AMA | Episode 6 - The AI Agent Takeover: From Information Overload to Actionable Intelligence (Featuring Soso Value)

2025. 8. 13.

In this AMA, our Marketing Lead Joules Barragan is joined by Jessie Lo, Co-founder of SoSoValue, to tackle one of the biggest challenges facing today’s traders: cutting through the noise. With markets overflowing with headlines, sentiment shifts, and price alerts, the real edge isn’t just getting information, it’s knowing what matters and when. Together, we unpack how SoSoValue News Bot — powered by Sahara AI, filters, tags, and prioritizes market updates to match each investor’s unique profile. From building smarter, personalized feeds to blending proactive alerts with on-demand AI insights, this session explores the tech, the strategy, and the future of actionable market intelligence in a world drowning in data. 

Link to Full AMA: https://x.com/i/broadcasts/1zqKVdQVLVnJB

Transcript

Joules: Alright. And we are live! Hello, everyone. I'm Joules with Sahara AI. And welcome to our Agent Takeover AMA series, where we explore how AI agents are quietly reshaping the way we work, trade, consume information, you name it. Today's conversation is all about how we can move from information overload to actionable intelligence and what AI's role in all of this is. In the markets right now, there's no real shortage of headlines, sentiment updates, price alerts. But the real challenge isn't getting the information. It's figuring out what information actually matters to you. To explore this, I'm joined by Jessie Lo, co-founder at SoSoValue. Together, we've recently launched the SoSoValue news bot, powered by Sahara AI, which is designed to filter through all of that noise and surface updates that match your portfolio, your goals, your timing, you name it. So, Jessie, thanks so much for joining us today.

Jessie: Thank you so much for having me.

Joules: Yeah, of course. And before we dive into the details of this AMA, Jessie, could you just start by telling us a little bit about SoSoValue's mission, how the News Bot fits into your vision for smarter market intelligence? Anything else you want to bring up?

Jessie: Sure, I guess in one sentence. Our general mission is to democratize a professional grade, at the Wall Street level, of investment intelligence and insight, and to bring it to the public. Because crypto is about empowering everyone. It's about democratizing assets. And we think that when we first come into crypto, on a given day, we need about 20 different tabs open. I mean, you need to pay for professional-graded reports for hundreds and thousands of dollars per month. And we believe that's like the anti-crypto approach to crypto. So the first step is we want to make information accessible and understandable and then actionable by average, everyday investors. And you're totally on point with the challenge that we face today. It's not the lack of information, but it is way too noisy. It's very difficult for us to separate what is truth and what is general chaos and noise and what is just scams. And to filter through that information, you either have to get through a really hefty paywall, or you have to spend a lot of time on it. And in order for crypto to go from a $4 trillion market into 10x that or even 100x that in the foreseeable future, which we all hope for, the number one challenge we need to cut through is to cut through the noise. So that's exactly the whole goal and mission behind SoSoValue and we first of all build a terminal that helps us filter through the information gap. And then we also build products like the on-chain Spot, ETF Source of value indexes and then Sodex as well to help democratize that access and for everyone to better participate in crypto and allocate assets into crypto. But that's down the road. The first step is to make crypto understandable.

Joules: Yeah, I love that. I've been in the crypto space myself since about 2015 now. Jeez, it's been a long time. I'm aging myself now. And it's interesting to see now how we were back then. I'm pre-DeFi summer, so things were crazy. And now that AI is in the picture, things are even shaping up kind of wilder than they were. And I've worked in the MEV high-frequency trading space for a while where on-chain data is everything and you're always chasing those opportunities and everything is seen in these milliseconds. And that's a completely different ball game for most people. Most traders aren't in that game. Most of us, like myself, wouldn't consider myself a high-frequency trader. The real triggers for us are news, what's going on in the news, what are the sentiment shifts, what key announcements are happening. That's the kind of alpha that we're looking for to spark that deeper investigation for us. And the problem is, and what you highlighted is that signal, those really important signals, are being buried. In a world where there's a new token launch honestly like every second, it's in my opinion, becoming impossible to keep up. And every trader now, they have more charts, more headlines, more sentiment feeds that they in a million years could probably never process all by themselves. And with all that information, people still make really shitty trades. I'm wondering from your perspective, what's the real cost of this constant information barrage?

