AMA with Amir Ashkenazi founder of Airtop, Adap.tv, and Shopping.com

@gabe It's easy to feel pessimistic about the future of the job market. While some job types will inevitably be eliminated, I’m actually very optimistic about AI’s overall impact. AI is elevating every job, shifting our mindset from simply completing tasks to thinking in terms of workflows and processes. Take sales, for example—it's no longer just about writing the next email to a customer; it's about designing a playbook that consistently generates and closes deals. I'm also excited about how AI is lowering barriers for those who want to build new things. The gap between an idea and a fully built product, app, or technology is shrinking—and that's a great thing.

@gabe @amir_ashkenazi But do you think with AI/automation enabling more productivity per employee that companies will hire fewer people?

@gabe @steveb Yes, this is a massive disruption. Some jobs are being completely eliminated, while others—like software engineering—are becoming exponentially more efficient, with one person able to do the work of ten. We’re already seeing companies announce hiring slowdowns for software engineers. But I also believe AI will expand the economy, create new types of job, and lower the barriers to starting new companies. As with any revolution, it is important to position yourself on the right side of it...

@aaronoleary great question... Looking back, a few patterns stand out in how emerging technologies move from niche innovation to mass adoption. First, there’s always an early phase where the technology is powerful but not yet accessible—only experts or highly technical teams can leverage it. Then, a shift happens when someone builds the right abstraction layer or tooling that makes it usable by a broader audience, like chat on top of an LLM. That’s when adoption really takes off. Another key pattern is that new technology rarely replaces old systems overnight. Instead, it first finds footholds where it delivers 10x improvements over the status quo—whether that’s cost, speed, or capabilities. Over time, as more businesses adopt it and the ecosystem matures, it transitions from being a competitive advantage to a necessity. Think about how e-commerce evolved from standardized products like books to fashion, cars and everything else Finally, as a founder, I always prefer to ride the wave of massive disruption—but timing is everything. The best ideas often emerge before the necessary infrastructure is in place. For example, Waze launched before GPS was built into phones, and when we started Shopping.com in 1997 as a comparison shopping service, there were only two or three e-commerce sites to compare.

@rajiv_ayyangar Great question and one of the most difficult aspects of building a company. Here is my take... as entrepreneurs, we have a story we tell the world about the problem we solve and how we’ll build a great business around it. As we progress, we gather more data points that shape our understanding of the market and our story. When you stop believing your own story, it’s time to pivot! Finding the new path isn’t all that different from finding a startup idea in the first place—it’s about identifying a real problem, testing assumptions, and iterating until something clicks :-)

@davefaliskie tools like Replit and Lovable absolutely wow me. At least until they get stuck on something, and then I need to debug the code... :-)

@amir_ashkenazi Nice those are on my list to checkout, I definitely will now

Hey @mikekerzhner, so happy to be here! I love the product and community you build. That’s truly a billion-dollar question. I won’t try to predict exactly how AI will evolve, but I’ll share some core assumptions I believe in. First, AI is completely transforming web automation. The days of manually selecting elements in a DOM tree and dealing with brittle scripts that break when a page layout changes are numbered. Platforms like Airtop already make it possible to automate anything on the web using natural language. Second, the demand for web automation will only grow. We’ll each have dozens of AI agents and automations running in the background, handling tasks on our behalf. These agents will need reliable web access to gather information and take actions, meaning a much larger share of internet traffic will come from machines, not humans. As for competition, even slight differences in philosophy or development lead to different products optimized for different use cases. I believe you’ll choose Airtop for some, and OpenAI Operator or Claude’s tools for others. But, of course, I still have to prove it.

@amir_ashkenazi thanks, love the answer. As an engineer, I truly hope that 2025 is the last year I write a Playwright or Selenium test.

@mikekerzhner let's make February the last month you write a Playwright or Selenium test...

@steveb we're mostly focused on selling to businesses, so most of the use cases are about increasing revenue, finding new customers, and automating internal processes. But I also love seeing how individuals use it for their own benefit. One of my favorite use cases is searching for a job and applying automatically. If AI is taking your jobs, it might as well find new ones for us...