---
date: '2023-01-22T10:57:34'
hypothesis-meta:
  created: '2023-01-22T10:57:34.532045+00:00'
  document:
    title:
    - Who Owns the Generative AI Platform? | Andreessen Horowitz
  flagged: false
  group: __world__
  hidden: false
  id: k3jJlJpDEe2r9LtfV5j0MA
  links:
    html: https://hypothes.is/a/k3jJlJpDEe2r9LtfV5j0MA
    incontext: https://hyp.is/k3jJlJpDEe2r9LtfV5j0MA/a16z.com/2023/01/19/who-owns-the-generative-ai-platform/
    json: https://hypothes.is/api/annotations/k3jJlJpDEe2r9LtfV5j0MA
  permissions:
    admin:
    - acct:ravenscroftj@hypothes.is
    delete:
    - acct:ravenscroftj@hypothes.is
    read:
    - group:__world__
    update:
    - acct:ravenscroftj@hypothes.is
  tags:
  - ai
  - generative ai
  target:
  - selector:
    - endContainer: /div[1]/div[1]/main[1]/div[1]/div[1]/article[1]/main[1]/div[1]/div[1]/div[1]/div[1]/ul[2]/li[1]/span[2]
      endOffset: 238
      startContainer: /div[1]/div[1]/main[1]/div[1]/div[1]/article[1]/main[1]/div[1]/div[1]/div[1]/div[1]/ul[2]/li[1]/b[1]
      startOffset: 0
      type: RangeSelector
    - end: 15074
      start: 14604
      type: TextPositionSelector
    - exact: "Vertical integration (\u201Cmodel + app\u201D). Consuming AI models\
        \ as a service allows app developers to iterate quickly with a small team\
        \ and swap model providers as technology advances. On the flip side, some\
        \ devs argue that the product is the model, and that training from scratch\
        \ is the only way to create defensibility \u2014 i.e. by continually re-training\
        \ on proprietary product data. But it comes at the cost of much higher capital\
        \ requirements and a less nimble product team."
      prefix: 'tive AI app companies include:


        '
      suffix: '

        Building features vs. apps. Gen'
      type: TextQuoteSelector
    source: https://a16z.com/2023/01/19/who-owns-the-generative-ai-platform/
  text: There's definitely a middle ground of taking an open source model that is
    suitably mature and fine-tuning it for a specific use case. You could start without
    a moat and build one over time through collecting use data (similar to network
    effect)
  updated: '2023-01-22T10:57:34.532045+00:00'
  uri: https://a16z.com/2023/01/19/who-owns-the-generative-ai-platform/
  user: acct:ravenscroftj@hypothes.is
  user_info:
    display_name: James Ravenscroft
in-reply-to: https://a16z.com/2023/01/19/who-owns-the-generative-ai-platform/
tags:
- ai
- generative ai
- hypothesis
type: annotation
url: /annotations/2023/01/22/1674385054

---



 <blockquote>Vertical integration (“model + app”). Consuming AI models as a service allows app developers to iterate quickly with a small team and swap model providers as technology advances. On the flip side, some devs argue that the product is the model, and that training from scratch is the only way to create defensibility — i.e. by continually re-training on proprietary product data. But it comes at the cost of much higher capital requirements and a less nimble product team.</blockquote>There's definitely a middle ground of taking an open source model that is suitably mature and fine-tuning it for a specific use case. You could start without a moat and build one over time through collecting use data (similar to network effect)