Investing in the public benefit OfficialAnthropic NewsOct 18Popular How AI giants will reshape global philanthropy WebAxios AIOct 19 Anthropic gives outside AI watchdog “employee-like” access WebArs Technica AIOct 18 What are self-replicating ‘AI worms’, and why is Anthropic worried about them? WebThe Guardian AIOct 18 Ben Horowitz: “The most dangerous thing in AI is Anthropic and Bill Gates” WebAll-In PodcastOct 18 Anthropic says its workforce has become less junior and is spending more time in meetings WebTechCrunch AIOct 17 Replit makes OpenAI its default over Anthropic, CEO says WebTechCrunch AIOct 17
Anthropic OfficialWebLearn & DoResources People Podcasts
All NewsTradeVideosVoicesCommunityResearchPodcasts
Subvolts take AI that improves itself could unlock scientific breakthroughs faster than humans alone. Jun 24, 2026
For you
Developers
Learn about advanced AI research concepts and how AI might drive its own progress.
Business
Understand potential future shifts in AI development and its impact on scientific innovation.
Hype note
3/5Notable

This interview features founders from a new AI company discussing a forward-looking concept for AI development.

AI + a16z
AI + a16z · Ep 102 Episode

Building Self-Accelerating AI with Mirendil

41 min Interview Hosts: a16z
0:0040:38

Audio streams from a16z. Your place is remembered on this device.

Go to the podcast episode page
Open AI + a16z in
Apple PodcastsOvercastPocket Casts
Follow the show
Apple PodcastsSpotifyOvercastPocket CastsAny Android app

From Subvolts

Summary and analysis by Subvolts editorial, AI-assisted.

Summary

The AI + a16z podcast features Mirendil cofounders Behnam Neyshabur and Harsh Mehta discussing their vision for self-accelerating AI. They explore how AI systems can contribute to their own development, accelerating scientific and technological progress beyond mere productivity tools. The conversation covers AI research, scaling laws, automated engineering, and scientific discovery.

Explained simply

This episode discusses AI that helps build better AI. Imagine a smart assistant that not only does tasks but also figures out how to become smarter itself. This could speed up discoveries in science and technology, benefiting everyone.

Why listen

This episode is for AI researchers and technologists interested in the future of AI development. Mirendil's founders, with backgrounds at Google and Anthropic, propose a paradigm shift from AI as a tool to AI as a driver of scientific advancement. Listeners will learn about the potential for AI to unlock breakthroughs in medicine and science by improving its own capabilities, a concept relevant to all frontier labs pushing AI progress.

On the mic
  • Hosta16z
  • GuestBehnam Neyshaburcofounder, Mirendil
  • GuestHarsh Mehtacofounder, Mirendil
About
AnthropicGoogleResearch
Key facts
  • Episode Title: Building Self-Accelerating AI with Mirendil
  • Show: AI + a16z
  • Length: 41 minutes
  • Guest Background: Mirendil founders previously led research at Google and Anthropic.
  • Core Question: What happens when AI systems can meaningfully contribute to their own development?
Show notes from a16z

Matt Bornstein speaks with Mirendil cofounders Behnam Neyshabur and Harsh Mehta about their vision for building self-accelerating AI.

After leading research efforts at Google and Anthropic, the founders started Mirendil around a simple question: what happens when AI systems can meaningfully contribute to their own development? Rather than focusing solely on AI as a tool for productivity, they argue that the most important application may be accelerating scientific and technological progress itself.

The conversation explores AI research, scaling laws, automated engineering, scientific discovery, and the challenges of building systems that can improve over time. They discuss the future of AI-assisted research, why they believe scientific progress remains bottlenecked by intelligence, and how more capable AI systems could help unlock advances across medicine, engineering, and the natural sciences.

 

Resources:

Follow Behnam Neyshabur on X: https://x.com/bneyshabur?lang=en

Follow Harsh Mehta on X: https://x.com/HarshMeh1a

Follow Matt Bornstein on X: https://x.com/BornsteinMatt

Check out everything a16z is doing with artificial intelligence here, including articles, projects, and more podcasts.

 


Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

Jargon decoded
self-accelerating AI
Behnam Neyshabur and Harsh Mehta
AI systems that can help improve their own capabilities over time.
scaling laws
Behnam Neyshabur and Harsh Mehta
Principles that describe how AI performance improves with more data or computing power.
automated engineering
Behnam Neyshabur and Harsh Mehta
Using AI to design, build, and optimize systems automatically.
Published
Show
AI + a16z · a16z
Length
41 min
Companies
Anthropic, Google
Summary by
Subvolts, using an AI model (how we work). Spotted a mistake? Tell us.
#Research

Best of AI + a16z

More on Anthropic this week

Questions people ask

What is Mirendil's core idea?
Mirendil focuses on building AI systems that can meaningfully contribute to their own development, aiming to accelerate scientific and technological progress.
What is the main application Mirendil envisions for AI?
Beyond productivity, they argue the most important application is accelerating scientific and technological progress itself.

Page generated Oct 19, 2026. Summaries are Subvolts' own; the audio, artwork and notes belong to a16z.