scikit-learn & data science you own
52:04 Share

Welcome to practical AI, the podcast that makes artificial intelligence practical, productive and access like this show you love changes interviews, days in on fridays and awesome talk show or your weekend enjoyment. Find us by a searching for the change log whereever you get your broadcasts, thanks to our partners at fly out I O launch your AI apps in five minutes or less. Learn how at why I O.

Okay, friends, i'm here, the new friend of hours over at time scale atha. So so atha r help me understand what exactly is time scale?

So time scales a post gress company. We build tools in the cloud and in the open sourcing system that allow developers to do more with progress. So using IT for things like time series and and more recently, air applications like crag in such an agents.

Okay, if our listeners were trying to get start with post grass time scale A I application development, would you tell them what's a good robot .

if you are developed about the you're either getting top late building and air application or you're interested and you're seeing all the innovation going on in the space and want to get involved yourself? And the good news is that any developed today can become an A I engineer using tools that they already know and love.

And so the work that we've been doing, a timescale with the P G, A, I project, is allowing developers to build A I applications with the tools and with the database that they already know. And that being progress, what this means is that you can actually level up your career. You can build new interesting projects, you can add more skills without learning a whole new set of technologies.

And the best part is it's all open source. A I N P G vector scale, our open source, you can go and spend IT up on your local machine via docker, follow one of the tutorials of the time stop log cutting edge applications like crag, and without having to learn ten different new technologies and just using post gress in the SQL cray language that you will probably already know. And I familiar ah so yeah, that's IT gets started today. H it's A P, G, A, I project. And just go to any of the time scale to reh either the P, G, I one or the P, G V scale one, and follow one of the tutorials to get started with becoming an a engineer just using progress.

Okay, just use post grass and just use posters to get started with A I development build rag, search A I agents in its all open source, go to time skill dot com flash ai play with P G A I play with P G vector scale all locally on the desktop is open source once again, time kill 点 com flash ai。

Welcome to another episode of practical A I. This is Daniel White neck. I am the CEO at prediction guard and i'm joined as always by my cohoes Christenson who is a principal, A I research engineer at lucky Martin. Are you going, Chris.

doing very well today?

Daniel has a going it's going great. I was saying that i'm really pumped to be talking about something that's near and dear to my heart um over many, many years because today we have with us yan, who's the CEO at probable and gome, who's an open source engineer at probable.

Welcome thanks for having us.

Well, Young and game are working on data science that you own, including projects like psychic learn, which is of course very near and dear to me along with other data scientists uh all around the world. So yon if if you could, since you're coming from the CEO prospect, us help us understand a little bit maybe for those that have heard of psychic learn or some of the other projects that you're involved with, but they haven't heard of probable. If you could give us a sense of what is probable as you mention, kind of in the lead up to this conversation, it's a slightly different kind of company that came about in different sorts of ways um then other types of start up. So yeah, if you could give a little bit of context, that would be great.

Well, very glad to be on show with you today. And a probable is a company that is typically known as a spin off from a research center in france called india. And in ria is the place where this technology, psychic on, has been developed over the last ten, fifteen years.

Not many people know that. And the project has been somewhat protected and sort of incubated within the research center. And after all that time, as you know, that I turn has been adopted or even probably anticipated in creating the field of data science because IT is applied math and essentially has created a sort of paradise, uh, for our data scientist approach.

Data science typically through two functions, fit and predict. And the front government has a national strategy for A I like many, many countries. And the government decided to double down on psychic turn and they came up with a budget.

They interested the research center with that budget but then they also asked for the project to be break given at some point and the teams said, okay, break hevin is fine but we don't do that um in the research center we don't break heven so when do we call entrepreneur uh to try to help us figure IT out and they called me so I have a track record as a software engineer and a an entrepreneur intact for the best twenty five twenty years but i'm not a data scientist so I did my due diligence and I sort of a dug deep to find out what this project a little about under the hood. Is there any good? Is the community any good? And of course, a psychic arn.

Is this a quite amazing uh, jewel of a technology that a every data scientists in the planets uses, that discovered that he was download one five billion times cumulatively, eighty million times a month, twenty two percent in U. S. Only three percent in france.

