Ep 556: Choosing the Right AI:  Agents, LLMs, or Algorithms?

Topics Covered in This Episode:

1. Choosing AI: Algorithms vs. Agents
2. Understanding AI Models and Agents
3. Using Conditional Statements in AI
4. Importance of Data in AI Training
5. Risk Factors in Agentic AI Projects
6. Innovation through AI Experimentation
7. Evaluating AI for Business Solutions

Keywords:

AI algorithms, Large Language Models, LLMs, Agents, Agentic AI, Multi agentic AI, Amazon Web Services, AWS, Vazhi Philemon, GenAI efforts, Amazon Bedrock, talent wars in tech, OpenAI, Google, Meta, Copyright lawsuit, AI training, Sarah Silverman, Llama, Fair use in AI, Anthropic, AI deep research model, API, Webhooks, MCP, Code interpreter, Keymaker, Data labeling, Training datasets, Computer vision models, Block out time to experiment, Decision-making, If else conditional statements, Data-driven approach, AGI, Teleporting, Innovation in AI, Experiment with AI, Business leaders, Performance improvements, Sustainable business models, Corporate blade.

Choosing AI: Algorithms vs. Agents

When selecting artificial intelligence solutions, it’s crucial to understand the fundamental difference between algorithms and agents. Traditional AI algorithms follow predefined rules to solve specific tasks—think of them as structured decision trees that handle one problem at a time. These are great for stable, repetitive use cases where the conditions and variables are known. In contrast, agentic AI introduces a dynamic layer: it combines multiple models that can work together, often supervised by another model, to autonomously determine how to tackle complex tasks. Agents are designed to adapt, make decisions, and even change their decision-making logic based on evolving input and goals. This flexibility makes agents powerful for use cases involving uncertainty, multi-step operations, or environments that change in real time. However, with this flexibility comes additional complexity, higher costs, and greater risk, particularly for organizations unprepared to manage or monitor these systems. Businesses should carefully evaluate whether they need the predictability of algorithms or the adaptive power of agents, keeping in mind their problem space, operational readiness, and appetite for innovation. Ultimately, the right choice depends on aligning the AI type with business objectives and risk tolerance.

Understanding AI Models and Agents

AI models and agents serve distinct but complementary purposes. A model, such as a large language model (LLM), functions like a “black box” that processes input and generates output based on training data. Models are often used for single, focused tasks like generating text, recognizing images, or making predictions. They don’t make independent decisions beyond the scope of their training—they simply respond to requests. In contrast, agents are systems designed to coordinate multiple models or processes toward achieving a goal. Think of an agent as a manager overseeing a team of models, using conditional logic to determine which model to engage and when. For example, an agent designed to plan a trip could use one model to check prices, another for weather, and another for visa requirements, integrating all outputs to make recommendations. Agents can even be enhanced with agentic AI, where supervisory models dynamically build or adjust the logic tree based on the task at hand. Understanding these differences is critical for businesses: models provide power and precision for defined tasks, while agents bring orchestration and adaptability for complex, multi-step challenges.

Using Conditional Statements in AI

Conditional statements, often described as “if/else” logic, are foundational in AI systems, especially in agents and algorithms. This simple but powerful structure enables AI to make decisions based on specific conditions. For example, in a customer service chatbot, a conditional statement might say: if the customer asks about a refund, then display the return policy; else provide the general FAQ link. In agentic AI, conditional logic can become highly complex, linking dozens of models or data points together. Conditional statements allow AI to follow branching paths depending on input and context, much like how humans make decisions based on different scenarios. In advanced systems, these conditions can be hard-coded by developers or dynamically generated by a supervising model in agentic architectures. This dynamic generation allows the AI to build or modify decision trees in real time, offering a level of flexibility that static algorithms can’t match. Understanding and managing conditional logic is key to building reliable AI solutions, as it ensures the system responds appropriately across varied scenarios without creating unintended outcomes.

