Bots now outnumber humans on the internet

3 trends in startups/tech/venture

📈 Bots now outnumber humans on the internet

• Before generative AI, bots such as Google’s crawler accounted for ~20% of internet traffic. As AI agents take the world by storm, bots now account for 58% of all internet traffic, marking the first time that humans are a minority online.

• This trend makes sense. When you—a human (hopefully!)—go on Google to decide which cruise to take, you might visit a handful of websites before making a decision. However, when you ask AI to do this research for you, AI sends bots to scour thousands of websites at once.

• The result? A new gold rush called “agentic commerce” where everything from researching options to clicking “buy” is handled by AI.

• Startups are starting to pounce on this trend, with new ventures popping up by the day to help companies not just become discoverable by AI (via a process called generative engine optimization or GEO) but to facilitate transactions with these AI customers — leading to new terms like “agentic payments,” “agent identity verification,” and “agentic security.”

🤖 AI is turning specialists into generalists

• With all the chatter about tools like Claude Code making software engineers obsolete, you might think that AI is primarily being used to turn a 10-day coding job into a 10-minute one. This is happening, but it doesn’t tell the full story.

• Researchers at OpenAI recently found that 44% of ChatGPT prompts at work involved people reaching beyond the scope of their day job, with marketers analyzing data, operators writing code, and managers moonlighting as researchers.

• Meanwhile, researchers at Perplexity and Harvard found that AI agents pushed people toward harder and more cross-field work rather than merely faster work.

• These studies all point in the same direction: interdisciplinary skills will become increasingly important.

💰 Concentration in venture capital grows

• I’ve written about VC’s increasing concentration of capital before here. Well, it’s still happening — and to an increasingly greater extent.

• In terms of venture funds: In the first half of 2026, two thirds of new venture funding went to just 6% of funds.

• In terms of venture deals: In the first half of 2026, “mega-deals” (funding rounds of at least $100M) accounted for 88% of all fundraising.

• The result? It’s now possible to become a unicorn by Series A. This is what the top 1% of startups are doing👇🏼

  

• The fact that so much money is rushing to so few places raises a question: what about the rest — the 99% of startups that aren’t inciting the most FOMO? My view is we’re in the golden age of overlooked opportunities, with smaller funds now having a big opportunity to pounce on deals with more reasonable valuations — and therefore generate more alpha in the end — through bets that go against the herd.

2 theses on what’s next

🤖 Reviewers will be the next to be automated

• So far, AI has been that A+ intern who does your every bidding without complaining or needing to go to the bathroom. A+ interns are great, but limited by one thing: their manager‘s availability to review their work and make sure it’s client ready.

• But what if the reviewer of the work could be automated too? We’d then unlock a whole new level of productivity because, for the first time, humans can be entirely removed from the equation, leaving robots to do everything from start to finish.

• I believe this next phase of automation is coming and we’ll start seeing the first examples within the coming year, if not even sooner.

🤔 The Google Doc-ification of the enterprise is coming

• AI agents running amok in a big company creates another problem: what do you do if something breaks? With humans, you’d know that it was one of your 200 software engineers — and specifically Steve who pushed his code last Tuesday at 4:00pm Eastern Time. But, what if you now have the equivalent of 200,000 AI agent “Steves” working 24/7? If something breaks overnight, you could be millions of lines of code behind by the time a human even logs on.

• The new need? The ability to look at every inch of your enterprise as you would a Google Doc, complete with the time stamp for every tiny change that was made and the ability to rewind — again like Google Docs.

• An estimated $30 billion is spent each year on the enterprise backup and disaster recovery market — but this market was built for human-paced changes. As the number of AI “Steves” explodes, I expect the market in the coming years to explode as well as more companies see the importance of having a constant and detailed log of nook and cranny of their enterprise.

1 tool I love

🎓 AI-first engineering

• Institutions like Microsoft, NASA, and Nvidia used to spend millions physically testing prototypes — like tweaking rocket wings at every angle to find the version that’s most aerodynamic.

• Rescale moves all of this physical testing into virtual simulations that are more cost effective and time efficient — enabling companies to deliver the best designs for rockets, planes, cars, and even drugs at a fraction of the cost.

• Learn more about what "AI-first engineering" actually looks like here.

More of my work 👇🏼

🎤 Decode videos | For top Berkeley and Stanford founders

• DECODE is the largest founder community co-hosted across UC Berkeley and Stanford. The DECODE annual conference focuses on helping founders in their earliest stages of starting a startup.

• I’ve had the honor of serving on the DECODE board since 2016. You can find videos from the latest conference I helped host here 👇🏼

🤖 Founders' prompt library

• My team and I have crowdsourced the largest AI prompt library optimized for founders – especially fractional founders — to help them build, sell and operate 10x faster and better.

👥 Pop up board

• I regularly host pop-up boards, a unique 45-minute session where founders ask their toughest strategic questions and get tailored advice from top builders and operators from Google, Microsoft, Meta and more.