So this is a project that is used, the holiday world. And the probable is essentially the spin off that takes all of the team, including game, here from the research center and turn IT into an open source company that inherited the mission that was initially given to the research center. And the mission is to build a sweet of open source technologies, including psychic arn, but above and beyond psychic karn as well for data science.

So the scope is large, the mission is novel, and this is what we're building essentially. So probably is a one year old company that has already started doing many, many things. And uh you know geo is the uh representative here uh for psychic on this technology that is a used again by every data scientist in the planet.

Well, this is brought up a lot of interesting questions on on my end. And I I really love the part of your pitch and at least how you framed IT on your website and in your materials online about data science that you own. In the open source side of this, which I know from experience, there can be some interesting chAllenges around finding business models that really work with open source SE technologies.

And we've seen technologies where companies start with the posture towards open source and then gradually become more closed over time. So i'm wondering from the leadership perspective, that sounds even in the way that this company was formed, that there is a posture towards stewarding that you know psych learn in these types of projects. But from your perspective, what what is your posture towards stewarding these projects in the open source side? And how do you view the business element of this to make a sustainable in the longer term?

So that is the hard question, but IT is the one that is important here. Psychic n is a technology that is again applied math. It's not rocket science but it's a played math uh and it's quite intricate.

The thing is uh the scientific community uses IT uh day in, day out and everyone dependent that. So typically when I, when I discovered red, this cover the project in the mission that was interested to the research center, I realized that this project is bigger than me, number one. Number two, the mission is took, actually create more open source.

Other words, you know, in twenty twenty four is even more acute. Typically, big tech keeps on a massing so much our concentration and we could argue that they do not distribute as much as they should. So that's not a judgment.

But IT is a fact and psychiatry is precessing a contrary. IT actually enables so many companies to to do data sense. So with that in mind, uh, before creating the company, we decided to craft sort of architecture for the company that would respect that. And so you know, before we created the company, before give me joined as a founder, uh, before even incorporating the company, we had a template that actually the the governance, the shareholding structure and also leveraging a new law in france that allows us to do the sort of b corp. So a company with a mission where the mission is clearly stated in the bylaws and that mission is to create or enforce for the science.

So in where we've created a sort of constrained environment that is unlike many companies because it's by design, this company by design has created god rails so that the governance cannot take this company too far on the right that, say, property or technology or even changing the license, that's not in the cards. And we've created a sort of mechanism where, you know, if we do not appalled to the mission, then we can actually lose some of the assets such as the brand. We are the official brand Operator, but the brand belongs to the research institute, right? still. So there are many mechanisms, trigger mechanisms that forces, including shareholders, that we, we, we would bring in to actually bind with a mission long term.

Ta, you ve read so many questions for me that I want to, that I want ask. I actually want to take just a moment and kind of go back because IT occurred to me, as we're talking about this, for some folks listening who may have never even use psych learn, they might have heard the name and stuff and you talk about of being applied math. Could you guys expand on that a little bit for somebody who hasn't had a chance to ever actually utilize IT themselves in terms of what is doing and kind of catch them up to us and the conversation a little bit? And then I A pear, a few more questions because you've got me you've got me really interested that you hit so many .

topics on that last yeah so maybe I can I can give like a big of background. So pilon basically attack line is a machine learning in baLance. So it's go back to the they say statistical route.

So the simple answer is we try to make prety predictive modeling. So try to use mathematics to from like data the board in the future to like give answer to a specific question to a specific polling. Uh the big difference with general ep learning is that all that you have that you have there are simple steps.

So there are fundamentals and deep learning bill on those but and are just like much more, let's take costly to train， costly to in the infant states. Uh oh not in the same scope as well and that is like a defect of choice when you want to have like doral data. So excelled spreading its uh data for train this way so that the the fact of way of train like to be able to eat those h created and give back some levels or some like regression, I say. And whatever is like image NLP or like this is like it's a deep learning and and transformers is more in that area. So we are more like back to what was machine learning like few days ago, but that many, many, many applications based.