Importance of Data in AI Training

Data is the cornerstone of any successful AI system. High-quality, well-labeled, and unbiased data ensures that models learn accurate patterns and that agents make sound decisions. Poor data, on the other hand, leads to unreliable outputs, biased results, and underperforming systems. In traditional algorithms, data informs the logic and thresholds for decision-making. In machine learning models, data directly shapes the model’s understanding of relationships and patterns. For agents and agentic AI, data quality is even more critical because these systems rely on multiple models interacting—each influenced by its own training data. Inadequate or inconsistent data at any stage can derail the entire system’s performance. Furthermore, data-driven decisions must be interpreted responsibly. As Michael Abramov noted in the podcast, being “pseudo data-driven” (where conclusions are drawn to fit a narrative rather than truly reflecting the data) can be more dangerous than not being data-driven at all. Business leaders must prioritize data integrity, ensure datasets are representative and accurate, and continuously validate outputs to align with business goals and ethical standards.

Risk Factors in Agentic AI Projects

Agentic AI offers exciting possibilities, but it introduces significant risks that businesses must understand and manage. Unlike traditional algorithms or standalone models, agentic systems operate with a degree of autonomy that can lead to unexpected or undesired behaviors. One major risk is complexity: as agentic AI systems combine multiple models with dynamic decision-making layers, they become harder to monitor, debug, and control. This complexity increases operational costs and can create vulnerabilities in safety, compliance, or brand reputation. Another key risk is sustainability. As Gartner reported, up to 40% of agentic AI projects could be canceled by 2027 due to high costs, unclear business value, or inadequate risk management. There is also competitive risk: large tech firms can quickly replicate and integrate agentic features into their platforms, undercutting smaller startups that attempt to commercialize niche agentic solutions. To mitigate these risks, organizations need strong human-in-the-loop practices, clear governance frameworks, and ongoing evaluation of both technical and business outcomes.

Innovation through AI Experimentation

Continuous experimentation is the engine of AI-driven innovation. As emphasized in the podcast, business leaders shouldn’t lock themselves into one tool, model, or architecture too early. Instead, they should foster a culture where teams regularly test new AI solutions—whether algorithms, LLMs, or agentic systems—against real business challenges. Experimentation helps uncover what truly works and what doesn’t, beyond the hype of the latest AI trend. It allows companies to discover unexpected efficiencies, build custom tools that address specific pain points, and stay ahead of competitors. Successful AI experimentation involves not only trying out models and tools but also sharing results internally to spark wider innovation. Leaders should encourage employees at all levels to explore AI tools, document learnings, and pilot new ideas. The key is balancing innovation with operational stability: experimentation should complement core business processes, not disrupt them without purpose. By making experimentation part of the organizational DNA, companies can stay adaptable and prepared for future AI breakthroughs.

Evaluating AI for Business Solutions

Choosing the right AI for your business requires thoughtful evaluation, not just of the technology itself but also of your company’s goals, capabilities, and constraints. Leaders need to start by clearly defining the problem they’re trying to solve—then assess which AI type aligns best. Algorithms work well for repetitive, rules-based tasks. LLMs are ideal for generating and interpreting language, offering natural language interaction at scale. Agents and agentic AI bring the power to coordinate complex, multi-step processes, making decisions dynamically. However, the more advanced the AI solution, the greater the need for data quality, governance, monitoring, and risk management. Businesses should also consider costs, team expertise, integration complexity, and long-term sustainability when selecting AI tools. Michael Abramov’s key advice is to experiment widely and choose what addresses your actual business needs, not just what’s trending. Evaluate AI as a set of tools to be orchestrated thoughtfully—not as one-size-fits-all solutions.

