One of the biggest challenges for startup founders is figuring out how to pay themselves and their first hires. It’s more than a bit of a mystery for most first time founders – something I get a lot of questions about. Since payroll often eats up more than 75% of a startup’s operating expenses, getting compensation right can make or break a company’s runway.

That’s why I put together the Startup Compensation Guide—real, data-driven insights based on actual payroll records from hundreds of venture-backed startups. One of the more surprising takeaways? At the seed stage, CTOs often earn more than CEOs, reflecting the intense demand for technical talent. Founder salaries scale with funding rounds, starting around $132,000 at seed and rising to $218,000 by Series B.

For early employees, I found that compensation varies significantly based on role, experience, and location. Engineering hires are usually the first and most expensive additions to a startup team. In major tech hubs like San Francisco, salaries range from $75,000 for entry-level engineers to over $230,000 for top-tier talent, with slightly lower pay in other cities. Sales, product, and design roles follow a similar pattern, but cash compensation is only part of the equation—startups also need to budget for equity, benefits, payroll taxes, and one-time costs like equipment and software, which can add 25-35% to the total cost of hiring.

Equity remains a huge part of startup compensation, and my research showed that early employees typically receive between 0.5% and 4% ownership, with later hires getting progressively smaller grants. Most startups follow a standard four-year vesting schedule with a one-year cliff, ensuring long-term commitment while protecting the cap table. Balancing cash and equity is crucial—more experienced hires may require higher salaries, while early-stage startups need to preserve equity for future rounds. By structuring compensation thoughtfully, using 409A valuations, and leveraging cap table management tools, founders can build competitive pay packages that attract and retain top talent while keeping their startup financially healthy.

I recently shared some thoughts in The Information‘s article, “How Long Can OpenAI’s First Mover Advantage Last?” The piece explores the rapid adoption of OpenAI among startups and raises the question of how long the company can maintain its competitive edge in an evolving AI landscape.

At Kruze, we’ve seen first-hand how quickly OpenAI has been integrated into the startup ecosystem. By October 2023, 57% of our clients were paying for OpenAI services, a rate of adoption that’s unprecedented. Startups are using AI for everything from marketing copy and Excel modeling to data processing and even training their own custom AI models. Check out our data on how startups are using AI.

However, from what I’m observing, it’s unlikely that startups will remain reliant on a single AI model for long. Many are already switching between different models with surprising ease. I believe that, moving forward, startups will leverage multiple AI models based on three key factors: cost, effectiveness for specific applications, and speed.

In the end, it’s going to come down to which tool provides the best results for each unique task. Startups are constantly balancing efficiency with expenses, and I think we’ll see companies become increasingly agile in swapping between AI models to optimize for these variables. The AI landscape is rapidly evolving, and while OpenAI’s adoption has been nothing short of impressive, its lead is far from guaranteed.

For the full story, check out the article on The Information.

I’ve been having a lot of fun with AI. One thing that I think is great is that you can ‘get advice’ from real people, at least you can get the AI to give you advice pretending to be those people… if they were famous enough to have enough stuff on the internet for the LLM to train on.

Here is some advice from a jerky Steve Jobs:

Well, well, well, it seems like we’ve hit a bump in the road, haven’t we? Let’s put a dent in this universe, shall we? Sit down, grab your favorite drink, and listen. If you thought this was going to be easy, you should have stuck to Sudoku.

First, let’s get this straight, products don’t fail, visions do. What’s your vision? What? Don’t have one? Then, there’s your first problem. Without vision, you’re just another no-name trying to hustle in the Silicon Valley rat race. No wonder the product isn’t shining. Remember, a product isn’t an item, it’s a solution to a problem. Is your product solving any problem? Or did you just push out something because your coding fingers were itchy?

Let’s talk about the product now. What’s wrong with it? What’s right with it? Can’t tell me? How do you expect your customers to figure it out? A product’s quality isn’t measured in lines of code or the time it took to design its logo. It’s about how it makes life easier, better, more enjoyable. In short, it’s about adding value. Is your product doing that?

And, oh boy, don’t get me started on your team. Are they A-players or just some people who know how to use a computer? If you’re the smartest person in the room, you’re in the wrong room. If your team doesn’t challenge you, doesn’t push the boundaries, doesn’t make you lose your temper once in a while, then they’re not the ones who’ll build the next big thing. Fire them. Hire the ones who can make your product a star.

Now, before I finish my rant, let me tell you something. It’s okay to fail. Heck, I got kicked out of my own company. But, I came back and made it the first trillion-dollar one. It’s not about the fall, it’s about the comeback. So, stop whining about your failure and start plotting your comeback.

You don’t need to be the next Steve Jobs. Be the first you. But for crying out loud, make sure you’re a damn good one. Get back to work now.

I’ve written a few pieces over on Medium about how I’m getting into telemedicine. For example, this one is about how I hoped telemedicine would adapt to help out during the COVID crisis. I wrote about how I liked the fact that the new telemedicine startup players have the infrastructure for diagnosis and care delivery, could provide access to anyone with a data plan (even in rural areas), were driving adoption of telemedicine, would make it safer to see a doctor (since you didn’t have to go to a crowded doctor’s office), could help with triage and treat other conditions during the crisis.

So far, I’d say that this is panning out, with players like Hims offering a large number of doctor visits for a huge number of conditions that go beyond their hair loss treatments and get into primary care, skin care and more. Pretty exciting stuff for the future of telemedicine and treatment delivery! I hope Hims and Roman and the like continue to innovate.