Podcast Guest: David Yakobovitch
In this episode, we spoke with David Yakobovitch about investing in AI startups.
David is a visionary leader in the AI and data economy, currently serving as General Partner of DataPower Ventures, a VC fund focused on Applied AI, Inference, and DeepTech. With over a decade of experience in technical enablement and data science, David has held key roles at tech giants like Google, where he led data product development for Google Ads and served as an AI Policy Ambassador. He is the host of the acclaimed HumAIn Podcast and a frequent keynote speaker at Carnegie Mellon and Columbia University.
David is also the founder of Tech Power Network, a community of peer-collaboration events in NYC & Silicon Valley, where innovation meets opportunity for founders, executives, developers, operators, and investors. He is the co-organizer of AI Realized Summit, The Enterprise AI Deployment Conference.
We covered topics including -
Why it’s important to learn the foundations and first principles on any subject
What business decisions he supported with data science at multiple startups and big tech companies
His career transition into venture capital
An overview of AI tech including AI Infrastructure, Applications, Agents and Inference
David’s advice for AI startups seeking investment
What DeepSeek’s entry into the AI market taught us
Why community building is so powerful
Highlights from the Episode
Using AI-powered tools
AI tools are being used extensively today, even for complex tasks like coding. This trend will only increase. But, it is important for users to learn the foundations and first principles related to a topic to be able to excel at applying AI.
Differences in how data is applied or utilized in startups versus big tech
The speed of making decisions is slower at big tech because they require slow and deep thinking about the process and the impacted decisions before making any change. You need to make sure that every dependency is solved for before making even a small code change. That requires time. An example of how decisions regarding data utilization are made in big tech are the 6-page research briefs required from product leads at Amazon.
Startups have to move faster because they need to achieve product-market fit quickly in order to survive and thrive. They have a much more compressed cycle for applying data and moving forward.
Describing AI terminology in non-technical terms
AI Infrastructure - (e.g. OpenAI, Anthropic) a technical layer which allows anyone to use APIs to connect to them and create their own projects. It is the most mature space in the current AI environment.
AI Applications - products or services built on top of the infrastructure layer that allow users to directly access them (e.g. Grammarly’s AI-powered features). This area might have the most opportunity moving forward because there is a human-in-the-loop, which allows for course correction as required.
AI Agents - refer to AI-based automation of human tasks like taking notes during meetings or answering customer calls.
Other categories of AI are humanoid robots and inference (analytics and dashboards)
Thoughts on DeepSeek’s entry into the market some weeks ago
Leading and cutting edge products are being developed in many parts of the world outside the United States. It is only a matter of time before they surface as competitors.
Although early reports indicated that DeepSeek’s tool was built at a fraction of the cost of development of other LLMs, those reports turned out to be misleading. In fact, the costs were on par with the costs of developing other tools like those by OpenAI and Anthropic.
DeepSeek is open source, unlike some of the other tools, which gives users easy access to it.
How to pitch a startup idea to David’s VC fund, DataPower Ventures
Use the Startup Intake form on their website to submit your startup information.
When identifying potential candidates for investment, DataPower Ventures fund looks for 1. product velocity (i.e. consistently building and shipping features) and that 2. the startup is validating what it is building to align with industry needs.
He also recommends that startups share how they are de-risking their product and business in their pitches or investor materials.
Building community
You need an important mission and an authentic ‘why’ around your community that resonates with your members.
Successful community building requires a focus on how you envision the world and on sharing that vision. “Build the world that you are creating for”.
To build community, start small and keep going!
References
DataPower Ventures - multi-stage investor in tech and AI startups
HumAIn Podcast - explores AI for consumers through fireside conversations with industry thought leaders; hosted by David Yakobovitch
Tech Power Network - peer collaboration and networking group
AI Realized Summit - enterprise AI deployment events
GAI Insights - advisory firm dedicated to enterprise GenAI
General Assembly - tech training school
Techstars - tech startup accelerator program
Y Combinator - startup accelerator
Share this post