
Sign up to save your podcasts
Or


Maya, Anika, and Priya open this week's episode after spending two full hours trying to fix their new audio recorder, which made Priya miss her Solidcore class for the eighth time this month. Anything for the pod though!
First they break down reinforcement learning as a service. Starting with what RL is and how DeepSeek popularized it’s efficacy. Companies like Applied Compute are trying to do for RL what AWS did for compute, except most enterprises can barely collect their own data properly, so the ROI of RL is an open question. The girls dig into why some companies like Sierra and Decagon have a clear use case for RL, while for others they may prefer relying on natural model improvements from the lab.
Next, the Iran War and what the closure of the Strait of Hormuz is actually doing to the world. Crude oil has hit $113 a barrel. Fertilizer shortages are tanking corn yields in Zambia and Sri Lanka. There's a mid-April deadline that has analysts worried. In all the chaos none other than Putin has ended up on top. The hosts debate what may happen next along with everyone on Kalshi that’s been trading on related markets. You can trade too: https://kalshi.com/sign-up?referral=getthecheck.
Finally, if Reflection AI would get the pod’s check. Reflection AI is the American open source lab rumored to be raising at a $25 billion valuation from JP Morgan and positioning itself as the western answer to DeepSeek. Maya and Anika debate how important it is for there to be an American open source model. Reflection has the model internally, but they haven't released it to the public yet. The pod is excited to see what they release and predict it’s coming soon.
Now…a word from our sponsors :)
00:00 Intro02:52 RL explained03:40 Why RL became popular (shoutout DeepSeek)05:01 RLaaS06:36 Why DoorDash bought an RL company07:23 The ROI on RL14:17 RL players17:48 Cursor using RL on Composer in real time20:18 Metis acquisition22:05 Iran War23:39 Historical context on Iran25:39 Closure of the Straight of Hormuz27:28 When oil reserves will start to run out29:43 Impacts beyond oil33:54 Reflection AI funding34:45 Open vs closed AI models41:00 Why Nvidia wants to be on an open source model
By Anika, Maya, Priya5
2020 ratings
Maya, Anika, and Priya open this week's episode after spending two full hours trying to fix their new audio recorder, which made Priya miss her Solidcore class for the eighth time this month. Anything for the pod though!
First they break down reinforcement learning as a service. Starting with what RL is and how DeepSeek popularized it’s efficacy. Companies like Applied Compute are trying to do for RL what AWS did for compute, except most enterprises can barely collect their own data properly, so the ROI of RL is an open question. The girls dig into why some companies like Sierra and Decagon have a clear use case for RL, while for others they may prefer relying on natural model improvements from the lab.
Next, the Iran War and what the closure of the Strait of Hormuz is actually doing to the world. Crude oil has hit $113 a barrel. Fertilizer shortages are tanking corn yields in Zambia and Sri Lanka. There's a mid-April deadline that has analysts worried. In all the chaos none other than Putin has ended up on top. The hosts debate what may happen next along with everyone on Kalshi that’s been trading on related markets. You can trade too: https://kalshi.com/sign-up?referral=getthecheck.
Finally, if Reflection AI would get the pod’s check. Reflection AI is the American open source lab rumored to be raising at a $25 billion valuation from JP Morgan and positioning itself as the western answer to DeepSeek. Maya and Anika debate how important it is for there to be an American open source model. Reflection has the model internally, but they haven't released it to the public yet. The pod is excited to see what they release and predict it’s coming soon.
Now…a word from our sponsors :)
00:00 Intro02:52 RL explained03:40 Why RL became popular (shoutout DeepSeek)05:01 RLaaS06:36 Why DoorDash bought an RL company07:23 The ROI on RL14:17 RL players17:48 Cursor using RL on Composer in real time20:18 Metis acquisition22:05 Iran War23:39 Historical context on Iran25:39 Closure of the Straight of Hormuz27:28 When oil reserves will start to run out29:43 Impacts beyond oil33:54 Reflection AI funding34:45 Open vs closed AI models41:00 Why Nvidia wants to be on an open source model

6,089 Listeners

10,182 Listeners

5,530 Listeners

6,477 Listeners

15,950 Listeners

146 Listeners

491 Listeners