Jessie: That's a really interesting perspective and exactly. Precisely as you said, the market has also gone through a few different cycles, and in different cycles it's driven by different things. And I think what sets this bull market run aside from all the previous one, is less about… in my head, the crypto market went through a few phases. In the very beginning, it's like a total wild west, you have those extreme idealists who wrote the right paper that we aspire to today and you have the pure scammers. That's like the wild west phase. And in that phase, information is definitely a bit chaotic and you rely on a very small circle of alphas, quote unquote. And then the second phase is more like a, we call it the casino phase. It's more organized gambling and there's a certain rule in place and there is a certain way to get high-quality information. But usually that comes with either a cost or it takes a while to find everybody who to follow on Twitter, who can send a proof of time, who hasn't rugged in the past three cycles. So it's very costly in terms of time or money resources. As of right now, I think we're really stepping into an era where the crypto market and the traditional financial market is accelerating into merging into one. And crypto has really become what we believe would be the future of crypto, which is an essential part of a financial building block. In order for that vision to happen, a few things need to happen. 

On the technological level, millions of dollars and many sleepless nights for the dev team building more functional blockchain, et cetera. But on the informational landscape, we haven't really seen that much of a step change. So that's where we play into the field, especially when we believe that this bull market is driven by Wall Street money primarily. This starting with the whole BTC ETF, Ethereum, and now many more ETFs passing, started the bull market engine. And now we have a lot more of the treasury plays and a lot of tokenized stocks, etc. So the two worlds are merging at a much faster speed and we need to bring the Wall Street-graded information level to the public at a much faster speed. And that's not only in terms of what you mentioned, the accuracy and timeliness of information, but it's also prioritization for investors. And that comes into the third piece, which is relevant. Information that's delivered at a totally irrelevant time and place is noise. So that's where AI comes in and is doing it a lot better than human agents, I would say. 

Because prior to Chat GPT actually SoSoValue existed because of the whole emergence of Chat GPT in the very first place, because before that we have 10 people from top-rated investment banking and research firms coming in to do research every single day and churning out reports on both the macro financial markets and the crypto markets. And we have morning meetings and then evening meetings every day to summarize those reports. It's very valuable information. But it's also not very scalable. It's not until Chat GPT happened and we're like, "Oh, wait, we can actually do this with much less humans in time." It's not long before then that we realized that maybe we can bring SoSoValue to a much bigger platform other than a smaller group of investors. That's where you can also see on our website at social.com, all of the news feeds are currently all cleaned, sorted, tagged by AI and to curate it in a way that makes sense to investors. And there's hundreds of hours behind the scenes of training the model. 

And recently you mentioned that we brought to the market the AI News Bot in collaboration with Sahara AI powered by technology to even further filter through what's important to investors and also what kind of investors you are. Depending on your risk profile and the type of investment behavior, you probably need a totally different set of information. With AI, we can also make tailor-made news feeds for every investor. For example, for you as a degen, you probably need a lot more on-chain analysis and more in the DeFi degen world. But for everyday investors, for example me, I'm a big index spot index embracer. That's exactly why we launched our own on-chain index fund as well. For that, you probably need a lot more macroeconomics input. When are the Fed's rate cuts happening and what is the consensus over the Fed's rate cuts and how is the war going and how is this pricing into BTC price, etc. So based on that, we believe the future would be that everyone should have a different headline and a different new stream. We already achieved that for Wall Street and we're really seeing that happening for crypto as well.

