Watch our exclusive interview with Benny Yufei Chen, Co-Founder of Fireworks AI.

Fireworks raised ~$327M, most recently at a $4B valuation to change the AI space, enabling the enterprises to own the model themselves.

We started by exploring Fireworks AI's mission of enabling companies to access, fine-tune, and deploy #AI models through a simple platform - and the significant #cost savings potential.

We then shifted to the #developer perspective, exploring how teams should choose a base model for Reinforcement Fine-Tuning (#RFT) and the impact considering the customization thereafter.

As many organizations still rely on closed-source models, we dove into the advantages of #OpenSource alternatives and the key tradeoffs involved.

We discussed the launch of #FireFunctionV2 and how it has changed the game for function-calling models and AI applications.

Other interesting topics we covered:

-  Can open-source models outperform frontier models?

- The tradeoff between cost, latency, and model quality and does it depend on the vertical?

- Advice for Founders

And guess what's is Benny's favorite #movie is? 🛸🌌🪐

Check our new website:

https://www.boardroomclub.net

Register for our newsletter

https://boardroomclub.substack.com

Interviewers: Yossi Garmazi & Ellie Bleiberg

#ai #tech #agent #RFT #OpenSource

Chapters:

00:00 Why Open Source?

00:34 Intro

01:14 From Meta to Fireworks

03:13 What is RFT?

04:39 Raising Money in the AI era

06:54 Fireworks Customers

08:48 Developers & RFT

11:02 Why move from Closed-Source to Open Source?

13:10 FireFunction V2

15:19 Open Source to outperform Frontier models?

16:33 Lessons from Meta Vs the new AI world

19:17 Tradeoff between Cost, Latency and Model

20:26 Lessons from the AI World

23:28 Maintaining rigorous valuation & Startup speed

25:09 Principles by Ray Dalio

26:48 Advice for Product Leaders

28:29 Mentors

29:11 Hobbies

30:02 Coffee Question

30:44 Favorite Movie