, Mosaic Ventures, GV Google Ventures, and Chapter 1.
- They have been investing in machine learning (ML) and artificial intelligence (AI) for the past decade.
- Their focus is on B2B software, but they also invest in consumer-focused businesses.
- They are interested in the infrastructure and application layers of AI and ML.
- They see value in bringing disparate sources of data together and creating new ways of interacting with products.
- They believe that the enterprise market offers more immediate value, while consumer applications may take longer to establish their business.
- They have made investments in B2B software with a focus on customer service.

- The speaker discusses the importance of customer service and the potential of using GPT-4 to improve customer experience and reduce operational costs.
- They also mention the potential of using data and analytics in customer success to identify risks and opportunities.
- In the B2C space, there is excitement around the democratization of services like executive coaching and the use of AI to provide creative experiences.
- In the B2B space, two areas of interest are guardrails for language models to ensure appropriate responses and the transformation of analytics within companies.
- The speaker also mentions the focus on democratizing access to education and professional skills through cognitive task analysis and AI.
- Infrastructure deals at a foundational level are also highlighted as important in the industry.

- The focus is on how Fortune 500 companies can integrate technology into their daily operations, particularly in areas such as customer service and cybersecurity.
- Open source software is being built and there is debate on how to monetize it.
- The strategy for monetizing open source software involves offering a managed cloud version and premium support.
- The success of open source software depends on its adoption by developers and its integration into enterprises.
- It is still early in the market, and the specific requirements for commercializing open source software are not yet clear.
- The founder and their ability to iterate quickly are important factors in AI investing.
- There will likely be successful horizontal and vertical businesses in the open source space.
- Open source software allows for transparency, collaboration, and organic growth, but there are also concerns about regulation and policy.
- Open source software aligns with the ethos of the blockchain world and the promise of democratized access to knowledge.

- Open source software allows businesses to benefit from the technological revolution by accessing and integrating existing software into their operations.
- The EU is expected to be a regulatory leader in AI safety, with the AI Act creating a compliance market similar to GDPR.
- Investing in companies that can solve compliance issues with the AI Act is challenging due to the uncertain nature of the legislation.
- Trust and safety, as well as data privacy, are concerns for enterprises deploying AI, creating opportunities for regulatory and employee policy solutions.
- Language-specific models may be less interesting for investment compared to multimodal, multilingual models.
- Europe has a competitive edge in AI due to its talented academics, research associations, and successful startups.
- Europe receives a significant amount of investment in Pre-Seed, Series A, Series B, and Series C stages.
- Companies in Europe are working on inclusive AI models that consider different languages, cultures, and biases.
- GDPR has already influenced enterprises to prioritize data privacy and compliance with existing laws.

- The GDPR regulatory framework has encouraged existing enterprises to prioritize data privacy and comply with existing laws, creating a strong foundation for integrating AI into businesses.
- European enterprises have the potential to adopt AI technology and leverage its use cases more than the rest of the world, giving them a competitive edge.
- AI can benefit businesses across the spectrum by enabling a culture of experimentation, allowing for quick deployment of internal products and proving value.
- Marketing and developer support are currently the most adopted use cases for AI in B2C and B2B respectively.
- Businesses should explore lesser-known AI companies and partnerships to fully understand the options available to them.
- AI reduces grunt work and improves productivity across all aspects of a business, from basic tasks to C-level training and coaching.
- Traditional industries are increasingly adopting AI, leading to significant productivity gains, such as a construction tech company increasing productivity by 26% in one quarter.
- Data quality remains a challenge for financial institutions and other enterprises, but there are tools available to help with data cleansing and preparation for AI models.

- There are tools available to help with the data life cycle, including data cleansing, ingestion, transformation, loading, and modeling.
- The modern data stack is becoming more standardized, making it easier to build applications that clean and standardize data.
- While automation is increasing, there is still a need for consultants to manually scrub and find use cases for data.
- OpenAI and other investors are looking for ways to automate the process of obtaining clean data.
- The data space is evolving, and there is a need for companies that can add structure to data lakes and extract necessary data from various sources.
- Responsible investing is a difficult question for investors as they balance the need to return capital with ethical considerations.
- Investors may pass on businesses that do not meet their ethical standards or may become outlawed in the future.
- Ethical burdens are acknowledged in early-stage investing, and the goal is to create net positive externalities.