Joules: Yeah, I love that you bring up a lot of really great points that make sense. I especially love what you talked about with the cycles and how crypto has really grown and now we kind of need that infrastructure that is almost like Wall Street level infrastructure. And it makes sense because I remember back when I started, it was easy to just go on crypto Twitter and I just see all the alpha I needed or I can be in a few different Discord channels and I got everything I needed. And I feel like at one point, that was easy to kind of do by yourself. And now it's just so much information and so much noise. And that's really where AI comes in to really just sort out all of that. So much so that you brought up another really good point. We're in an age where we have so much information that literally humans can't even keep up. Having those real human analysts that can go in and make these reports just isn't even scalable with how much data that we have. That is crazy. And I can't imagine what it's going to be like a few years from now. So being able to kind of get in on the ground and really start to build this is amazing. 

And you guys, if you haven't checked out SoSoValue and their news feeds, definitely do. It's pretty cool. That brings me into the next topic that I really wanted to discuss with you, which is these smarter feeds. I'd love to know how AI is powering these smarter feeds. For example, we have this new SoSoValue x Sahara AI bot, this new feed. I'd love to know behind the hood, how does that actually work? How does it filter this mountain of market data and surface the updates that are only relevant to you and your portfolio, your risk tolerance, what have you?

Jessie: Yeah. So actually in the very beginning when we come up with that newsfeed, all of their work is kind of dirty and not sexy at all. Before talking about personalization, I think in crypto newsfeed even step one is a really big task in terms of finding out what is real news, what is pure opinions, especially based on tweets, and what is a long-form research piece and what is just noise, the paid content and the noise. Even sorting through that in 10 different languages that we offer on the site and different news outlets formats, we have Discord, Twitter, the Googles of the world and then all the AI news feeds and then all of the crypto plus financial news feeds, even sorting through that, tagging them accordingly, correctly, consistently. We have spent a lot of resources and time on that collectively as a team. And once it's all tagged, which you can also see on our AI bot and our website as well, we're also attributing to all of the news channels because for one particular news, you probably will be reported in many different ways. And then how do you recognize that it's the same news rather than pushing it 10 times? That is also another industrial challenge that we're facing. 

Again, not sexy at all, but some of the fundamental layers that we have to sort through. And once we have that in place, we are able to get through a little bit more nitty-gritty. That's a more fun part of the challenge. We're sorting through customer behaviors. The first layer is language. Where do you come from? What kind of tokens or pages have you been interested in putting into your watch list? Or are you more of an ETF dashboard person or are you more a meme coin only investor where we are able to gather a lot more customer insights through our websites we have been running for a year and a half now at least. So based on the insight we're able to sort through, "Hey, this might be the information that you would need more on a daily basis." And then so, that's why we have the "For You" tag. 

And on the bot as well, you can configure on a very basic level what kind of information you need at what frequency, do you need more of a price alert or if you're more of a macro news person and in what language. So yeah, this is the three-step change that we're bringing into the crypto news AI world. It's been through a lot of time and sweat into this. Hopefully we're still evolving as AI is constantly. Every single new model happens, we have to reconsider if our model is the latest one and more up to date and how to scale it up at a reasonable cost is also another challenge. So yeah, behind the scenes there is a lot. Even though the front end might be a very simple Twitter bot or a TG bot but behind the scenes is a lot of work.

Joules: Awesome. I'm curious with all of this customization, do you have any examples of success stories where this level of personalization and this AI influence has really helped traders cut through that noise and make better decisions?

