240. What Non-Technical Innovators Need To Do About DeepSeek

DeepSeek is a new AI that's 44 times cheaper than ChatGPT and completely open-source.

While tech blogs obsess over the technical specs, there's a bigger story for non-technical innovators that no one else is talking about.

In this episode, you'll discover the hidden opportunities DeepSeek creates for founders and innovators who don't code.

Listen to learn:

• How to build genuine expertise around DeepSeek while others are still figuring it out (time sensitive)
• Smart ways to work with developers who are excited about DeepSeek
• Why DeepSeek's affordability means you can test your ideas without breaking the bank
• Practical steps to build custom AI solutions without needing a PhD

Whether you're a founder tired of paying premium prices for AI, a corporate innovator looking to stand out, or an entrepreneur wanting to experiment with AI, this episode shows you exactly what to do with DeepSeek right now.

Learn why the real DeepSeek opportunity isn't in the code - it's in how you use it to get ahead while everyone else is still talking tech specs.

Resources mentioned in this episode:

• 227. How to Use AI in Product Development: A Non-Techies Guide
• 196. Why All Businesses Depend on Open-Source Software

Timestamps

00:00 Introduction

02:59 DeepSeek vs. Other LLMs

06:14 Opportunities for Non-Technical Innovators

09:05 Open Source Advantages

11:56 Engaging with the Developer Community

15:14 Global Innovation with DeepSeek

For more career & tech lessons, subscribe to Tech for Non-Techies on:

• Apple
• Spotify
• YouTube
• Amazon Podcasts
• Stitcher
• Pandora

FREE COURSE:

5 Tech Concepts Every Business Leader Needs To Know

Growth Through Innovation

If your organisation wants to drive revenue through innovation, book a call with us here.

Our workshops and innovation strategies have helped Constellation Brands, the Royal Bank of Canada and Oxford University.

Transcript

Sophia Matveeva (00:00.04)
If you're a non-technical founder or if you're an innovator at a corporate and you want to know how DeepSeek can help you, then this is the lesson that you need to listen to and this is the lesson that you need to watch.

Sophia Matveeva (00:17.294)
Welcome to the Tech for our Techies podcast. I'm your host, Tech entrepreneur, executive coach and Chicago Booth MBA, Safiya Matheer. My aim here is to help you have a great career in the digital age. In a time when even your coffee shop has an app, you simply have to speak tech. On this podcast, I share core technology concepts, help you relate them to business outcomes.

And most importantly, share practical advice on what you can do to become a digital leader today. If you want to have a great career in the digital age, this podcast is for you. Hello, smart people. How are you today? This week, I have been playing around with the DeepSeek LLM. And to be honest, I've been finding it quite exciting. I'm excited about it because it basically works.

just as well for what I've been using it for as the big US model, basically just like Chad GPT and like Claude, but it is 44 times cheaper. Have you ever had a product when you then found basically a substitute that is just as good that costs 45 times less? I mean, everybody should find that exciting I think. Anyway, so

I'm sure you've seen that there's loads of news about DeepSeek around right now, but this lesson specifically is for the non-technical innovators. So if you're a non-technical founder or if you're an innovator at a corporate and you want to know how DeepSeek can help you, then this is the lesson that you need to listen to and this is the lesson that you need to watch. So here's a quick introduction.

So there is a Chinese hedge fund that has created this large language model called DeepSeek. And basically it has just come out into the world. So it's been used already, but it's just become basically super famous and massively downloaded. And I've been testing DeepSeek and ChatGPT and Claude to basically do the same tasks. And I've been comparing how they all work. And to be honest, DeepSeek is

Sophia Matveeva (02:31.36)
Really, really good, as I've mentioned. also DeepSeek, the company has chosen to make their code open source. And this basically means that you and developers can look at their code so like they can look at how the model has been built. And then based on that knowledge, you can build other stuff on top of it. And if you don't know what open source means and you've just heard that term bandied around.

That's totally normal. You know, if you're not a developer, why would you know about this? So I actually have a lesson on what open source means earlier on in the show. So we're going to paste a link to the open source lesson in the show notes for this one. But basically open source means that everybody can see it. And because everybody can see the code, that means you can essentially take that code and then you can build on top of that. So you can make

new things on top of what's already been created. basically the two main things you need to know about DeepSeek is that it's much cheaper and it's open source and both of these things have implications for you as a non-technical innovator. know Silicon Valley is up in arms, what are you going to do about it? So number one is that this is a really good opportunity to establish yourself as a thought leader.

on AI and digital transformation. And this is totally a hack. So there's massive opportunity to talk about DeepSeek now. And you don't even need to say anything like that. It's so intelligent. And people are going to listen and repost. And honestly, this is what's happening to me. So for example, I have been contacted by major international press for my views on DeepSeek versus the US platforms, literally as a user.

right? Not as somebody who is a technical expert. So if you just basically do that same comparison yourself today and write about it on LinkedIn for like a couple of paragraphs, the likelihood of you getting some traction and being noticed and being known as somebody who experiments with new technologies, that's basically a really cool site guys to focus on to capture right now rather.