- Early stage investing involves taking risks on potential business, market, and team
- Ethical checks and understanding of regulations are important for AI investments
- Collaboration with regulators helps shape policy frameworks
- Placing oneself on the XY graph between monetization potential and regulatory compliance is crucial for early stage investors
- The panel concludes and thanks the audience for their attendance.

- Panel of investors from Local Globe, Mosaic Ventures, GV Google Ventures, and Chapter 1
- Focus on investing in machine learning (ML) and artificial intelligence (AI) for the past decade
- Interest in B2B software and consumer-focused businesses
- Focus on infrastructure and application layers of AI and ML
- Value in bringing data together and creating new ways of interacting with products
- Enterprise market offers more immediate value, while consumer applications may take longer
- Investments made in B2B software with a focus on customer service
- Importance of customer service and potential of using GPT-4 to improve customer experience and reduce costs
- Excitement around democratization of services and use of AI in B2C space
- Interest in guardrails for language models and transformation of analytics in B2B space
- Focus on democratizing access to education and professional skills through AI
- Importance of infrastructure deals in the industry
- Integration of technology into Fortune 500 companies' operations, particularly in customer service and cybersecurity
- Monetization of open source software through managed cloud version and premium support
- Founder's ability to iterate quickly important in AI investing
- Concerns and opportunities related to open source software
- EU expected to be regulatory leader in AI safety with AI Act
- Trust, safety, and data privacy concerns in AI deployment
- Europe's competitive edge in AI due to academics, research associations, and startups
- GDPR's influence on data privacy and compliance
- Potential for European enterprises to adopt and leverage AI technology
- AI's benefits across businesses for experimentation, deployment, and productivity
- Marketing and developer support most adopted AI use cases
- AI reduces grunt work and improves productivity in various industries
- Data quality challenges and tools available for data cleansing and preparation
- Standardization of modern data stack for easier data cleaning and standardization
- Need for consultants in data scrubbing and finding use cases
- Automation of obtaining clean data being explored
- Need for companies to add structure to data lakes and extract necessary data
- Ethical considerations and responsible investing in AI
- Collaboration with regulators to shape policy frameworks
- Importance of balancing monetization potential and regulatory compliance for early stage investors
- Panel concludes and thanks audience for attendance.

Speakers: 
Jamesin Seidel, Investment Partner, Chapter One
Luna Schmid, Partner, Google Ventures
Chandar Lal, Principal, Mosaic Ventures

Moderator: Ash Arora, Partner at LocalGlobe

The Oxford Generative AI Summit is a two day multi-stakeholder expert convening hosted at Jesus College, University of Oxford to dive deep into the use cases, implications, and future of generative AI & society.

For more info: oxgensummit.org

So we are sort of, after a lot of back and forth we are just gonna stick to this completely my class and happy to start.

I look after blockchain and Gen AI investments at early stage.

We're just gonna kind of go around and do a series of introductions as well as everybody would tap into a little bit of where we are spending time.

And this is not a very sort of quintessential formal panel.

We love sort of interaction from the audience as well.

And happy to save some time towards the end, hoping that somebody can just give me a timer mark.

Hello everyone, hopefully you can hear us in the back.

My name is Chandra, Principal at Mosaic Ventures.

We are an early stage venture firm based in London, investing all over Europe, and have been backing machine learning founders for now the best part of a decade, and that's only accelerating, as you can imagine.

Much of our focus is B2B software, but we do a bit of consumer stuff here and there as well.

We've been investing through the kind of waves of NLP, computer vision over the last decade.

And it feels like the rules about what we get excited about has changed a little bit as We're kind of entering this new era of AI.

So we made 3 investments so far this year.

None of them announced publicly, but I'll be telling you a little bit about them over the next hour or so.

Well, hopefully I can also be heard in the back of the room.

We're an independent, non-strategic investment arm.

And we are also early stage investors who have been investing in ML and AI for the better part of a decade.

So a lot in common here across the table.

We historically have invested roughly equally across life sciences, enterprise, and consumer-focused businesses, and we've got offices across both the U.S.

And here in London where we cover Europe.

My name is Jameson, and I'm on the investment team at an early stage fund called Chapter 1.

I cover consumer, B2B, and enterprise across all verticals and mostly all the AI startups at this point just because that's where the industry is.

It's a US fund, but we have investors across Los Angeles, New York, and I'm in London.

Quick note on my background, I have a data science background and a sports betting background.

So I was at Twitter for 5 years as the technical lead for the platform manipulation team, so focused on build out ML models and systems, check bots and election reference, and then before that as a professional sports bettor for 3 years for a high-stakes sports betting team.