Jessie: Yeah, actually funny enough, the features that people thank us the most are usually the more simple and less of a fancier one. For example, the DAX and on-chain data we actually spent a lot of money sorting out, but the feature that's mostly used and people thank us for is still the simple stuff like sector rotation and how to track it across a period of time on the entire crypto Internet. We haven't really found this information and that's exactly why we created it. First of all, if you go to our front page, you will see a very top-down structured information architecture, kind of how we would process the market in the traditional stock market. What is the overall market cap? How does that translate into different narratives and different sectors? And how is that sector particularly performing over a period of time? And against BTC where we believe will be the beta and against Ethereum, a very simple chart, but no one else is doing it. For example, if I ask you what is the market dominance of layer 1 or NFT or even AI, as for our topic today, you'll be able to find out on our front page very easily and then to track its performance. And then when you click into that specific sector, you'll be able to see a whole basket that we categorize within the narrative of our sector. So this will help investors very easily track their rotation. Where is the market feeling right now? 

For example, as of today, AI has jumped 2.98% the market dominance like 0.36 percentage for example. And within AI you can see what tokens follow through it. So this is super simple, but again, dirty work, a very logical framework that no one else cared for to do it a year and a bit back. So we did it and people loved it. And a lot of media partners are quoting it for free every day without our logo on it. So there's a lot of free media mentioned there by influencers and crypto media and even macro media in the traditional stock market. And another piece that we know in the space is the ETF Net Inflow. Again, a super simple indicator. But even to come up with the indicator of the net inflow as the singular one metric to look at is where we add value, because there are so many different things to track. And on a given day for an ETF, BTC, spot, ETF one topic, you have 10 different charts. But in the end, you will see this one chart that we created from day one is the one that's mostly fast, widely circulating because we came up with this one easy thing for investors to track. 

And why does it matter? Because it signifies how much money is coming from the traditional stock market into the crypto market, the more long-term, healthy, reasonable risk profile money coming from stock into crypto. That's why we think it matters the most. And that's why we plot it in the way we do to help you track the daily net inflow. And as soon as all of the numbers, the Blackrock, the Neck of the world, publish a number, we will aggregate it and plot it automatically using our own proprietary framework so you can see it in a really timely manner. For example, today the total net inflow would be 65.95 million. And once the number hits today, we will update it immediately. And this has time and time again proven to be the forward indicator of where the BTC price would go. And the same for Ethereum. So again, simple is the best and to achieve simple is not easy. That's what we find out.

Joules: Yeah, I feel like the simpler you try to make something, it's usually the harder it is to do it. It's the same in content, the same in engineering, same in pretty much everything. But it's always the simplest things that are the most impactful and effective, as you're well aware of. That's cool. I love that, 100%. And I love when you can get free media attention too. So that's always a good plus. I wrote a note here because I love what you said here was "behind the scenes, it's not sexy at all." I completely agree with you. Behind the scenes, a lot of it is always so messy and crazy. I think one thing that I want to point out between a lot of the AI projects that we see today, when you think of AI, you think of these sexy interfaces where you're like actively interacting with the AI, right? You're like asking Chat GPT, "Hey, let me paste my investment strategy. Is this a good investment strategy?" And you're actively interacting with it. A lot of what you guys are doing and a lot of what's going on in the bot that we have together, the AI is actively working in the background. You don't have to actively prompt it or anything like that. It's just kind of quietly listening to your social media feeds and looking at what you're doing and kind of adjusting based on that. And I'm curious about your opinions. What do you think about this invisible assistant version, this model, what do you think about that and the influence of AI? And do you think it's more valuable for people, especially traders, than manually having to interact with AI?

Jessie: Yeah, that's a really good question. How to interface with the AI, I think it fits different types of investors at different times. We actually do have both. One is the more actively providing you what you need, the proactive AI. On our website as well, we have Socatis AI, which is more of a natural language interacting with your GPTs of the world. And what's different is we're feeding the AI with all of the real-time market data, all of the real-time news, all tagged, sorted and cleaned up. We're feeding that stream and all of the live charts as well. We're feeding that to our AI agent model so you can interact with it in real time and ask very pointed questions and it will structure an answer in a way that makes sense to investors. Kind of like a report. Plus all of the charts plugged in and every single market data piece is real time. So if you ask a Chat GPT a question regarding price, it might be a lag, but our AI model is the latest. But we see people interact with these two things in a very different way. So there is certain information that just makes sense to be pushed to you because of the whole high relevancy and timeliness requests, especially for price alerts. Two seconds later or two minutes later might not be relevant anymore. So for those more timely, relevant stuff, you generally would frame it in a passive push. So you don't really need to do anything. And for that to be highly relevant to individuals, a lot of the inferences need to be made and a lot of the personalized data needs to be put for that model to be relevant. Otherwise it's just spam. For example, if you don't really follow a certain type of altcoins and I push 100 different price alerts to you, it will be spam.