Sophia Matveeva (04:52.716)
And if you're thinking, okay, but I don't know what to do, just do something really simple. Listen, you could just go into your kitchen right now and you could see what you have in there. And then you could go to Claude, chat GPT and DeepSeek, open all of them and to all of them say the same thing. Like you are a chef. This is what I have in my kitchen right now. Help me come up with three ideas for a healthy and nutritious lunch and just see what they do. And then.

You can write about it on your LinkedIn post that this is the output, this is what I learned, these are my impressions. And if you do that, tag me in on your LinkedIn, I would love to see how you're doing this. Okay, so that's number one. Number two, you can obviously just save money by switching to DeepSeek. So if you're already using ChatGPT, if you're already using Claude's Anthropic, you could get rid of one of them.

or get rid of both of them. So I'm paying for both of those and I am considering what I'm going to do because there's literally no point of paying for both of them whilst there's this really, really good, cheaper model around. So number three, talk to your developer connections about what they are doing with DeepSeek and what they're excited about and what they're planning to build. So as I mentioned, DeepSeek is an open source model.

And developers get really excited about this sort of thing. So I bet that if you're working in a corporate, there are developers in that corporate already thinking about, what can we do with this model? How can we adapt it to what we are doing in this company? So if you're working in a corporate, you're a corporate innovator, literally just go and find the developers and just say to them, guys, what are you doing with this? What do you think about this?

How could we use this? What are your ideas? And even if you don't understand what that should tell you, the fact that you are going and the fact that you're asking, often that's so meaningful because developers love talking about the stuff that they're inventing and business people often kind of just ignore that and their eyes glaze over when they think it's too scary. So even if you just go and you ask and you start the conversation, that's going to be a really good way to build relationships. And if you're not in a corporate, then

Sophia Matveeva (07:12.024)
find some developer friends and ask them the same thing. So first of all, this could be an opportunity for you to get involved in some really interesting frontier projects, right? Because in every tech company, there are the tech people and the non-tech people. So just going to find out what people are building on top of the DeepSeq source code could be an opportunity for you to get involved in something really interesting, really frontier, potentially experimental, potentially really successful. But even if that doesn't happen,

Having a community, like having your own developer friends is really, really useful for whatever you want to do. Whether you want to start a company, whether you want to scale a company, whether you want to invest in tech, or you just want to have a really great career in the digital age. Having a group of developers who are going to help you understand concepts and who are essentially going to help you understand some things, maybe even help you build something.

or help you screen developers that you're hiring, that's a really good asset. So I'm a non-technical founder and I make it a point to have really strong developer connections and really good developer friendships. and developers love telling you about what they're working and on what they're excited about. So ask them about DeepSeek and this will help you make some developer friends. Okay, number four, DeepSeek, as I have mentioned several times, is much cheaper.

So that basically means that if you are an early stage founder, a solo founder, or if you're working in a corporate and you're just not shaking the CFO down for an innovation budget, well, you don't really need one anymore. So this is really good. You can start experimenting with AI solutions basically without the financial risk. So you can use DeepSeek for data analysis, or if you have basically a bunch of reports, if you have a lot of data.

or maybe have a super long document that you need to get through. Or, you know, I recently had to analyze the reports of 60 of my students from my Tech Fun on Technical Founders course. And I wanted to spot patterns, you know, what did people find interesting? What did people find surprising? So I literally, put all of the reports into all three models to test how they compare. And to be honest, DeepSeq was really, really good. So you can do this kind of thing without worrying about the cost now.

Sophia Matveeva (09:34.976)
Also as a startup, that means that you can compete with larger companies because you don't have to worry about the budget. So what does this mean specifically? For example, if you have an idea for a prototype, if you have an idea for an app rather, whether you're in a corporate or whether you are just doing this for yourself, you previously used to have to work with professional designers or learn to use Figma, which is a professional design software. You can learn to use it.

It's just a bit fiddly. But now you can use, you can do some prototyping, actually quite a lot of prototyping with ChatGPT, with Claude and with DeepSeq. But because DeepSeq is basically so much cheaper, you can get your prototyping done much, much faster. And then you can analyze results without worrying essentially that you're going to run up your costs. And I've done a full episode on how to use AI to build

prototypes, so check that out. Just go back and actually we'll just post it in the links below. So what's another thing that you can do? Since DeepSeek is open source, you can hire developers to start building on top of that. And I don't want you to think that hiring developers to build on top of DeepSeek is going to be super hard and super expensive. You can basically hire mid-level developers.

who know how to use Python, you don't need super expensive maths PhD. So this is why DeepSeek lowers the cost of innovation and it lowers the cost of AI innovation. And for non-technical founders, this is really important because if you're non-technical, the likelihood of you like building an AI model by yourself is pretty much nil, right? So not only is DeepSeek just cheap in itself,

cheap or almost free basically in itself. Getting people to build on top of that, to build on top of it isn't super expensive. Like you don't have to compete with OpenAI for talent because if you did, well, you'll probably lose, right? Unless you have their massive budget. So actually it's cheaper in two ways. It's cheaper because you don't need to pay the same subscription, but it's also cheaper because you can get mid-level developers with

Sophia Matveeva (11:56.622)
Python experience and some AI experience, and there are actually quite a few of those people around now, to help you build a custom solution using DeepSeq's code. I hope this makes sense. So what can you start doing? There are some things that you can do. You can basically build a custom chatbot. Maybe you already have a website and people are getting confused on that website. Or maybe if you're selling something really expensive, then you can build a custom chatbox, a chatbot. You can automate some of your workflows.