So just from my end as well, I hope you're very similar to the rest of the panel.

We've been spending the sort of the better part of the last decade looking at ML based companies as well.

We've deployed a little over 700 million in Europe alone in the last 6 years, backing a majority of startups.

I have about 50 AI ML companies that are working for you all.

So, yeah, I'm happy to kind of dive right into it.

I'd love to know where all of our investors are currently spending time in the AI ML space.

We spend so much time in front of our computers that like, even interaction is So weird nowadays.

Gosh, this is a lot of questions, a lot to unpack.

But so I would say there's, of course, the really obvious areas that generative AI is changing with, you know, summarization, content generation, and specifically with startups, you have the horizontal startups that are kind of building tools across the stack, whether that be in observability or kind of, you know, all these tools for to bring large language models to enterprises or consumers.

And then you have the vertical sections, whether that be large language models in law, insurance, or whatever vertical it may be.

So spending time across all of those, I would say probably 50-50 between B2B, so companies that are building tools to sell into businesses, and then B2C, straight to consumer.

And then within that have been really focused on personalization, edge computing, and then and also kind of creative infrastructure.

And to keep my answer short, I'll cut it there and then we can always get into it later.

2 main areas where I spend my time are 1 on sort of the picks and shovels layer in the infrastructure side.

So my partner, Vidoo, actually just announced a deal on context AI earlier this year that's on the analytics side of things, obviously early stage because everything in this world is early stage right now.

And the other area where I spend most of my time to be honest is on the application layer.

So I think all of us here probably agree that there's gonna be a massive amount of value that's gonna be accrued here across the board of generative AI.

I think the challenge here though is that not all of the value actually will accrue to the net new startups.

So I think that on the paradigm, we never saw AWS coming out with a consumer facing layer.

And we see that with open AI, they've obviously got massive amounts of revenue there.

We also see a lot of existing software players that are able to leverage what they're seeing in generative AI and add it onto their products and create massive amounts of value.

So I think then what do you think about what accrues to net new startups?

I think we're worst focused on is where with AI, the application layer looks fundamentally different from how you used to interact with products or software tools in the space.

And then secondly, where data is different and more disparate.

And so some examples of that is where, so for example, if you think about how you thought about sales previously, you had a bunch of outside-in data that you would get from the Zoom info of the world, and then you would also get a bunch of data from your internal tools like Salesforce.

I think there's a lot of value in bringing those disparate sources of data together, and that exists across the board in plenty of different areas.

And so that's where we're spending a lot of our time on the application layer.

Yeah, I guess if we continue with that framing of tools and infrastructure on the 1 hand and applications on the other, Looking at tools and infrastructure, there are some spaces where it's really hard for startups to create value.

NVIDIA's printing cash because it's the bare metal that everyone's building on.

It's hard for a startup to compete with that kind of efficiency and economies of scale.

When you move up to the foundation model providers, you've got this arms race of capital that's getting raised by the open AIs and the coheres and the anthropics of the world.

For an early stage investor to bet on a new entrant there feels very tricky, especially given the valuations they're seeing.

Looking at the other tools that you can build in, there's this question of where does free open source software win, and where can you build something that is a commercial proposition?

So, you know, we've been asking ourselves, will a vector database be the next MongoDB?

Or is pgVector gonna do the job for everyone who just needs to stand up an application quickly?

So there it's, again, for us at least, very hard to make a call on how you actually build a company versus build an open source project.

So, I think similarly, the application layer is where we spend more of our time.

And there, like, the key questions are very case by case.

A lot of it is retention, like, loads of users are getting excited by generative AI products.

They're getting on board, they're trying stuff.

You know, Jasper AI acquires hundreds of millions of users within the enterprise very quickly.

Does it keep them or are people just experimenting and sharing off in a week?

That's 1 of the big questions that we're asking.

So anything that's deeply embedded into the enterprise is therefore attractive to us.

And anything that's just an entirely novel UX that people stick around on for a month or 2, we kind of start to get excited.

I think at our end as well, I like to actually call it the T flow of AI.

So this is very similar to what I think all the panelists here have touched upon as well, which is across the entire spectrum, everything about semiconductors, like everything about NVIDIA.

You have, like I said, on the infrared middleware side, you have data lake houses, you have DevOps, different kind of vector databases, observability companies that are coming out all the way up to like a B2B or a B2C consumer application where that touches across different industries.