So for different types of investors, different information delivered to you in a timely manner, we kind of call it the mom model, "mom knows best," it would just give you furs or give you a jacket whenever you need. And if it's overcaring, it becomes nagging and annoying. So we kind of want to make it into a friend that knows your needs the most. For the active pushing ones and for the passive natural language model ones, what it needs is to serve investors with a clear ask in their head. For example, for Chat GPT, if you have a really good question versus a really vague and general one, you will have a completely different set of answers. So in the other, the input-based model, what we're trying to do is to help investors ask the better question. So we have some passive prompts, for example, to help you come to the answer that you need. And based on your question as well, we will guide you through, "Oh, is this what you want to ask? Do you want to know more about this?" to slowly guide investors, especially the one that's less familiar with crypto, to ask the right question, to get to the right level of information. So yeah, in terms of input, output, in terms of cadences and type of investors, I believe these two models have their own benefits and downfalls. And we're quite careful in terms of, especially in the passive pushing side, what information to push at what time would be the most relevant challenge to walk through. That's where we are working together with Sahara AI to perfect that experience.

Joules: Yeah. So you guys are really offering the best of both worlds and I love it. I love what you guys do. It's why we partnered with you because you're doing amazing work. I do want to make sure we are running a little low on time. So if you have any questions, please ask them in the chat. We already have a few, so let me start going through those before we run out of time. I do have somebody asking. There's a ton of AI-powered news products on the market. What makes SoSoValue so special?

Jessie: Oh, that's a really great question. I guess we're back to question one. We think what makes us different or sets us apart is where we're coming from in the very first place, and we haven't wavered from it since day one, is that we are extremely investor-centric. So whenever we turn out a new feature or a new model or a new AI bot, the first question we ask is how would it help investors to stay on top of the market, to make the right decision at the right time and to enjoy crypto gain in the long run? How do we serve that purpose? And that's exactly how we come up with the three product pillars. We have the SoSoValue terminal, kind of like an information hub to help you understand crypto and to have insight as actionable. And we have the SSI, which is the SoSoValue index, which is an on-chain spot. You can think of it as an on-chain spot ETF where investors, no matter where you're coming from, can easily participate in crypto. Because we learned from the stock market, the Warren Buffett of the world told us for average investors, the best way to earn in the long run is to buy and hold indexes that passively. The passive strategy. 

So that's exactly why we invented that. And we have been running it for a year and a half now. We have 200 million TVL within the first few months of launch. So there's a natural product market fit there. And as of now, I think it has consistently been in the market, especially the Max 7. And a third piece, we're launching Sodex, which is our own on-chain order book, Dex, allowing people to buy and yeah, basically buy assets that they like, premium assets that they like at ease in an easier manner. So the all three pillars all serve one purpose where we want to democratize assets when it comes to information or investment assets we want to offer to investors. So that really also comes into all of the news feeds and AI bots that we created. We were all answering that question because there are enough casinos out there. There's enough trading bots, there's enough a way to help you generate yields, for example, or to do very specific things. But what we're seeing is lacking in the market is a pretty holistic approach. We want to solve that information gap in the first place. And that's exactly where we're coming from in terms of our AI offering here.

Joules: Awesome. We do have a few more questions. I'll try to summarize some of them because we are running low on time. There's a few that are asking, "How do you make sure the bot doesn't filter out something important that I should know?"

