Podcast Summary: Bloomberg Businessweek

Episode: BioRender CEO on Anthropic Deal, AI for Complex Imagery
Date: November 20, 2025
Hosts: Carol Massar, Tim Stenovec
Guest: Shizaoki (Founder and CEO of BioRender)

Overview

This episode centers on the intersection of artificial intelligence (AI) and scientific visualization, focusing on the capabilities and challenges of generating accurate scientific imagery. Shizaoki, founder and CEO of BioRender, joins the hosts to discuss her company’s innovative approach, their partnership with Anthropic (makers of Claude LLM), and the paramount importance of accuracy and openness in scientific illustrations.

Key Discussion Points & Insights

1. The Importance of Data Quality in AI (Start–03:12)

• "Garbage in, garbage out": Tim and Carol open by discussing this computer science principle as it relates to AI image generation, emphasizing how flawed input data leads to flawed outputs.

  • Notable Quote (Tim Stenovec, 01:42):

    "But we talk about it a lot with AI... if data that goes into something is flawed, the data that comes out is flawed. It's only as good as the data that goes in."

• AI's tendency for "hallucinations" is highlighted—where language models or image generators can provide incorrect or absurd outputs, sometimes intentionally for fun, but often problematic in serious settings like science.

  • (Carol and Tim, 01:53):

    "It's like when an LLM gives you an answer that you know is totally off base or gives you a picture image that's way off the mark as well."

2. BioRender’s Mission & the Anthropic Partnership (03:12–05:19)

• Shizaoki’s Perspective on AI in Science:

  • BioRender aims to make AI safer and more reliable for scientific communication by ensuring accurate visuals—a critical need since scientific misrepresentation can have severe real-world consequences.
  • The Anthropic CEO (Dario Amodei) is referenced, suggesting that AI could drastically speed up biological progress:
    • Notable Quote (Shizaoki, 03:32):

      "Dario Anthropic CEO, said that I could compress the next 50 to 100 years of biological progress into just 5 to 10, which means that the bottleneck really shifts from computation to human comprehension."

• Current Capabilities and Gaps:

  • AI excels at generating discoveries in text, but scientific visuals remain a challenge due to their need for precision.
  • Mistakes in scientific images (e.g., inaccurate protein folding, misdirection in metabolic charts) can have dangerous implications:
    • Notable Quote (Shizaoki, 04:36):

      "Or even like one arrow moving in the wrong direction in a metabolic pathway could mean that we're feeding the tumor instead of starving it."

3. BioRender’s Platform & Approach (05:19–07:08)

• What is BioRender?

  • Software that allows scientists to quickly create accurate, beautiful biological diagrams.
  • Positioned as a "Canva or Figma for biology," aimed at users unfamiliar with specialized scientific graphics tools.
  • Prior to BioRender, most scientists used general office tools like PowerPoint to create visuals, which was inefficient and often inaccurate.

• User Experience:

  • Scientists can drag and drop anatomically correct components, zoom in for detail, and effectively communicate complex interactions—far superior to the previous process involving basic shapes in PowerPoint.

• Founding Vision:

  • Shizaoki and her co-founders identified a significant communication gap in scientific fields and launched BioRender eight years ago to bridge it.

4. Customers, Revenue Model, and Industry Uptake (07:08–07:41)

• BioRender has become profitable early on, with rapid uptake in the scientific community.
• Primary users include researchers, leaders in pharmaceutical companies, biotech firms, academic institutions, and publishers.
  • (Shizaoki, 07:13):

    "Our customers are primarily researchers, including leaders, actually... in pharmaceutical companies, biotechs, academic institutions and even the publishing world."

5. Copyright, Sharing, and the Open Science Movement (07:41–09:24)

• Addressing the issue of copyright in scientific images:

  • There's a move towards openness, similar to collaborative platforms like GitHub and Wikipedia.
  • Scientists are eager to share their diagrams, shifting away from a "who published it first" mentality toward accelerating scientific understanding and error correction.
  • Notable Quote (Shizaoki, 08:32):

    "They just feel so compelled to change the narrative and change the status quo of how science is communicated today that they are willing to share the work that they do in BioRender back into the community."

• The BioRender community actively spots errors, updates are quickly made, and corrected visuals are shared—promoting collective scientific advancement rather than competition over intellectual property.

Notable Quotes & Memorable Moments

• On the Vital Role of Visuals in Science:
  • (Shizaoki, 04:19):

    "We've seen horrific diagrams of legs with extra bones or maybe less obvious mistakes, but more dangerous ones are like protein folding the wrong way. It can actually mean the difference between a functioning cell and Alzheimer's disease."

• On the "GitHub for Science Visuals" Vision:
  • (Shizaoki, 08:56):

    "So it's almost looking more like instead of the Canva or Figma, more of the GitHub or even Wikipedia because they share their work back into our library and either we spot errors or the community is quick to jump on those errors. We update those and then they get back uploaded into the repository."

Timestamps for Important Segments

• 01:40 – "Garbage in, garbage out" and AI hallucinations
• 03:12 – Introduction of Shizaoki & focus on scientific imaging in AI
• 04:02 – Partnership with Anthropic and the compression of scientific progress
• 05:19 – What BioRender is and its core use case
• 07:08 – Who BioRender’s customers are and its business model
• 07:41 – Copyright open sharing, and collaborative science
• 09:24 – Episode wraps with a forward-looking note on open scientific communication

Tone and Style

The conversation is approachable, lively, and focused on demystifying a highly technical area for a broad audience. Shizaoki is enthusiastic about open science and making advanced visualization accessible, while the hosts keep the discussion grounded with practical and ethical considerations.

Conclusion

This episode highlights AI's growing role in scientific visualization, the unique challenges involved, and the importance of accuracy in advancing research. BioRender, through its tools and open library, is changing the way researchers communicate, accelerating science, and fostering a spirit of openness and collaboration.