So that is where like the upper spectrum of the peak comes in.

So yeah, I think that's pretty much how we are sort of evaluating the space as well.

And definitely on the enterprise side, a lot of value has been expected to accrue over the short period of time, while the consumer applications might take longer to kind of find the defensibility of their business and then go further.

Awesome, so I think we've all touched upon a little bit around consumer applications as well as enterprise applications.

Maybe it's worth double clicking in a little bit on sort of the B2B as well as the B2C side of things and what are the different kind of use cases that all of us are seeing, starting with you.

Yeah, so we've made a couple of investments in B2B software that sort of generated AI native, if you will.

The way that we're kind of framing this is, first of all, where is there really hard, measurable cash ROI, especially in this macroeconomic environment?

And then where can you build really deeply sticky software?

So 1 space that we spent a lot of time thinking about and made an investment in is customer service.

I'm sure everyone in the room has been on the phone to their bank or to an airline and they've gone through the painful, you know, press to speak to this person about this kind of problem.

Let me give you my national insurance number, I know you don't understand it, it's a pretty nightmarish experience, and that costs a lot operationally in terms of the number of people you need to run a call center, and it costs a lot in terms of customer experience, you know, the number of customers who just walk away with a terrible UX.

So we found a team in Berlin that's cracked this problem using GPT-4, using a proprietary layer they've built themselves for text-to-speech and speech-to-text, and with an architecture that seems to scale really well across different verticals.

And it took a little bit of benchmarking to figure out like, are these models and is the fine tuning they've done really distinctive and allowing you to deliver way more than you could just get using an LLM directly?

And commercially, that company has been taking off in what is traditionally a slow-moving enterprise sale precisely because they can show the customer just how much OPEX they save on day 1 and potentially how much revenue they uplift as well.

And so that kind of no-brainer enterprise sale is definitely an exciting 1 for us.

So I've given 1 example already of a space in B2B where we've seen a lot of net new businesses.

Another that we look at is similar to the customer support, but actually in the success side.

So there was sort of a gen 1 of the game sites of the world that have built something within customer success.

Where do we have expansion opportunities?

And again, thinking through the trend of using data that we have now to be able to leverage that and get new insights from it.

I think customer success is another really interesting space where we're spending a lot of time sort of thinking through, we can now marry product analytics, sentiment from email, sentiment from our CRM, and all sorts of other outside-in signals to sort of understand where do we have risk in our existing customer base, where do we have opportunity, And then more of a prescriptive layer, what do we do with that?

And what is most likely to prevent this customer from churning or encourage this customer from expanding?

So that's 1 area on the B2B side that there's a lot of opportunity.

And on B2C, I think a lot of the opportunity that I find really exciting is actually around the democratization of things.

And so if you think about executive coaches, historically that has been something that's been limited to your C-level execs, your VP level, someone who you can actually spend that significant investment to get a coach in place.

I think there's probably an opportunity to think through, how do we leverage training data to then bring that to someone who's more, becoming a manager for the first time, or someone who's really just in your rank and file, and bring that sort of coaching to someone, and bring the answers that would help them sort of develop earlier in their career.

There's plenty of spaces, I think, where we're all spending time, and we're only just scratching the surface here.

I'll go for B2B first, and then if I'm not 25 minutes over, we'll talk about B2C.

But Yeah, so for companies that we're interested in, that are building out tools and software that are selling into other companies, 2 areas that are particularly exciting for me.

So they talked about customer service bots and a bunch of chat bots and pretty much all this tooling.

But we also have to think about how to make sure that the models are responding in an appropriate way.

So, like, 1 of my favorite examples is Kevin Rose from the New York Times.

He did an article about him talking to a chat bot, and the chat bot declared his love for Kevin Rose and then asked him to break up with his wife.

And so that's obviously, if you're talking to a customer service bot and you're trying to return a shirt but then you ask it something random, you'd also don't want it to be responding to in certain ways.

So that's kind of a silly example but a more serious 1 is, you know, examples around there, for national eating disorder, in order to day, they put a chatbot out and it started giving diet advice to people that were interacting with it.

And so there's a lot of questions around guardrails and companies that are trying to solve the guardrails around large language models so that if a business does, you know, is implementing 1 of these models that they also have the appropriate guardrails so that they don't get in a situation where the model's telling the customer something that it shouldn't be telling it.

And then kind of a second example for B2B, it's analytics times large language models has just been totally insane.