Well, considering how incredibly popular in the foundational in the data science world, could you could you kind of give me a little bit of a landscape view? And i'm not sure which of you would be the right one day answers so you can pick between yourselves. But a little bit about um kind of how that fits in to the data science, the landscape with A I coming in just that with people listening, they can come to go.

H I see how IT fits into the the many organizations and tools that are out there. How do you think about that? Uh, for that and they know get back a little bit more to the organizational stuff that are you talking about few minutes ago.

So maybe I can answer like partly, which he is by giving uh use cases and and to to see that uh for in with a partner as well that we worked uh over the years uh to give like where do you find machine learning and prince, a machine learning in can be found in s here where you want to are you know you want to know if the drugs works or not then if you want to find the this is as well in some type of data IT could be as well like for the detections in banks in in insurance uh predictive intent uh and those stack of like a all applications that we have since like many years.

So let's says the use case, these are very large. And what bring like pilon on that is that this is not we I mean, he is not for one of the use case. I mean he would have thoughts from the being to be general enough you can apply to any of those use cases and to come back to, let's say, classification and regression programs that say or unsupervised as well but um like that you can appreciate to anywhere. H in that. So maybe and something more you add perhaps .

also the micro level is to say you know to say you can and does a lot fans, including deep learning. But in to be to be Frank, when you want to do deep learning, typically you'd go to pitch or tens of though. But for everything else, psychiatry, in other words, in the great A I family about horizons, there is machine learning.

And within machine learning, you have declining. Within deepening, you have other categories are going them, such as know transformer based models that lead to elim's. So it's basically, you know russian doll of sorts. And psychology is the biggest provider of algorithms in the machine learning space.

In fact, if you look at the downloads, typically psychic arn is downloaded as many times as pitch and tensor for combined, which is crazy because now everyone is talking about, you know, L, L. Lambs, of course, but also deepening because depending is currently in a spring state, not quite a winter yet. So of course, deep earning and genre is a wonderful breakthrough.

That being said, you know, I like to simplify sometimes the eighty twenty prieto distribution. So I had the intuition that eighty percent of the use cases out there use psychic ern when he comes to machine learning. And people actually tell me, know, yeah, you're wrong.

It's more like nineteen, twenty five percent, right? Because in terms of you technology that is robust, that is tried and true, that is used to actually you know turn a profit or return on investment banks and insurance companies, right? H give me with mentioning fraud detection, fraud detection type uses and that actually says money that actually, you know banks would be losing money without that.

So IT is actually quite essential. But again, it's a great much right. So it's like to earn is only a fast data to this category of problems.

Was our friends, i'm here, a friend of mine, a good friend of mine, Michael Green, nich CEO and founder of work. O, S. Work of us is the all of one enterprise.

S, S, O. And a whole up. More solution for everyone from a brain is start up to a enterprise and all the AI apps in between. So Michael, when is too early or too late be going to think about being enterprise ready.

It's not just a single point in time where people make this transition. IT occurs at many steps of the business surprise sign on. The example of you usually don't need that until you have users. You're not onna need that when you're getting started and we call in enterprise future.

But I think what you'll find is there is companies when you sell like a fifty person company, they might want this actually if especially if they care about security, they might want that capability in IT. It's more of like S M B features even if if they're tech forward at work west, we provide a ton of other stuff that we give away for free for people earlier in their life cycle. You just don't charge you for IT.

So that off kit stuff, I mentioned that identity service, we give that away for free up to a million users, one million users. And this competes with off zero and other platforms that have much, much lower free plants. I'm talking like ten thousand, fifty thousand, like we give you a million free because we really want to give developers the best tools and capabilities to build the products faster, to go to market much, much faster.

And where we charge people money for the service is on these interpreters things. If you end up being successful and grow and scale up market, that's where we monetize. And that's also we near making money as a business.

So we really like to alive our incentives across that. So we have people using off kit that are brand new apps just getting started. Companies and White combination side projects, hackathon things know things that are not necessarily commercial focus, but could be someday they're of future proving their tax ac by using work of us.

On the other side, we have companies much, much later that are really big who typically don't like us talking about them. There are logos know because there are big, big customers. But they say, hey, we try to build this stuff where we have some existing technology.

