LLMs, OpenAI Dev Day, and the Existential Crisis for Machine Learning Engineering
Jeremy Howard (Fast.ai), Shreya Shankar (UC Berkeley), and Hamel Husain (Parlance Labs) join Hugo Bowne-Anderson in a Vanishing Gradients livestream to talk about how LLMs are changing the worlds of data science, machine learning, and machine learning engineering, including (but not limited to!)
* How they shift the nature of the work we do in DS and ML,
* How they change the tools we use,
* The ways in which they could displace the role of traditional ML (e.g. will we stop using xgboost any time soon?),
* How to navigate all the new tools and techniques,
* The trade-offs between open and closed models,
* Reactions to the recent Open Developer Day and the increasing existential crisis for ML.
As a data scientist or ML engineer,
* What are you supposed to do as an ML engineer now?
* What kind of skills do you invest in?
* Should you really even focus on training models?
Let's find out!
00:00 Prelude
01:40 The panel begins
03:44 Introducing our guests Shreya, Hamel, and Jeremy!
05:37 A bit more about our guests and why they're so darn excited about LLMs
12:04 Is Jeremy Howard actually excited about LLMs? Plus monopolies, feedback loops, and regulatory capture
19:02 How do our technical roles change with the advent of LLMs?
27:00 ChatGPT as a transformational product experience
32:30 The importance of transfer learning and the end of fine-tuning
36:10 The crisis for ML engineering and the rise of the AI engineer
42:01 AI engineering and low code no code vs hacker skills and lots of code
46:00 Do you need to dive into Typescript and other languages these days and can ChatGPT help?
50:33 Have fundamental software engineering and CS skills became more important?
56:16 The impact of OpenAI developer day
01:06:45 The future success of open source models?
01:12:28 AI Safety, AI responsibility, and regulatory capture
01:17:23 How to choose between OSS models and vendor-based APIs
01:24:12 Wait, what does open source even mean when it comes to LLMs?
01:26:08 Notebooks vs LLMs -- friends or foes?
01:29:08 Our guests advice on what to do with LLMs next!
Jeremy Howard (Fast.ai), Shreya Shankar (UC Berkeley), and Hamel Husain (Parlance Labs) join Hugo Bowne-Anderson in a Vanishing Gradients livestream to talk about how LLMs are changing the worlds of data science, machine learning, and machine learning engineering, including (but not limited to!)
* How they shift the nature of the work we do in DS and ML,
* How they change the tools we use,
* The ways in which they could displace the role of traditional ML (e.g. will we stop using xgboost any time soon?),
* How to navigate all the new tools and techniques,
* The trade-offs between open and closed models,
* Reactions to the recent Open Developer Day and the increasing existential crisis for ML.
As a data scientist or ML engineer,
* What are you supposed to do as an ML engineer now?
* What kind of skills do you invest in?
* Should you really even focus on training models?
Let’s find out!
00:00 Prelude
01:40 The panel begins
03:44 Introducing our guests Shreya, Hamel, and Jeremy!
05:37 A bit more about our guests and why they’re so darn excited about LLMs
12:04 Is Jeremy Howard actually excited about LLMs? Plus monopolies, feedback loops, and regulatory capture
19:02 How do our technical roles change with the advent of LLMs?
27:00 ChatGPT as a transformational product experience
32:30 The importance of transfer learning and the end of fine-tuning
36:10 The crisis for ML engineering and the rise of the AI engineer
42:01 AI engineering and low code no code vs hacker skills and lots of code
46:00 Do you need to dive into Typescript and other languages these days and can ChatGPT help?
50:33 Have fundamental software engineering and CS skills became more important?
56:16 The impact of OpenAI developer day
01:06:45 The future success of open source models?
01:12:28 AI Safety, AI responsibility, and regulatory capture
01:17:23 How to choose between OSS models and vendor-based APIs
01:24:12 Wait, what does open source even mean when it comes to LLMs?
01:26:08 Notebooks vs LLMs — friends or foes?
01:29:08 Our guests advice on what to do with LLMs next!