So to give you kind of, to talk about personal experience, so I joined Twitter in 2016, and 2 months in I joined the spam team and focused on bot detection, but

my first 2 months, I was, my first project I was asked to do a tweet analysis.

So this is 2016 after Twitter had IPO'd and they asked me, how many tweets are in each country?

You know, how many, what are the trends around?

And like, it was just totally insane to me that they didn't know any of these questions.

Everyone was really wowed by it because it was all this new information and I think we're now at an age where data and metrics like that is a lot more accessible.

You have amazing companies like Mixpanel or Amplitude where people can find their daily active users or anything like that.

But now we're shifting into a world where you'll just be able to, in your business, you'll just be able to ask it, how many daily users are there?

And all of that data will be mapped and graphed internally.

And so the user interface internally at companies around analytics is just going to be transformed.

And then I'll quickly note B2C, because I feel like I've been talking a long time.

But on the B2C side, of course, like just chatGBT, OpenAI, laying out chatGBT to the world was like changed everything overnight that they just made it a consumer app and it just crossed a billion users, or a billion, sorry, not a billion users, a billion revenue, yeah, yeah, a billion in revenue.

So obviously that changed the consumer side, but we've been spending a lot of time on human enhancement.

And I know that I just listened to the media panel and they were talking a lot about kind of AI within the creativity world.

And so just to define human enhancement, digital technology that's augmenting or changing or kind of modifying the human form.

So for Snapchat has around 200 million daily active users and with Snapchat 63% of snaps that are sent have some sort of filter on them or AR.

It's a really insane number and it's a lot of, you know, how people are using apps these days, they want some sort of like different experience or creative experience.

And what's cool about AI is, in general, AI is that now able to more easily provide and kind of create these new fun, creative experiences for consumers.

I think at our end as well, very similar to I think what all the investors have touched upon here as well.

We are also very keenly focused on democratizing access to a majority of the users in the world.

So that is coming from, like you said, a perspective of a lot of knowledge and sort of professional skill set concentration at the top 1, 5, 10 percent layer across the world.

And that has resulted in complete mismatch in terms of opportunities and access the majority of the educated world gets to have.

So That is where this sort of technique of cognitive task analysis, which was apparently invented by the US military about 3 decades ago, is 1 of the things that's used across C-suite level exec coaching.

It is used across a lot of top universities and how professors are delivering those classes, et cetera.

And there's this actually very famous case study on mob mentality, if anybody's from Harvard here as well, which became a course now in that curriculum, where they are talking about how Hitler back in the day was able to mobilize such a large country to actually stand up for something that not everyone really believed in.

So those are the kind of things that cognitive task analysis addresses.

And with AI, you can not just kind of focus on guardrails around such a technique being taught to the right people, but also helps a large number of people realize their true competence.

So that was again like a very interesting space that we've actually invested in a company that we found based out of Oxford, actually Oxford-Cambridge team, that we have done recently as well.

And then I think on the infrastructure side as well, there's a lot of, like you said, Chandra, there's a lot of deals happening at a fundamental foundation level, where there's, that's a very capital intensive business, right?

But if you think about it, like back in the day, AWS, Google Cloud, etc, they all captured a very big market share of the fundamental in front of the internet that we all use today.

And it's going to be harder to have like up 200, 300 companies have that kind of market share in this business.

So again, the focus is more on sort of businesses, as well as like how would Fortune 500 companies integrate such technology into their business day to day.

And there's a lot of fine tuning infrastructure across customer service, a lot around cybersecurity, where bots and bugs can be sort of self-curated and created automatically, like running different kind of softwares within the internal CRMs of the business, that's also been very, very interesting on the B2B side for us as well.

Awesome, so just switching gears a little bit, and we can kind of come back to sort of this sort of topic that I touched upon across sort of end use case for fortune 500 companies.

But just switching gears, you know, there's a lot of sort of companies and infrastructure that's being built out right now that is open source.

And there's been a lot of conversation around how it is very important to build in public,

It's a very democratic way of building companies.

But at the same time, from a business perspective, how do you think they'll be able to monetize it?

You know, There's this entire debate around build versus buy as well.

So what has your sort of key takeaways been around how open source softwares can have, let's say, sales-led or enterprise-led growth?

So yeah, the fund that I was at before over in the West Coast, we were investors in a couple of the darlings of open source software, MongoDB, Confluent, and there seems to be a playbook here, which is you build out an infrastructural set of tools, you launch them out into the community on GitHub, people start building, people start tinkering, they're adopted from the bottom up by developers and then embedded into the enterprise because there's enough kind of groundswell around them.