You can build a predictive analytics tool. So for example, if you're working in marketing or you are thinking of creating marketing solutions and you want to basically create a tool to figure out which ones of your customers are going to be the highest value customers, you can use DeepSeek with a mid-level developer to basically create that custom tool for you. And to keep costs low, you can hire freelancers and also go offshore. You know, you are not, as I said,

For this, you don't need to compete with OpenAI for Silicon Valley salaries because they are ridiculous. So you can actually get really, good people from Eastern Europe, from Nigeria, from the Indian subcontinent to help you create this. And another thing that I recommend that you do as a non-technical innovator is that you get involved in the open source community talking about what to build on DeepSeek.

And basically the open source community is a bunch of developers who are talking to each other online, usually on Reddit, usually on Reddit and on GitHub basically. Those are the two places where they hang out about what they can do and about how they are using DeepSeek in order to build new stuff on top of it. So if you're not a developer,

Obviously reading all of these developer forums is going to be incomprehensible. So I don't suggest that you just sit there and just read all this stuff because you're not going to understand all of it. But you will begin to understand some of it. So do go and have a look, see what people are talking about a bit. Don't spend all your time on that. But even if you don't understand all of it and you understand like 20%, that's still pretty good. Also.

Sophia Matveeva (14:16.79)
The fact that you know that these communities exist is also really important because what happens is when there is some sort of new tool being invented and developers start using this new tool to create stuff, what often happens is that people get stuck creating something. So I don't know, they have this new tool, they're building some code and then they don't know how to use the code to basically create a new feature.

So this happens quite often. I'm sure it's going to happen with DeepSeq. Basically it happens with everything. What developers then do is they go to a developer forum where other people use this tool and they say, I'm trying to build an analytics feature. I'm trying to build a photo editing feature, whatever. I'm trying to build this feature using this code. I'm running into this problem. Has anybody solved this problem before? What do you recommend I do?

I want you as a non-technical innovator to know that these communities exist because when you're speaking to developers, I want you to find out are they in these communities? What communities are they in? What are they learning about? And where do they think the discussion is going? What are developers generally saying? Because being a developer, it's a big part of being a developer to basically to understand how to use new tools and to understand how other people are using new tools.

Get involved in the open source community, know roughly what they're talking about, know roughly where these communities are. You know, if you're feeling super, like, I don't know, you're feeling super vocal, you can even start speaking in them. I never do it, but I do like to go and have a look what's in there. Okay. So the key points for you are that DeepSeq's source code is open source. As we have mentioned, it is also low cost.

So basically, if you have been thinking about using AI, but you just haven't done it, and you've been basically making up all sorts of excuses for yourself, those excuses are just no longer relevant. I have given you an example of how you can just get started, literally get some recipes for what to have for lunch from what you have in your kitchen. And once you get started, once you get going, then you can move from your kitchen to what stuff that you actually need to solve for your business.

Sophia Matveeva (16:43.246)
So you have no excuse not to do it anymore. Secondly, you can now actually create new custom things, new custom chatbots or new custom AI for your project, which is going to be especially relevant for basically regional innovation. So what I have found, so I've been working in different parts of the world. So some of my audience is in the US, some of my audience is in the GCC and my students are in both.

And what I have found is that when really large American companies are being used, know, when large LLMs are being used in places basically outside of the Anglo-Saxon world, that sometimes their results are not that great. Not, basically not that maybe culturally sensitive and not well adjusted to a different region. And, you know, it's annoying if you're in this different region.

But it's also pretty normal because at the end of the day, these models are still super new. so, you know, they are, trained by a specific set of people on a specific set of data. So of course they are more geared toward that particular American Silicon Valley type of market. But what's good about having this deep seek model and being able to build on top of that is that if you are outside the U S say you are in Singapore, say you are in Dubai, say you're in Saudi Arabia.

So you're from Nigeria anyway, you get me. Then you can actually build solutions that are relevant to your culture and that are relevant to your language, for example. And this I think is going to have a really, really interesting impact on global innovation. start using the stuff, start talking to developers about using it and start creating things that are relevant for you and for your

target market. So I'm really excited about what you're going to do with this. And thank you very much for being with me on this episode. Thank you very much for watching. I really, really appreciate the fact that we are on this journey of discovery together and I'm really, really curious about where it's going to go. So my dear smart person, if you have found this lesson useful, you know what I'm going to ask you to do? Use your hands to either press like.

Sophia Matveeva (19:05.002)
on YouTube or leave the show a rating and a review if you are listening on Apple and on Spotify. That kind of help that basically really helps me and my team put this high quality free learning content together for you. All right. On that note, thank you very much for being with me. Have a wonderful day and I shall be back with you this time next week. Ciao.