Real Alpha
Try Pro free

All-In · 3 min · Jul 11, 2025

Human-labeling businesses may have a short runway

KRKeith Rabois

Keith Rabois argues that demand for human-labeled data may last only a few years as models learn to label for themselves.

In this episode

Full episode
  1. 6:35Food automation can cut kitchen labor below 10% of revenuePro pick 8 min
  2. 21:03Compute-first learning is overtaking human-engineered AIPro pick 8 min
  3. 46:22AI browsers may be a costly detour from better products8 min
  4. 30:04Human-labeling businesses may have a short runway3 min
  5. 1:17:33The ruling cleared workforce planning, not implementation3 min
  6. 1:03:33Super PACs can now fund campaign ground operations6 min
“the truth is i think there's a very short half-life on human-label data.”
KRKeith Raboisat 30:08 · All-In
30:08

Takeaways

  1. Human-labeling demand may last only one to three years.
  2. Models may learn to label data as well as people.
  3. Assess data vendors by durability, not just current revenue.

About this segment

Keith Rabois argues that demand for human-labeled data may last only a few years as models learn to label for themselves. For investors, current revenue may not mean a durable market.

KR

Keith Rabois

  • Chamath Palihapitiya, Host · Founder & CEO, Social Capital

    Led Facebook's growth team; his firm backed Slack and AI chipmaker Groq.

  • Jason Calacanis, Host · Founder & CEO, LAUNCH

    Early Uber angel; sold Weblogs Inc. to AOL; hosts This Week in Startups.

  • David Friedberg, Host · CEO, Ohalo

    Founded The Climate Corporation, sold to Monsanto for about $1.1B in 2013.

From “Grok 4 Wows, The Bitter Lesson, Third Party, AI Browsers, SCOTUS backs POTUS on RIFs”, Jul 11, 2025. Watch the full episode

Answers “Which jobs will AI replace, and which are safe?” See all 121 answers

Answers “What will AI look like in the next few years?” See all 155 answers

Transcript

Transcribed by machine (Whisper) from the episode’s audio, so a word may be off here and there.

  1. 30:10

    should we invest in companies like scale or mercor or any of these surge the truth is i think there's a very short half-life on human-label data. And so everybody who's investing in these companies, just looking at revenue traction, really didn't understand that there may be a year, two years, three years max when anybody uses human-label data for maybe anything. Because we hit the end of human knowledge or just the collection of it is 99% done.

  2. 30:43

    Or you train on it so well that you don't need to label anymore. Like the machines know how to label as good or better than a human. And so like we're seeing this in the self-driving space is labeling was huge, right? You would have a three-dimensional sort of scene that's created by video plus LIDAR, let's say, okay, I have to label all of these essentially what become boxes. like i've identified objects you're you're some of the players in the in the autonomous software space autonomous vehicle software space are no longer doing any labeling because the machines are doing it all just broadly it'll just be built into the chipset that this is a stop sign like it's like we know what a stop sign is we don't need the millionth time it's like those captchas like you're like find the stop sign or what's the traffic light and eventually the machines are

  3. 31:38

    just way better than humans at identifying these things for you to be very practical when you see a stop sign you don't have to identify that it's a stop sign you just see that every human when they encounter a stop sign 99.9 percent of the time they hit a break and they never act so nobody actually knows it's a stop sign it's just that hit a break when you see something that looks like this Yeah.

  4. 32:00

    it's just a vibe. It's a vibe. I would just say that that's like intuited knowledge versus like the expressly labeled human knowledge. The question for me is, if everybody was so reliant on human labeling initially, if you're an investor now, when you see these Grok 4 results, how do you make an investment decision that's not purely levered to just computation? So if you look at these results, does it mean that there's 300 to 1,000 basis points of lag between just letting the computer's vibe itself to the answer versus interjecting ourselves?

  5. 32:43

    If interjecting ourselves slows us down by 300 to 1,000 basis points per successive iteration, then over two or three iterations, you've totally lost.