And then you sort of ask yourself, OK, we have tons of users.

And the way you do so is potentially by offering a managed cloud version of what your product is, so that enterprises can feel comfortable with how it's hosted and deployed.

You offer premium support so that slightly less technical personas can interface with the database or the infrastructure itself.

What we now have are a bunch of open source tools for building with LLMs, you know, whether it's Lang chain or it's PG vector or, you know, kind of pick your poison quadrant in the vector databases world.

And so each of these founders will kind of look at the previous generation of open source business products that became massive enterprise companies and broadly work on similar principles when it comes to, you know, how do we convert this into an enterprise SaaS product that we can monetize through subscription.

It's hard to call a lot of this at the moment because the space is so early.

Even if we've got a lot of developers building, it's not yet clear what the specific requirements are by product or premium support for specific hosting of a database on infrastructure that's being provisioned by somebody else.

So we're in this very early stage of the market where we have historical precedents for how you commercialize open source software.

We can sort of broadly apply those to the open source projects that are getting tons of adoption today.

But the reality is you don't really know until you see traction in the enterprise.

And so as an early stage investor, you've got to take a bit of a leap of faith.

If you believe that the tool is good enough that it's going to become an industry standard, you sort of believe that you get the users on board, and over time, you find various ways to commercialize.

Just to kind of add to that, I call it like taking a shot in the dark with a little bit of light coming right before it hits the dark board.

That's basically how right now sort of early stage open source software AI investing looks like.

Yeah, I agree with that, but I also think that's the case sort of across AI investing in general.

If we had this panel 3 months ago, all of our answers across the board would probably sound very different.

And that's, and part of that is, you know, again, when we think about investing at the earliest stages, we think about founder and we think about what they're trying to build.

And I think in this space, more than ever, I think founder is the most important thing and we look at something that we call clock speed, right?

How quickly can you take in data, iterate, change?

Even the really successful businesses today, they've shifted, changed multiple times as the paradigm has shifted.

So I would say similar again to the horizontal versus vertical question that we had before, I think similar to previous paradigms, there's gonna be a lot of great successful horizontal businesses.

And maybe those are customer success or support, maybe that's a couple others.

And there's gonna be a lot of great vertical applications would be that in law, be that in we're looking at things in VFX and sort of Hollywood, there's a lot of great companies that are being built across the board.

And so I apply that again to here where we're talking about open source versus closed source.

I think 1 thing that's interesting is that in this new paradigm, people have gotten to their first thousand GitHub stars in a matter of 5 minutes, maybe even 1 minute.

People have gotten to a certain amount of adoption much faster than we've seen elsewhere.

And so that I think we often think of Where we invest in our sweet spot is really the series A as early signs of product market fit.

And I think there's a lot of signs that people are interested in the problems that we're looking to solve.

And there's still the question of there's it's again similar to previous paradigms people everything will fall into open source that can be commercialized, open source that will always just remain open source and won't be commercializable, and then closed source.

And so we're continuing to look at everything that gets traction and again that happens.

That's a lot of things that we're looking at every single week and making evaluations on a case-by-case basis

Yeah, definitely a great call all the points are for sure so I guess like quickly kind if people aren't familiar in the Around like what closed sources and what open sources so open source open repose anyone can download the technology, use the software, and modify it.

So you can say, hey, I found a bug, push the code, and then the bug that you fixed can now be solved within the open source software.

And so obviously, as everyone noted, there's huge benefits to transparency and collaboration and also the 1 to many.

With closed source or normal companies, you have a salesperson saying, hey, I'm trying to contact you to look at my tip to sell my software with open source people can just find the software and And And it's very like bottoms up a lot of its organic growth.

Some of the most, of course, like some of the most valuable companies right now are open source.

And I guess it, There's so much magic to open source.

I come from the sports betting world, and that could not be the more opposite thing.

Like, everything we worked on, it was like you couldn't talk about with anyone.

And I'm still, like every single time I use any sort of open source technology, I'm blown away that people are writing code and there's like the ethos of software is to open source it.

And I've also spoken to a lot of, I mean, billion dollar funds that say, if open source isn't the future, I mean, we're, our funds probably going to 0.

See all OxGen AI transcripts on Youtube

Investor perspectives on generative AI: Oxford Generative AI Summit 2023