But sort of unhappy with the developer that built IT maybe has left. I was talking last week with the company that does over a billion in revenue each year and their skim connection, the user provisioning was written last summer by intern who's no longer obvious that the company and the thing doesn't really work. And so we're looking for a solution for that.

So really wide spectrum will serve companies that are in you know, their office is is in a coffee shop. They're living in them mothe way through. They have their own building in downtown san or a new york or something.

And the same platform, the same technology, the same tools on both sides. The volume is obviously different. And sometimes the way we support them from a kind of customer support perspectives is a little bit different.

There needs are different, but same technologies platform, just like a eight of U. S, right? You can use eight of U. S, and pay them ten dollars a month. You can also pay them ten million dollars a month, same .

product or more, no matter where you're at on your enterprise ready journey work. OS has a solution for you, the trust about complexity. Copy A, I, E.

Indeed, so many more, you learn more and check them out at work. O S dot com as W O R K O S dot com. Again, work with that com.

So yon your kind of arty going there and I I love the direction that you're going with this, but I I think maybe I could take up A A softball for you here. I'm personally passionate about the the answer to this question and you probably have A A Better view on IT. But there is there might be people out there maybe listening to this podcast.

We are thinking, well, now that we have G A, I, we have large language models. I could put in a prompt to one of these models to do fraud detection or to find entities in text or to make some prediction of a classification. And you know, sometimes that works. And so maybe there's people thinking, well, there's these general purpose large models out there.

How does that change the way that something like psych learn plays in in industry? And I personally would argue and and think that this actually makes psych learn more valuable, if anything, rather than than less valuable in the terms of the ways that IT can be combined even as a tool that's orchestrated with geni models. But i'm i'm curious you're perspective on this from the business side and maybe geum has some ideas on the technical side.

yes. So um psychic and typically is this one technology that is patron manual, in other words, the logs to everybody. Um in fact, there is another stat when you look at the you know the figures that are public, by the way, the number of dependencies.

So psychiatry is actually used by nearly nine hundred thousand projects and given up. So there is nearly a million projects are dependent psychic. Um and there's A A new law that I discovered recently, someone mention that or in this effect, which means that something that's been used long will remain important for long enough.

So not saying that second learn will go the way of cobo, the psychic learn is here to stay and we are with the community, the gardens of that. So we're going to make sure that psychic arn remains there forever for companies that actually needed in a stale version. And of course, geum and the team are building up new features as we go, right? So there is a uh, dedicated effort, and I should say that we have carved out a nearly ten people in the team are doing only that, contributing to psychiatry and the other associated libraries.

Now your question, Daniel, is is you know whether psychic on will be absolute and say a number of years is because general proposed technology has made IT you know be relevant in some ways um number one, psychic extremely frugal. IT actually works on C P U. And IT is well controlled, well understood.

Uh it's actually quite predictable in some ways where deep turning is usually known as a black box where it's really, really hard to introspect. And so second and does produce for certain category of problems, things that I actually working quite well, more so than language models for sure today and and more so than any sort of deepening based technology that we understand today. Now IT is possible that with additional data, additional training and techniques and even evolutions and and the transformer base model, we could improve and probably render absolute second.

But to us and you and I, we talk about that. And with the team, we also experiment without limbs. And we are also trying to figure out how we can use these new technologies to actually help of first personal, and that is the data scientist.

So we are a technology provider to help data scientists and increasingly saw the data scientist in enterprises because we will be creating valuable services and solutions so that we cannot generate revenue to sustain our mission. So the goal for us is to actually project ourselves while contributing to open source, but also create a sort of business value proposition, not the similar to red hat, because that is closest type of company that we identify with. And then the spirit .

to that point that are making right there, i'd like to get back to something that you said earlier that, that feels like you're kind of tying back to at anyway there. And that's that you talked about, you know the mission to create more open source uh, and the mission that you're trying to create this environment that you're describing by design, you said, which is you and that with psychic learn here to stay for the long hall, it's going to be something that you is not going away soon. It's solving such a high perceptive problems. Could you describe a little bit about kind of what you're thinking around that in terms of further developing this particular for and the ecosystem around IT, uh, so that we have to benefit you know for for many years to come? How how are you .

approaching that? So the the company is built with multiple business units, if you wish. That is a big word for, but we have multiple uh revenue lines and multiple actin even within the open source team they did so gone perhaps can elaborate on some of the other libraries that we support that compliment psychic cil.

