In an era where artificial intelligence is rapidly transforming industries and reshaping the future of work, few voices are as uniquely positioned to offer insight as Misha Leybovich. With a background spanning from aerospace engineering to AI product development at Google Labs, Leybovich now leads 8P Strategy, a specialised AI consultancy and product factory.
In this exclusive interview with Stewart Tinson, project director of AI-360, Leybovich shares his perspective on the current AI landscape, its potential applications across various sectors, and the balance between innovation and practicality. As we stand on the brink of an AI-driven revolution, Leybovich's insights offer a glimpse into how this technology might solve old problems in new ways and reshape industries from the ground up.
Tinson: Can you give us some background on what you focus on?
Leybovich: I focus on trying to apply AI to solve old problems in new ways. AI is our greatest tool so far, and while there's a lot of hype out there, there's also a lot of genuinely new capability that we couldn't do before. I'm trying to figure out how to apply AI in as many niches as possible.
I currently have a company called 8P Strategy, where I work on this. Prior to this, I was at Google Labs building AI products around the time of the ChatGPT moment when AI started to feel more human and do things you would have previously needed to call someone for help with.
Tinson: What areas of building with AI interest you enough to go back to your entrepreneurial side?
Leybovich: It's important to think about the new AI capabilities we have now that weren't widely available before, such as:
Understanding deep insights in data: You can now upload a CSV file to ChatGPT and get analysis ideas and visuals generation, lowering the barrier for who can participate.
Generation: AI can generate text, images, music, video, code, audio, and even solutions like generative design. It lowers the barrier by allowing people to describe their output goals rather than needing all the skills to use complex tools themselves.
Sensing: AI can now watch hours of video, extract interesting things from images, and understand audio tone, massively scaling what humans could do before.
These capabilities lower barriers and allow AI to come up with solutions given constraints, leading to personalisation and customisation of everything digital, or tasks that can be digitised. As the cost of production comes to zero, we'll see more and more niche things catering to individual desires and professional needs.
Tinson: In terms of different job roles, functions, or industries, how do you even begin the process of identifying opportunities for AI?
Leybovich: It's a challenge because there are tens of thousands of industries and jobs, and millions of tasks. I didn't want to be limited by my own personal professional experience, which is a small fraction of all the areas where work is needed. However, with AI, we can now begin to ask the right questions and break down large questions into manageable chunks through step-by-step reasoning and orchestration.
I'm looking at different industries by size, roles by popularity, market caps, and more, trying to strike a balance between where there's money and where there's less competition. AI also allows you to take multiple shots on goal because the cost of production is going down.
One of my rules of thumb is that I basically don't know anything about anything, and neither does anybody else. The only way we know is by trying and iterating based on real market response. So I'm looking everywhere because it’s now possible, even though most things will be dead ends. I'm drawn to what's new that no one could do before.
Tinson: How do you balance customisation with scalability when applying AI to specific job roles or industries?
Leybovich: As a builder, the challenge is ensuring that the AI is providing a genuinely helpful answer and using the right level of creativity or the right model. It's still more of an art than a science right now.
Another challenge is that things are becoming less defensible in terms of raw technology, so we need to anticipate competitors showing up more easily in the coming years. Switching costs are also becoming incredibly low due to AI making it easier to move data and reducing interface lock-in.
In terms of the approach to scalability versus customisation, certain products that are super specific to existing industry workflows will likely still be in the domain of companies that have built products for that industry to date, with AI being built into them.
However, there's a class of products that require small changes to be applicable from industry to industry, such as tools that read PDFs and provide insights. These may fit better into existing workflows if they have useful documents for that space built in and know how to ask the right questions.
It's unclear whether we'll have swarms of super specific purpose-driven products that can communicate with each other, or mega products where everything is built into one interface. We might be in an unbundling phase where AI can do more specific things at a lower cost, giving you exactly the tool you need for a specific task. However, fragmentation of tools could be an issue, so there may need to be systems that allow these things to work together, like a toolbox or toolbench to organise purpose-built tools.
Tinson: Are some industries or job roles better suited to adopting AI, and is there a concern for potential job displacement?