Podcast Transcript

Jordan Wilson [00:00:46]:
It seems like AI is sometimes just alphabet soup of the buzzword of the day. Right? Yeah. We all want to use AI and Jet AI and LLMs and what happens when AGI comes or ASI. Right? But I think it's important to first understand the basics. Right? And not just rush toward what everyone else is using. So today, we're gonna be breaking it down and how to choose the right AI. And we're gonna be talking about algorithms, agents, and large language models. I'm excited for today's conversation.Jordan Wilson [00:01:25]:
I hope you are too. What's going on y'all? Welcome to Everyday AI. My name is Jordan Wilson, and this thing is your daily livestream podcast and free daily newsletter, helping us all not just keep up with what's happening in the world of AI, but how we can all actually use this information to get ahead to grow our companies and our careers. If that's what you're doing, you're in the right place. It starts here with the unedited, unscripted live streaming podcast, but where you are actually going to grow is on our website. So please go to youreverydayai.com. There, you can, yeah, go, listen to and watch more than 550 episodes for free. It's a free generative AI university, but, also, you need to sign up for today's, daily newsletter.Jordan Wilson [00:02:05]:
It is free. We're gonna be recapping not just the best insights from today's episode, but also everything else you need to be the smartest person in AI in your company. Alright. Before we get started, I'm bringing back the AI news for today. Podcast audience, let me know in livestream audience. Sometimes I do the the the news right before. Sometimes I don't. I've taken a little bit of break.Jordan Wilson [00:02:26]:
Let me know if you want it back, but let's just go ahead and, go into the AI news for today, June 27. So first, Amazon Web Services has lost a pivotal AI leader as the talent wars intensify in techs, in tech. So, AWS has lost Vazhi Philemon, a vice president who helped lead its GenAI efforts and the Amazon Bedrock platform according to reports. So Philemon's departure follows eight years at Amazon incomes as competition for top AI talent accelerates with companies like OpenAI and Google setting the pace and reportedly sometimes offering companies and meta. We've been talking about all this buzz of a $100,000,000 annual contracts. So Amazon continues to invest heavily in AI, including an $8,000,000,000 stake in startup Anthropic. Amazon has recently rolled out Nova and Sonic AI models, expanding capabilities in text, video, and image generation. According to its CEO, Andy Jassy, Amazon advances in a Gentic AI could lead to fewer traditional corporate jobs as automation replaces some tasks even as demand grows for new roles in AI development.Jordan Wilson [00:03:43]:
Alright. Our next piece of AI news, Meta has won a key copyright lawsuit over AI training. So a US judge has ruled in favor of Meta, dismissing a copyright lawsuit filed by authors including Sarah Silverman and, T. Nesi Coates. Hopefully, I didn't get that name wrong, who alleged that Meta unlawfully used their books to train its AI model, llama. So the judge said the plaintiffs failed to show that Meta's AI would harm the market for their works, leading him to call Meta's use of the material fair use under current copyright law. However, the judge did emphasize that his ruling does not mean all AI training on copyrighted material is legal, noting that using such works without permission could be unlawful in many situations. And this comes, just kind of hours after Anthropic won a similar, ruling about their use of, essentially training their model on books.Jordan Wilson [00:04:42]:
Alright. Last but not least, a little one for developers. So OpenAI has unveiled its deep research model and webhooks for its API. So OpenAI has announced two major updates to its API, the introduction of its deep research mode and support for webhooks, the company just announced on Twitter. So the new o3 deep research and o four mini deep research models are the same advanced version that power deep research within ChatGPT. So, essentially, you have a version in the o3 or the o four mini. So if you are one of the countless companies building on top of OpenAI or maybe if you're using services like I mean, you probably don't even know, but your bank is probably using OpenAI's API. Right.Jordan Wilson [00:05:31]:
So now, the capabilities are really gonna be expanded now with this deep research rolling out to the API. The models also come equipped with features like MCP and a built in code interpreter. And with the rollout of webhooks, developers can now receive real time notifications for key API events, including competed responses and fine tuning jobs. Alright. For those stories and a lot more, again, go to our website at youreverydayai.com and, check out in the newsletter. Alright. Let's get to the real stuff here. How the heck do you choose the right AI? Should we all be using agents, multi age like, multi agentic workflows, agentic workflows, traditional algorithms, large language models.Jordan Wilson [00:06:18]:
I don't know. It's a question that we're always, talking about and something business leaders are always tasked with. So, let's bring on someone to help guide us, through today's conversation. So, please help me welcome to the show livestream audience, if you don't mind. We have Michael Abramov, the CEO of Keymaker and Key Labs. Michael, thank you so much for joining the Everyday AI Show.Michael Abramov [00:06:40]:
Hi. Hi. Thank you for inviting me. Alright. Well, thanks thanks for having us, Michael. So first, before weJordan Wilson [00:06:46]:
get into it, can you tell us a little bit what you do at KeyMaker and KeyLabs?Michael Abramov [00:06:50]:
So, actually yeah. So KeyMaker and KeyLabs are data labeling plat platform and the and the, service provider. So what we do is we prepare datasets for training the models, whether it's computer vision models, whether it's LOM models, or any other, AI. Actually, we are preparing the whole like, all the training materials for for that.Jordan Wilson [00:07:13]:
Yeah. So and yeah. The the the the training materials and the data, it's it's the hot topic. Right? Even mentioning that in the, in the AI news today, but maybe let's let's fast forward to the end. Right? So for companies, you know, it it's hard to keep up with everything that's available. So there's, you know, traditional AI algorithms. There's there's large language models. Now there's these models, these large language models on the consumer side that are agentic, and then you have literally agents.Jordan Wilson [00:07:42]:
So where do companies start? How do you choose the right AI?Michael Abramov [00:07:47]:
Oh, that's amazing question, and, I don't have a a, like, very specific answer to that because, you know, it's like choosing your life partner. When you're choosing your life partner for life, you might be 20 years old, or, like, any other age, and then, your requirements at that specific moment are for something specific, but then, your requirements change over time. Right? And you wanna choose the the person who's gonna be suitable for different age ages. Right? The same with AI. You have to think about it. Do I choose it for short term? Do I choose it for long term? Do I, you know, what kind of tasks it should, solve? Also, there is some kind of, overpromising, on in social networks because you I mean, people like us look at LinkedIn every day, right, on the LinkedIn timeline, and you see people saying, oh, I just, you know, I just replaced 50 employees with this. I just replaced six employees with that. And you kind of sit and think, hey.Michael Abramov [00:08:54]:
What am I doing wrong? Where did I, you knowJordan Wilson [00:08:57]:
Yeah. You're why do I still have employees? Why like, right? Like, whatMichael Abramov [00:09:00]:
am I doing? Yeah. Right. How does it work? Like, what where did I did I see? And then you try to do it. You try to play with some tools, and they are they they might help you here. They might not help you there, and you you know? So it's it's very it's actually very hard, and I think there is no one one very specific answer. But, what I do believe you like, every person, nowadays, should do is, play a lot with AI. I mean I mean it. Play it yourself.Michael Abramov [00:09:32]:
Don't listen to anyone. It's not like, okay. I'm a CEO. Right? And I I'm managing a company with 480 people. And, my my team is pretty big. I have pretty pretty busy day, and I don't have time for for stuff. But my calendar is always closed for half an hour per day to just to play with tools, just to play with things. Okay? And when you do, just register to all the platforms, try to see how it can help you.Michael Abramov [00:10:00]:
You're always gonna go to the to to your biggest pains. That's, like, that's natural. That's organic. I mean, you don't have to see it and think, like, what am I doing? Like, what am I testing first? My emails, you know, elaboration the emails or my Slack problem. You will always go to the things that, you know, the first first, it will always go to things that are painful, the the most painful. Now after you play with it by yourself and you see how it helps or it doesn't help you, it ended up now, Jordan, for for for for for reference, it doesn't matter whether it's agent, agentic, LLM, BLM. I mean, just put all this terminology aside. Right? Try to find what helps you, what solves your pain, your problem.Michael Abramov [00:10:48]:
Once you do, give an example to your employees, to your to your peers, to your colleagues. You can just come up and say, hey. I'm using this. I'm doing this, and it helps me. Mhmm. Right? And then and then it's you know, you're gonna you're gonna see how to you know, you're gonna see what's best for you.Jordan Wilson [00:11:06]:
Yeah. I think I think there are some great points there. Right? It it sounds simple, but I love that it's like you're actually blocking out the time, experimenting with the latest technology, seeing what works, and then sharing about it. That's that's literally, you know, a great, kind of road map. It's a similar road map that we share all the time, so I love that. But let's even though you don't need to, you you know, define everything, I think it'll be helpful for our audience because it can be confusing. Right, because you hear these things about agents, and then you hear these things about multi agentic AI. Yeah.Jordan Wilson [00:11:42]:
Then here, large language models. And now these large language models have agentic capabilities. Right? Help us with the definitions. What the heck is an agent? What is a large language model?Michael Abramov [00:11:52]:
What's an algorithm? Yeah. That's that's amazing, question. And, I think that the okay. We we speak here about three things, first of all. LLM is, think about it just like a black box, right, that you can ask any question and it can answer you any, answer. And that's it. That's all it can do. So let's call it model.Michael Abramov [00:12:13]:
And, agents, think about it as a automated multimodels. So you you can have, like, five or 10 or whatever amount of models, and then you can put them on the dashboard and draw some arrows from one to another with if and else. I think many people are programmers, like, software developers here and might understand the the the if else terminology. But, anyways, anybody else can also understand that. And, now think about it. You want to go to Europe and you have built your, travel agent. Right? So so you ask, hey. What about going to Paris next week? So what's what is it what is it gonna do? It's gonna go to model that checks your, bank account if you have enough money.Michael Abramov [00:12:58]:
It's gonna go to weather, model that will check the weather in Paris, and we'll see if it's good for you or not. Right? So so let's think about agent as, just multiple models that know how to speak to each other and, how to do this, you know, if else inside the the agent. Now agentic AI is a much more interesting concept. It's it's a concept. It's not a, you know, specific tool or specific thing. Agentic AI is when you have an agent, but all of these if elses are being decided by another model. So you don't it's not hard coded that you have to check the bank account and the weather in Paris and the flights on, you know, different, flight companies. So the the the another model, which let's call it supervisor model.Michael Abramov [00:13:50]:
We'll look at the problem at the at the task, and we'll decide what if else is it wants to put inside the thing and where it wanna it wants to go and what it wants to do, and then apply agents inside the system.Jordan Wilson [00:14:05]:
Yeah. So help us break down a little bit more for our our nontechnical audience. Yeah. Explain the if else, you you know, conditional statement. Like, what does that mean, specifically in the context of, you you know, AI or, you know, large language models. Right? But explain that if else conditional statement as it pertains to AI.Michael Abramov [00:14:24]:
So every yeah. Okay. Okay. So every, kind of decision will go through through some kind of decision trees, and you can take an example of, your day to day life. So you you you wake up in the morning, and, you want to take your child to a school. Right? Let's say, your kid to to school. Now you have a lot of if else's. So first, if else, did did he wake up or she? Right? If yes, ask them to take a breakfast.Michael Abramov [00:14:56]:
If no, wake them up first. Right? And then do they have a fever? Maybe they have a fever. Right? If yes, stay at home and then, you know, rest. If no, let's go to, you know, let's brush teeth and go to school. Stuff like that. So there's lots of decisions. Let's call decision three, because it's hierarchical. Usually, it's it goes, like, from top to down or or whatever direction you choose.Michael Abramov [00:15:21]:
But there is initial condition, And then there are lots of final, items that are dependent on your decisions that you made on the way. And all of these decisions will be made by if and else. Mhmm. If this then else something else.Jordan Wilson [00:15:42]:
Yeah. I think I think that's a great way, and even I I love the example that that you just played, but that that just goes to show and emphasize ultimately how right? Even, like, what your company does, like data. Right? And making sure that you have the structured data, to help answer those if else conditions. Right? Like, you you know, whether it comes to, you know, traditional algorithms, you you know, large language models, agentic AI. Right? Whatever it is. How important is having your data correct?Michael Abramov [00:16:12]:
Oh, that's the most important thing. So, I had this problem in my company that I was, requesting from all of the people who work with me to be data driven. Now let let me just explain what data driven we mean. Sometimes you can come and tell me, hey, Michael. Most of the people in the world are afraid to lose their job now. Okay? And, when you say most of the people in the world, like, how many? Like, is it 88%? Is it 90%? Is it business people? Is it people from United States? Is it like, what kind of people are you talking about? Now the second thing is, what does the most mean? Like, where did you take this information from? Is it reliable? Can you rely on this information? Okay. You got it from Gartner or you got it from Google search. So, like, who wrote it? Right? And, maybe it's your, your personal, you know, your personal afraid, and you you are afraid of losing the job.Michael Abramov [00:17:08]:
And now you extrapolate this this, you know, emotion on most of the people in the world. Right? I don't know. That's that's, so you have to be data driven. You have to come and say that, you know, I read, like, 12 different articles from different sources, and that's, how I why I think that's true. Now, there is another problem too. So after I asked most of the people in the company to be data driven, and I explained what data driven means and how to acquire data and how to look at data, etcetera, I got reverse problem, the the mirror problem. People what they did was they were super data driven. They took data, and they, have built conclusions on the data, and they would come to me and say, hey.Michael Abramov [00:17:52]:
This is, you know, this is the problem, and here is the data that proves it. And that was super funny to see that the the data didn't prove it at all. It was their perception of the data that proves it. Okay? And, and that's if I give an example, we could say, you know, when you say, the sun has fallen and that's why the I don't know. I I even don't have an example. Sorry. But but but you know what I mean? I mean, you can relate things that are unrelated, and many people are tending to relate everything because people have to explain every single thing. And, so this pseudo data drivenness is even worse than not being data driven at all.Michael Abramov [00:18:46]:
And, yeah. So we prepare data for for machine learning training. We have lots of problems over there. We have lots of, misunderstandings of how the data should be structured, how it should be labeled, how it should be perceived. Even even if you structured it really well and you labeled it pretty pretty well, the so the developers of the model might, you know, do wrong things on the on the training,Jordan Wilson [00:19:17]:
stage. So, you you know, one thing you mentioned in there, Michael, was something about, you know, whether you're getting your information from from Google or Gartner. So I wanna ask you, here in a second about a recent Gartner, study on a Gentec AI. But before we do, we're gonna take a very quick break for a word from our sponsors.Jordan Wilson [00:20:08]:
So as we, you know, go back and forth between algorithms, large language models, and agents, it's it's no surprise over the past, you know, year or so, the rush has been toward agentic AI. And it seems like every single company, even if maybe they don't need it or if they don't even fully understand it, they're trying to dive all into a, like, agentic AI. And there's a recent Gartner study that we covered in our newsletter, yesterday about that predicted that 40% of agentic AI projects will be canceled by 2027 either due to, high cost, unclear business value, or inadequate risk controls. So I'm not gonna ask you to predict a a a percentage, Michael, but Mhmm. You know, one thing, that that caught me from that study is risk. Right? Can you talk a little bit about the difference between risk in large language models, algorithms versus agents? Because in my mind, agents, it can get kinda risky if you really aren't, kind of have a strong human in the loop connections.Michael Abramov [00:21:15]:
