00:29:51 · Luca Ferrari, Anish Acharya
[00:29:51] S2: I... Sounds so great. [00:29:52] S1: Yeah. Excellent. Hey. So you mentioned proprietary data, which I think is actually under discussed these days with everything happening in models. You know, there's a quip that many companies are sort of trading for less than the value of their tokens, which is the value that they would have if they were simply acquired for training data by models. Do you think that's true? Do you think proprietary data has changed?
00:30:13 · Luca Ferrari, Anish Acharya
[00:30:13] S2: So I I I haven't thought enough about this and I don't know enough about it, but I will say that claim strikes me as potentially a little bit misleading. Meaning, the value of data within new pools declines pretty rapidly. It's like if today I write a new blog post on the internet, nobody's gonna give me money for that. It's private and I'll go to open... I'm about to, you know, I may publish it, I may not. Do you want to give me it on top? You know, the first blog post they ever got access to or could ever access, they would probably have paid a thousand dollars for it, but now zero because incrementally it doesn't add anything to the corpus of knowledge that you can train AI models. So yes, maybe there's now a need for plenty of data on how people do project management or whatnot. The first companies selling that maybe make a bunch of money, but I think over time that trends to zero as the, you know, the quality of incremental training you know, returns a decelerate for incremental training. But I'm I'm in no position to determine whether, you know, companies could sell data for more of their enterprise value. I'm not I'm not certain about that. Interesting. Let... [00:31:24] S1: Let's double click on the kind of operating system you use at... Within Bending Spoons to help improve and optimize all of the companies. Can you talk about some of the key components of it? And which of the components do you think would have been most useful to you as a single product founder versus are very, you know, specific to what you do with the portfolio?
00:31:24 · Anish Acharya, Luca Ferrari
[00:31:24] S1: Let's double click on the kind of operating system you use at... Within Bending Spoons to help improve and optimize all of the companies. Can you talk about some of the key components of it? And which of the components do you think would have been most useful to you as a single product founder versus are very, you know, specific to what you do with the portfolio? [00:31:42] S2: So there's pretty much everything that you need to run a digital business. We built in house with the exception of highly, I think universal tools like Gmail or Slack and these sort of things or Google Drive, we're happy to use those that others build. We don't have highly specific or personalized needs. But everything else from data processing, the AB testing platform, all the marketing automations, user lifetime prediction, credentials management, recruiting tools, it's... The list goes on and on depending on how you count, it's over 50 different in house tools. They... We build them ourselves and and the reason why this has been valuable for us is, well, the obvious way is we we save on vendor costs quite a lot actually, but but it's not something I'd say is a game changer. [00:32:36] S1: Mhmm.
00:32:37 · Luca Ferrari
[00:32:37] S2: The two game changers are one, we make them super fit for purpose and very specific to our needs and these tools are natively integrated with one another, so the sum of the parts is much greater than what you could possibly accomplish by buying these different tools from different vendors, because each is better for what we exactly need to do and they also work in a more harmonious way together. Also because we can... There are this big monolithic foundation, we can swap out the foundation at an acquired business quite rapidly. It's actually much faster to change the entire thing in one or more or less the entire thing in one shot than removing one piece and then you place the alternative and you need to plug it to everything else, then remove another piece, plug it to everything else. So actually swapping the entire thing out tends to be faster. Now of course this is not particularly relevant to a single product company because they they don't need to buy new companies so that's that's not what they need. But they could probably benefit from a bunch of these tools. A lot of the value is unlocked by having the right culture and the right talent because like every tool ultimately the user makes a ton of difference and so depending on what company we're talking about they could be able to take advantage of it or not but some are more let's say autonomous and they don't require a lot of effort and quite powerful for example we built something called ALT ALT, like alternative spooner, that's a... It's basically an agent that lives in Slack and you talk to an ass, automatically has access to all the same tools and assets as you do. It's the same GitHub repos, same folders drive and whatnot, and you can task it to do things for you. Anything you could do in theory could do, and increasingly it can do and does. So you can tell it, for example, to run a data analysis for you and it will go fetch the data. If it doesn't have access, it will escalate a request to our credentials management team and potentially our IT security team if relevant.
00:34:41 · Luca Ferrari, Anish Acharya
Get get the data process it show your graph and commentary you don't have to ask anybody else it you get it in five minutes say many man hours and certainly a lot of delay in the exchange of information I experienced an interesting situation where a general manager of Evernote for us in a channel where we post feedback on the product she wrote to her agent I've seen this bug can you please go to Moros which is our customer support tool to check whether other customers have reported it then go to the code base find a root cause code a fix and ping Marco which is our engineering lead on that product to review that fix and fix it and and it got done and that would I mean it it was a 100 times better than what was possible before. So that's maybe a tool that I think could be close to a game changer for a lot of companies. [00:35:32] S1: That sounds really powerful. As you think about the way models get deployed within the company, do you think there are any job functions where the model will be doing a 100% of the work in the near future, or do you imagine it being a kind of lever for individual productivity as you just described?
00:35:46 · Luca Ferrari
[00:35:46] S2: 100%. So a customer support, for example, we have evolved our... All of our customer support managers from being more of like agents answering, you know, responding to tickets. Mhmm. Now we are for that we pretty much a 100% use either AI or for the residual portion of the tickets external vendors and our customer support managers are more and more similar to say a product manager trying to architect the ideal response process, response rules and user experience and this is just the beginning, know, we're investing heavily in this to make it much better, we'll include at some point support via phone calls, WhatsApp, So that's an obvious strength for LLMs because it's mostly about digesting a body of knowledge and exchanging communications with people and LLMs are have been trained on data that's quite similar to this and and already today are pretty good at it. Another area where we were we have seen the percentage of tasks taken by AI increased massively is data science, data analysis. Now our data scientists are doing much more here too, architectural work, building... Writing software and building tooling and processes for, say, the budgeting process or a recurring dashboard, but the actual fetching data, cleaning data, plotting data type of work is done. I don't have a precise figure, but I'm pretty confident at least 90% by AI. And then coding of course, we were at 0% written by AI two years ago, we're now... Last time I checked at 94%. Wow. Some companies, you know, depending on how aggressive you wanna be, you could be... If you're a good company, it could be at seventy and ninety nine, but no good company should be, I don't know, below 70 right now in my view. Yeah. There's no point.