Uh, you know that's one way to answer the question. But also, we are building a new product, which I called reversible sense. So we are building a product that will provide additional value to data scientists.

And the goal is to create a sort of I don't want to use the term cool pilot because that IT is too close to olympic, but IT is the spirit. We are building a companion to augment the work of data scientists all the way, two teams. So that is an additional product on top of psychic on because phidon just works.

And so we don't want to change that. And contrary to a company that would build a sass solution with a propriety approach, we want to say, okay, whatever you guys use is fine. We need to find a way to add new value and some of IT will be open source, certain modular. But for those companies that have more money than time, that in more service than beyond their own, we'll have a solution for you and we'll make your life easier. And you know, data scientists are a new breed, is a new type of job, is not been around for very long.

And in a way, when I talk to people, so you know I been in code forever and we you know this where the the developers is um when they get hired, they are turn key in some some ways, right that they are get environment and they know how to peer code and that's all pretty standard. But when you talk about a da scientist, sexy, quite artisan, it's an art and a science at the same time. And your your manipulating two objects, actual code, but they are scientists are not codes.

And you manipulating actual data, a start code, it's patterns. And so data scientists are have a difficult task, which is to combine these two things and create value for the enterprise. And then they talk to business units.

And then I, what do I do with this model? I do have print production, right? So there is a huge common room to solve, and that's what we're going to do. Additionally, two, building up in source that are modules that people can use. Maybe you you can elaborate on some of the other libraries that are key uh to actually help.

yes. So we have so we have the open team and so we work for like many years on second already, but we see the importance and as a community, which is importance of putting mother into productions and as well getting closer to the data sources. So we are just like uh, working on libraries that should like make those come together.

So for us, we have a library that uh is on the area upside that his school scops that we we want to be to make uh like the persisting more sewing some way. But we look at well on how to bring databases like uh secure words into like closer to the machinery models. So like how can you trust some data we states with different tables and how you can be in new item words we are caring so much about, like seal for some, and how you can bring this into psychological in psychology as well.

We want to improve, like many ways, visualization, evaluation, inspection of models, which is on the top of, just like training algorithm. Because this is so, we want to augment all those aspects in like beyond those and either is inside, learn either. This is like library connected to psychotic.

So the one before it's called scrub, by the way. So like scribing data, so it's scrub. And scribes are two debris that the we look at.

So you know, we're of talking about the libraries now.

You know you have this uh, robust open source contributor community built up around psychic learned in the various projects within a um how does probable work with those? How have you guys set up that relationship? What does the government look like on that with with because you have both your courtine that you alluded to earlier ah that's working improbable you know at probable on this but you also have that larger open source community. How how does that all work? But can you to tell us how that of all I imagine it's quite mature by now.

and that's the point. The the maturity means that by design, we decided to not affect the license of second learn. We're not branching IT out.

We're not care for IT. And so the governance of cyclical and being so saying already means you don't touch IT if you don't broke, don't fix IT. So the the governance is unchanged.

So the the center of gravity was at india, the research center, but also involving people all over the world. I don't know. Give me how many contributors? Maybe two hundred.

Oh, even more. I think like you know you you have more than that yeah maybe like you maybe free from that. Uh and the core team is like to say half of IT may be around friends around paris, around provably, but then is like another half of ten east person around the world that contributes very like almost every day that say uh by community with the community and as as a Young tal nik, we didn't want to change that.

Nothing change in that regard. So anything that we actually did more, we we we did more uh to bring transparency. So to explain to people. So now that we can we feel that because when private entity we need to communicate, what are we doing and on what are our our road and which the committee items are we going to work on just to like bring more trusts said that, I mean, we don't like go like in the dark and that nobody knows now what we are doing. So we try to ready to pat and to every six months to mention which of the item that are defined by the community. They are not defined by by power, but from the items, which one we have the capacity to work with, the human resources that we have, let's say, at hand. So we really want to to to show that and .

and by design the open source team that is full time. On pattern in other open source libraries means it's a cost center to the company. So that cost center is by design, and we know that the cost we have to cover.