Yeah. So okay. So the risk is not in LLM or agents. The risk is if I I mean, we speak about the risk for the businesses who develop who who try to, you know, make money on developing these agents and selling them, to someone. Now, I have this idea of, I call it corporate blade. It's a huge blade that goes, like when the new when new technology evolves, everybody is running out to to implement some kind of, you know, raptors or or plugins or some kind of things on top of the new technology. Right? And we we see a lot of it with AI. Lots of my friends, lots of people out there are trying to build startups, with AI.Michael Abramov [00:22:05]:
Now what usually happens in the first stages of such technologies, when they are still not stable, Some projects die because of what you said, because they are not sustainable or something else. But some of them which are, which are, you know, successful, they also might fall. And they might fall because the huge the giants will take the idea, and for the giants to take the idea and to implement it is, like, one week, two weeks of of work. So if you think about perplexity, perplexity is a multimillion dollar start up, and I think it's multibillion dollar valuation already. But all it does is it's wrapping, ChargeGPT, and then it adds a little bit, better search, like, you know, capabilities on top of it, and also some some different user interaction dynamics. But it's nothing that ChargeGPT like, the OpenAI can can do in two weeks. Right? And if they like the idea, they can do it and just, you know, blade cut out all of these startups that did, you know, very interesting things. And we see it we see a lot of this.Michael Abramov [00:23:17]:
We see a lot of agents, a lot of, mass tutors, and, you know, personal assistants, psychother like, psych psychologists, you know, therapists, AI therapists, etcetera, or or calendar assistants that are being wiped out by the next feature of llama or or entropic or ChargeGPT, etcetera. And, maybe maybe Gartner's, research relates to that as well, not only, you know, to unsustainable businesses, or bad ideas.Jordan Wilson [00:23:54]:
Yeah. And, you know, so I'm I'm I'm curious. Even if if you could walk us through. So, right, you said, how many employees are at your company again? It's, four eighty.Michael Abramov [00:24:05]:
480 employees.Jordan Wilson [00:24:06]:
So 480 employees, and you obviously, you you know, specialize in data labeling for machine learning. Yeah. So I'm guessing that you've had a healthy amount of AI use, right, to make make that assumption. Right?Michael Abramov [00:24:20]:
Yeah.Jordan Wilson [00:24:21]:
How are you even deciding? As a CEO of a of a growing company, how are you deciding, hey. When do we step outside of the traditional, you know, decision tree algorithm to large language models to agents? How are you making those decisions?Michael Abramov [00:24:36]:
Push my people every day to innovations. Now that's something, special about, like, my team, because I think in the beginning, they were, not, I mean, they they were very curious about doing that, but at some point, they were exhausted because they said, hey. Too much innovation. I mean, let us, you know, be a little bit in a stable position for for some time. And, now we found a balance. So there is balance between, like, routine job and, coming up with innovations how to make this routine, work, like, more performant, let's say, etcetera. But the main thing here is that I'm not saying, hey. Come up with LLM idea or a agent idea or a Genica AI idea or something else.Michael Abramov [00:25:29]:
I'm not kind of trying to say use this or that tool. So I'm just saying, hey, guys. Here is the whole toolset. We even have once in a week technical education, kind of session where we show up new things, everyone shows to another, etcetera. And then I say, here is a tool set, and you have problems. You have every day you have daily problems. You have fires that you, you know, work with, etcetera. Just see if if if there is a better tool than you're using today.Michael Abramov [00:25:59]:
That's it. K? I I'm not I'm not asking every person to be a software developer, agent developer, or, you know, stuff like that. And you I mean, you'll be amazed by how many people who have never, you know, did any if else in their life, they take this this tools. They sometimes they ask questions. Sometimes they ask ChargeGPT, hey. I I have this problem. How do I solve it? Sometimes they ask their friends or or myself or or YouTube, and they build amazing things. Amazing.Michael Abramov [00:26:32]:
Many times I'm just saying, hey. This is a startup. I really have such examples, and I'm not gonna spend your time on that. But, I mean Yeah. That's amazing.Jordan Wilson [00:26:42]:
And if if we look at kind of the the linear progression, and, obviously, this goes over the course of of many decades. Right? But if we say algorithms led to large language models, which led to, agents. Right? And then there's a lot that probably comes after this.Michael Abramov [00:27:00]:
Yeah.Jordan Wilson [00:27:01]:
