AI SAFETY TWITTER UPDATES

← Front page

Nathan Lambert lectures on risks and access controls in open models

2 people · 6 posts · first post · latest

Nathan Lambert discussed how to manage risks and control development in the open model ecosystem during his Harvard lecture. He detailed model types, licensing, and the use of revenue gates and usage restrictions by model makers. Discussions following the lecture focused on the efficacy of distillation and the impact of local models.

All posts · 5, most important first

  1. @natolambert research ·
    My first open model lecture, covering all the many pieces that build the ecosystem, and the open questions/debates determining where it goes next. How does distillation fit in? How do we manage risks? How do we see access to models changing in the future? I'm excited to revisit these questions about every 6 months, to give a state of the union on the field. Chapters 00:00 Introduction 01:36 What is an open model? 08:47 Why should we care about open models? 18:52 How competitive are the best open models today? 25:28 How are open models built? The role of distillation 34:52 Are people actually using open models? 41:47 What comes next with open models? 47:32 Conclusions What questions did I not answer that you have?
    72 likes · 10 reposts
  2. @natolambert research ·
    I estimate a few months closer to the frontier. I discuss the arguments for and against distillation mattering in the lecture
    1 likes · 0 reposts
  3. @natolambert research ·
    Yup, local models surging got a shoutout, and the model makers do care, it’s incredible marketing for them
    1 likes · 0 reposts
  4. @natolambert research ·
    10 likes · 1 reposts
  5. @natolambert research ·
    Nathan's youtube arc, like & subscribe?
    3 likes · 0 reposts

Reposted by @boazbaraktcs.