Leybovich: Most people are not yet thinking deeply about or using AI in their daily lives, even in tech circles. We're still in an early part of the adoption curve, and most people haven't developed an intuition for AI yet. Incidentally, this is an area I’m also working on with weekly examples of "magical new capabilities" to inspire personal AI use at HowIUse.AI.
I don't think most jobs are fully at risk yet, but tasks are increasingly being addressed by AI. This has always been the case with technological advancements: take older examples like the automation of butter churning. Tasks that AI is good at – generation, analysis, sensing, and increasingly agentic things – will obviously be addressed more by AI as people and companies seek to get ahead.
If a significant percentage of a job’s daily tasks can be done by AI, the natural implication is that fewer people are needed in that job to do the same amount of those tasks. This doesn't necessarily mean widespread job loss, but roles will evolve, and new tasks will emerge. Some roles, like transcription, may be fully enveloped by AI if that's the sole task.
The nature of roles will change as technology advances, just like there are many fewer blacksmiths now than in the past. There is often pain during transitional moments, which is where societal systems come into play to deal with technological disruption. In the long run, humanity will figure out how humans plus AI is better than just humans or just AI alone.
Tinson: Do you see opportunities for cross-industry knowledge transfer between niche AI applications?
Leybovich: Absolutely. AI is really good at finding plausible solutions in whatever space you direct it towards. Just as recipes can be generated by combining characteristics of different cuisines, AI can be good at abstracting the nature of problems and apply solutions from one space to another.
Intractable problems have often been solved by generalists who bring perspectives from other fields. AI can simulate this by applying approaches from one industry to fundamentally different industries that share the same shape of root problem. While many attempts may be nonsense, some will uncover solutions that would have taken a uniquely experienced individual to discover.
Human creativity is about finding particular solutions within a massive space of possibilities. AI can help us explore that space faster and better, even though the cost of computation and analysis is still not zero. It takes human taste and guidance to navigate the creative potential space, but AI allows us to zoom around it in a much faster racecar.
This combination of human creativity, judgement, and taste, with AI's ability to process more information, will lead to the development of new genres and the solving of problems that would have required a super special person and a super random set of circumstances to come up with. I think that the upside of this will far outweigh the downside overall.
Tinson: What is the future of AI development in terms of specialisation and fragmentation of software applications?
Leybovich: The future is about customisation. Everyone will get exactly what they need. Right now, software is expensive to develop, so it meets the median needs of an industry. We'll progressively get to software that meets the needs of a sub-industry, then a company, then an individual, and finally, the specific task someone is doing right now.
This is wonderful because it means everyone will have tools that allow them to get things done faster and better, enabling humanity to produce more cool stuff that makes all our lives better. There will be downsides to the process, but the upsides are magical, and I'm excited to help build that future along with everybody else.
About Misha Leybovich
Misha Leybovich is a versatile entrepreneur and technology leader with a diverse background spanning artificial intelligence, space technology, social media, and management consulting. He holds dual master's degrees from MIT in Aerospace Engineering and Technology & Policy, as well as a B.S. in Engineering Physics from UC Berkeley.
Currently, Leybovich is involved in multiple AI-focused ventures. He's the founder of 8P Strategy, which specialises in "niche AI transformations" across various industries. He's also the CEO of Partylab.ai, creating AI-driven "unforgettable human experiences" as IRL social events, and a writer for HowIUse.AI, sharing insights on "magical new capabilities" and practical AI applications.
His professional experience includes roles at major tech companies like Google, where he worked on AI products and marketing technologies. At SpaceX, he served as Director of Starlink Sales, contributing to the company's satellite internet initiative.
Leybovich has founded several companies, including Bigtent Creative, which used innovative tech for political engagement, and Meograph, Inc., where he led the development of social AR experiences and secured multiple patents.
Earlier in his career, he worked as a management consultant at McKinsey & Company, focusing on technology and public sector clients. He has also conducted research at prestigious institutions like Cambridge University and MIT, contributing to publications on industrial policy and commercial spaceflight.
Leybovich's diverse skill set encompasses leadership, entrepreneurship, product development, and strategic thinking, with a consistent focus on leveraging cutting-edge technologies to create impactful solutions across various sectors.