But what are some of those unknowns? Right? Because, obviously, the pace of of of innovation is going very quickly. We've heard the smartest people in the world, as an example, say, hey. Once we get to, quote, unquote, AGI and multi agentic AGI, things really escalate there. But what are those kind of, unknowns that may come after whatever next on that linear graph?Michael Abramov [00:27:25]:
Yeah. That's my favorite question. Because we have, we have known unknowns, which is like, oh, when we get to AGI, like, every every person will lose their job, and the AJI will do everything for us. And, in that case, we have two direct two directions. Direction number one, we're all gonna die because AJI is gonna kill us. And, direction number two, we are going to live on the welfare because AGI will work. We're gonna have, we're gonna get pay payments from the government, and we will have a lot of time for creativity to be, you know, artists, musicians, etcetera, etcetera, etcetera, because we don't have to work. We don't I mean, we have enough, resources to to exist without working.Michael Abramov [00:28:07]:
So and and there's lots of speculations, like, around these two ideas, and I would call it known unknowns. Now there are some unknown unknowns, and, a good example of it is imagine the times when people were flying on zeppelins, right, or on the balloons. I don't know how Balloons. Right? Yeah. Yeah. Air balloons. Yeah. And the the the capacity of one basket in the air balloon was, like, 20 people.Michael Abramov [00:28:32]:
And then the engineers, they were thinking, oh, in hundred years, people will fly on the air balloon with 100 people capacity. And in 200 people, it's gonna be 500 people capacity. Right? So it's gonna be like a bigger balloon. They had they could never, imagine an airplane. Right? That was the unknown unknown for them. That's something that you can't even imagine. Now, for us in AGI or in overall AI, I don't know how you you you name it. Right? There might be unknown unknowns in, I don't know, teleporting.Michael Abramov [00:29:06]:
Did you think about, like, teleporting? Would you like to be I I mean, for Oh, absolutely. Yeah. Yeah. I mean, either way. If I could choose the next feature, I I would love to purchase in this, game world, right, it was it would be teleporting or invisibility. Right? So maybe maybe we can get there. And and, maybe maybe we can get to something that we can't even think about right now. I don't know.Jordan Wilson [00:29:31]:
Yeah. So so so, Michael, we've we've covered a lot in today's conversation, But, you know, as as we wrap up, what is the most important, piece of advice or actionable insight, that you have for business leaders out there that are maybe just scratching their head when it comes to, you you know, algorithms, large language model, agentic AIMichael Abramov [00:29:55]:
Yeah. Agents. Like, what do people need to focus on to make the right decision for their business? Experiment, guys. I mean, people experiment, experiment, experiment. And your everyday should look like you you are starting up a new company. Even though your company is amazing and your company makes lots of money and you are super successful in in your company, you should think about it. Today, I'm gonna create a a something new. I'm gonna create some new idea and experiment with it.Michael Abramov [00:30:23]:
And then you can just implement it in your company. You don't have to open a new company for it. And, I would say that most of the interesting things that are happening, inside Keymaker in terms of, improving performances in the in the teams, etcetera, etcetera, are you can take it and make a separate start up out there. Mhmm. I mean, we won't spend time on it, but that's all because of experiment. Yeah. You have toJordan Wilson [00:30:49]:
experiment. Awesome. Amazing. Amazing advice, and a very insightful conversation. Michael, thank you so much for taking time out of your day to join the Everyday AI Show. We really appreciate it. Thank you. Bye bye.Jordan Wilson [00:31:01]:
Alright. And if you missed anything, don't worry. We've got it all for you. A lot of great insights and advice there from Michael. So if you miss anything, it's gonna be in our newsletter. If this was helpful, if you're listening on the podcast, please make sure to follow the show and subscribe. Drop us a note as well. And And then when you're done with that, go to youreverydayai.com.Jordan Wilson [00:31:20]:
So make sure to check out the recap for this podcast. We're gonna be dropping some additional information that we probably didn't have time to get to as well as keeping you up to date with everything else you need to know. So thank you for tuning in. Please join us next time for more everyday AI. Thanks, y'all.