So we we will cover IT through different types of activities, for instance. And this was something that was done in the past where brands uh wear sponsors. So either they hire someone that becomes a core developer and they're naturally sponsoring someone to build up this technology or they were giving money as a donation to the research center. But now that the team is with us, we are translating this to into a contractual sponsorship homework. And so you know brands who wanna contribute to psychic on and help us compensate for salaries, we will get something in return.

Uh, exposure and a and and you say actually put more money into IT, then we'll have a conversation around the road map, find a way to make IT converge in a win win kind of way because game, for instance, can say, you know this brand wants us to do something, but IT makes no sense for the community, then we won't to take our money for the sponsorship type of business, right? However, if companies want to pay us to do a certain type of paid for software, will will look at IT, but that's a different branch of the company. So we've really clearly separated.

And by design, we know there there's a cost to IT. And that cost is actually if we are doing well, it's compensated by the fact that we have done good by the brand. In other words, hopefully, the the community will actually resonate with what we are doing.

And so they'll pay us back by actually appreciating what we are doing, which will Carry the message further. So we think that there is a cell ful filling prophecy if we actually keep adding value to the whole scheme as opposed to removing value. And I will not name certain projects that i've chosen a different way.

But on the other hand, going back to the governance of the company, when the company flips and becomes VC funded or only VC funded, vcs require a sort of return on investment that is too radical. And so that sort of forces a change of posture with visari the community and and the licensing ing scheme. In our case, we've actually created a structure that is baLanced in terms of shareholding groups.

And so we will ultimately have that's the goal of the structure. The architecture is to have as much money from public support than from private support. So it's again sort of baLanced.

When we start a podcasting back in two thousand and nine and online stores, just the further thing from our minds now we have much that change. And you can go there right now some t shirts, and it's all power by shop of what do we do before shop of fire. I'll tell you, we did nothing.

We couldn't sell. There were other ways. Of course, they were very hard, very difficult shop. If I let us build up in entire front end, obviously branded like change logue is it's amazing. Merge that change com.

And our favorite feature is we use their API to generate a new coupon code a as coupon code for every gas becomes our forecast. And they get a free t shirt from our merge door. And that's so cool. They choose the shirt they want. They use the coupon code IT arrives, we have charged to them and life is amazing. But also you can go there right now to merge dot change, love that com and buy some press yourself and that's A S so you and get the same check out we use with dollar month period at sharp fy 点 com flash practical A I all lower case go to sharp fy dc flash practical AI to upgrade selling flash go ai。

So as we come back out a break here, i'd like I want to turn to kind of a fun question for you. And I like i'd like each of you to take a swing at IT because it's not specifi C2Being the CEO or bei ng doi ng the tec hnology its elf. If each of you could describe kind of a cool use case something fun or interesting or that's really captured your imagination with psychic learn um and kind of shared that with the listeners in terms of something that that just kind of really took you as your thing, i'd love to hear i'm expecting IT to be a bit different coming from each of you in your different roles. But i'd love to hear kind of how how you see that and what's the thing that stick .

out your mind and you start because I have to think about IT. No.

so it's a very technical one that say, but so during my piece I was doing classification. So uh which is something that I was trying to find people that has a specific type of consent or process concert of people that didn't have IT.

And uh, inside that space you had one thirty specific and which discord imbaLance data and this what introduced me as a key to cycle because I had that pray I was using segment follow the specific issues and how to take up down those those type of issues. And uh, what is really funny is that so it's all I got interested to schone and speak, for instance, with the developers. And I develop one library which could, in baLance, learn that is a merging as well.

We psych n like is competitive in some ways. And for many years I maintain that package, even when I was my teen years, with a psychologie years after years, we did everything by is a book basically in the library. We implementing the arguments that we insecure res, and everything was fine until that as part of india.

A now probable we we have as well time to educate our set. And two, try to as well then bring fruit documentations of psych, learn to explain some conceptive people. And by doing this, we find out that most of the research there didn't look at the prime property.