Jessie: That's a million-dollar question. Behind this one simple question, we have done a lot of iteration. I don't know how many people have been with us since day one. We actually improved a lot since our first model. We're definitely improving in our back-end inferencing model. And we're also improving with every single new AI model coming out. We're also trying out new things and how to solve exactly that problem. There are two problems, two ways to predict the problem. And one is the over-information, the over-pushing information to you and sorting through what is relevant is that one. And the second is how do you make sure that you have the overkill? I think the overkill one is slightly easier to solve because we actually have a lot of alarms and different layers of approval in place. For example, at first we actually have a pool of information that's properly tagged and sorted with our AI. Then once we realize if the AI agent has negatively filtered our certain information, our human agent will step in to verify whether the information is relevant and whether it should be pushed live. And we have a very easy mechanism for us to fact-check and then in effect to push the news in a very timely manner, like within minutes I would say, because we're sourcing from multiple media sources. Usually if the news is super important everywhere, cross-check that better. Different weighted news sources. We actually have our own weighting system to give credibility to different media sources. And if more grade A media sources have reported on it and different influencers have mentioned the same topic, we usually will push it through. But if it's below that threshold, we will involve the human agent into fact-checking and relevance gauging before you push it live. So that's how we balance out information accuracy, relevance, and timeliness triangle.

Joules: Awesome. That's really cool. I'll ask one more question. Let's see. "Could this same concept work for other fast-moving industries beyond finance?"

Jessie: I believe so and I believe it's already happening. We kind of have an AI agent that's for each vertical that's already developed by aspiring people in the field. And we really want to stay in our lane because we all know that AI works best the more niche it is. And to be honest, in the crypto world today is already a really broad data source we have to source from. It's the financial market data, the macro data, the crypto data and everything crypto. We have the crypto media, crypto KOLs like Twitter, Discord and everything. So even within this niche, we have a lot of expertise that needs to come into place to gauge what is relevant and what is good. We actually have our content committee. We have more than 10 people, some of them are actually our active users giving us feedback and helping us tag and improve the new sources to sort out what is relevant. And there's a lot of upkeeping as well. There would be a crypto media or KOL out there that was really relevant and churning out great reports in one cycle and the next day they drop that or start taking money from projects and become like pure ads. So we actually have involved a lot of constant updating and management of the new source and pipeline. So we believe that the more focused that we are, the better outcome it is, especially in AI, right? "Garbage in, garbage out." So we, so that's why we want to stay hyper-focused, hyper-relevant and to go back to our mission where we want to serve investors well. And that's how we structure our resources and time on a daily basis.

Joules: It's crazy. You mentioned how crazy crypto kind of already is with information overload. Because I was just thinking about, I was talking to some friends in other sectors, other verticals, my non-crypto marketing friends and it's crazy how much more just being in crypto, you have to be an expert in everything and not just crypto, but in finance and tech and you have to be kind of partially a developer. It's so much. But I love it. I wouldn't trade it for the world. That's so cool.

Jessie: Cool.

Joules: Yeah. Thank you so much again to you, Jessie from SoSoValue for joining us. And to all of you who tuned in today, if you haven't tried it yet, please check out the SoSoValue News Bot on Telegram, powered by Sahara AI. It's free for the first month and it's built to give you only the updates that matter most to you. No noise, no doom scrolling, just good old-fashioned updates. Jessie, do you have anything else that you want to highlight? Anything you want to shout out?

Jessie: I'm not sure if it's my Internet or... Yeah, but I think I got the gist of it. Thank you.

Joules: Is there anything you, is there anything you wanted to shout out before we leave?

Jessie: Yes, definitely. Try out the bot that we created together with SoSoValue and Sahara AI. And then, yeah, stay tuned and use our product and give us feedback and we'd like to grow and learn.

Joules: Yeah. Thank you everybody. We'll see you next time.

Jessie: Thank you so much. See you next time. Bye.

AI