And by communicating, we've overall, dave, we just like phone out that a huge part of the thing was just wrong and that you should look at this in another way. And then it's pretty funny because we've this we found like some useless self that was an imbaLance. But then now we have like Better contents. We went to conferences to explain these problems and people start to tell us, oh yes, I agree, that's right. And and he spend that you come and say that whatever you are doing, like five years ago, ten years ago, is actually like up so little, not good. Or I mean that that we we then expect from there and is something that I find very like fun when you do open source, because you can like you are here to call you to something and just to bring like the best of what you do to everyone and everybody will be like thankful for that even I mean, and you are not defending your own like h let's say entity paper or like that. All what is true and and for me that's like one experience that comes from my P D from like no eight or nine years ago to up where I am now and then like I I see like an evolution where I was very, very good people and you could like correct area that you doing the best and I agree that will beneath its everyone afterward because that standing in size documentation of secular or even in size of library and then like everybody will just like you, it's as a million of users will be affected and say, oh, actually that's good and this something that I would have staying up, for instance, free, that wouldn't have have happened because you wouldn't have like time or be quickly enough because you won't been like in in in the books. They the pressure books and and good ideas but that's the one.

And electives is good. I might be the C E. O, but I do have the imposter syndrome because psychiatrist is so impressive. It's day in the out, time in a team and and gomis, very humble and you know, very discrete, but the amount of knowledge and the amount of techy that is trapped inside the libraries mind growing and you haven't met to the other members of the team, it's pretty much very, very hard to compete in terms of the amount of CPU cycle that going there. So psych one is the gift that keep them giving in some ways.

And the team is just out of this world and nice, and it's just a pleasure to work with our team, uh, all the time. Now the more I discovered second turn and the more I find IT amazing a because of what the brands meets to people. And so last weekend and today, actually, we just released, if you allow us to actually put the the link in the notes.

of course, absolutely.

we released the very first official psychic on certification program. And what's amazing is that we so this is the first time so we're doing is that by step and system works, people can register that can pass or fail the test, but without advertising. We had, like within a couple of days, six hundred regiments strains all over the world.

But last time, india, I should. Because people in india, they do work also remotely for other clients across the world. And so they do need A A stamp of approval to showcase their ability to provide a service. So very interesting that this brand almost instantly and promote the sort of service that is valuing. So that's the one thing.

But then on the more technical level, um I fall enough with one new feature that came out with one point five uh of psychiatry by another cofounder and uh and corded approach me and that is the call back feature. why? Because second learn in fact is a platform.

IT is a platform and the cold back feature allows us to to provide extensions, if you wish, where people can hook into the inner workings of state term as they are building new models. And in fact, I find that to be essential because we are entering an age of the liability with regards to AI. Companies need to be able to introspect.

They need to actually find out why the model is producing such such results. And so introspection is is critical. And as I said earlier, depending is sort of a black box type of approach, which I love, by the way, again, in one thousand hundred and two building decorating models in the middle of the winter of AI at the time, a psychic is actually quite introduction cc, quite transparent for goal, as I said.

And so callback x are yet another feature that provides actual introspections into how we build models. Because talk about insurance companies, uh, flood detections, you've get human beings at the end of the of the spectrum being panelled by algorithms. And so that is critical.

And I think we we all feel a very important uh need, uh, with these features. So again, thank you. Turn the gift that give and giving and am impressed every day with a bit of an input or syndrome because that team is just so, so powerful with this tool.

And you know, speaking of of this, uh, team game, I get through a question at you. People out there been listening to this. They are kindly, okay, I want I want to dig into this.

So you're going to get some new developers, uh, they are going to come how should they engage? How should they fine and get started in the projects to develop. What's a good on boarding path for those developers?

Praise of best on boarding pass is like if you have a chance that inside your local community, there is some people that do what we call first time contribution to ensure or like coding prints, go speak to those people because I mean that we help you to get on board. But then like if if you are beyond computer and then you don't know where to start, is where we have like documentations that describe what do we call contribution.

Because contribution is not only coding, you could be speaking, they begging, commenting, organizing prints and and those type of things. So a and and we have like what we see to us contribution on how you can help, basically, and where you can be. So of course, the natural thing is to come and could, and then we explain you how to to start you that.

So this is like documentation uh china uh as a documentation repair and uh and afterwards, everything is online and public, so there is nothing pride. So is we have different channel of communications. The many one is a github and is being through the issue ranking or the put request, like depending on which size you are and called repair will be, I would say, twenty four hours over twenty four because we are around the world. So that's that's like if m city that somebody else in australia, in the us. That we can like just answer to you and and then were just like give you feedback and and then and is where you join you start you should not be shy and you should not be scared of making a mistake because we are not judge mental that we all started by by that set of like saying like I don't know what i'm doing and I need to ask people what should I do and that's a Normal step and afterwards you just grow with the community and then the committee bring you over mean. But the most difficult thing is there is the first step like engaging and saying, like, uh, so i'm import in syndrome as well but that people say I don't want I mean like this like those very killed people, they will never want to speak to me and that's not the the case so just come and just try your best and then people will just communicate with you for sure.

Great guidance there as we wind up. I'd like to get for each of all for both probable and for psychic learn kind of what you're think about for the future. And you and i'll let you define what time being the future is, you know, whether it's you know a few months or years out. But I really like to to wind up paint us a picture of when when the duties of the day have finished in. You're just relaxing and you're thinking about what's possible going for what do you think .

about go with the mission? The mission is bigger than me, began to bigger than us. Uh, and so that's why the governance create self sustain models. Of course. Uh, you know it's not trivial.

There's a other work to achieve the mission long term, but that mission ends up with an IPO in other as this company is not meant to be sold or wrapped up. The the goal is to do an IPO so that this company can Carry on with a mission and allowing people to invest. And be part of that story.

And that's why earlier Daniel ask a question about uh you know um investors and all that. So we do have seventy individual investors, including people were contributors or our contributors to say you turn who don't have the chance yet to be employees full time of the company. So the goal is to create this sort of dynamic vehicle.

And if we look at the north dark, there is no such company today that IT is the provider, open source machine learning technology. That company does not exist. And we aim to be that because we need that in an age where there is too much concentration, uh, within just a handful of players.

That's not okay. It's not okay for the global south is not OK for europe, which is lagging behind, there is not even OK for the U. S.

The us. May have big tech, but that's not okay as a single model. We need people to own their data science. That's why that is our time line.

I was good here. What are your thoughts?

Yeah so maybe more on. So i'm probably I am really thinking that we have emissions that say to help more data scientists, but I will speak more about like about psychology and under ecosystem. So for me, the mission is we should stay, focus on what's happening, how there and make sure that second, still the events.

So we have the final model. That's fine, but we need as well to understand where this is deployed and how this is used because we can make such progress that bring to make IT easier, to bring databases to cyclo, to bring cyclone models into production and to reduce friction and everything and and as well bring values on understanding. The more that I mean, we are speaking about A I X as well in europe now.

So i'm i'm sure like this, like plenty of IT say, are you are where we cannot read an impact and then there is well technical that move very fast to, for instance, before we knew pandas, now this is put out. So we need to move like in the fraction of sequences. How do we like deliver values to the user that just makes the switch? And still can you psych lean? Like can I can we do I accept those things.

And then so we have to make this audit of what's happening. So this is difficult to say where we will be in five years because in five years, we have all those things that can I say, we have a full chain of machine learning that pray will be here. So we should be aware, but we should be aware of whatever IT moves very fast around us.

Everything that was well said to gentlemen, you guys have done a fantastic job of of teaching the rest of us about this. And thank you very much for coming on the show today.

And pressure.

All right, that is our show for this week. If you haven't checked out our change log newsletter, had to change love 点 com slash news。 There you'll find twenty nine reasons, yes, twenty nine reasons why you should subscribe. I'll tell you reason another teen you might actually .

start looking forward to, sounds like twenty .

eight more reasons are waiting for you at change. Log up com slash news. Thanks again to our partners at flight, to I O, to break master cylinder for the beats and to you for listening. That is all for now, but will talk you again next time.