scripod.com
Why Dwarkesh is Wrong about Computer Use + How OpenAI shipped its Jev competitor in 1 Week

Overview

Shownote

Highlights

Transcript

Chapters

Pins

Why Dwarkesh is Wrong about Computer Use + How OpenAI shipped its Jev competitor in 1 Week

Latent Space: The AI Engineer Podcast

3 DAYS AGO
Latent Space: The AI Engineer Podcast

Latent Space: The AI Engineer Podcast

3 DAYS AGO
OverviewShownoteHighlightsTranscriptChaptersPins

Shownote

Three months ago Dwarkesh, who has been posting incredible blogs and episodes about RL, posted a framing question for his video essay on RLVR which upset a lot of Computer Use folks: We are no strangers to learning in public and are no strangers to the st...

Highlights

OpenAI’s DevDay announcements frame a rapidly evolving agent stack: persistent cloud computers, faster and more capable Computer Use, responsive tool-calling APIs, and infrastructure for long-running agents.
03:01
Every Dot gets its own cloud computer
05:23
Computer use is 180 degrees different now
10:00
Computer use can see the whole application
12:32
Computer use is already faster than average humans
15:36
Agents are now trusted with real-world payments
18:16
The agent becomes its own QA
20:18
Models no longer have to pause while tools run
24:26
Jev inspired OpenAI’s Decisions API sprint
27:40
The Decisions API starts with Luna, not a new model
30:49
The Decisions API is already classifying support tickets
33:34
A 12-hour cache guarantee is coming
35:47
Compaction is built into the Agents API harness
38:39
Building an AI-native cloud

Chapters

OpenAI DevDay: Dots, GPT-6.1, Agents API, and Decisions API
00:00
Dots and Personal Cloud Computers
02:52
Why Computer Use Is “180 Degrees Different”
04:59
From Sky to Self-Debugging Computer Use Agents
06:04
How Computer Use Sees and Operates Software
09:24
From Faster Than Humans to Superhuman Computer Use
12:09
Agents API: Trust, Permissions, and Safety
16:03
Computer Use for Coding, Testing, and QA
17:31
GPT-6 APIs, Async Tool Calling, and UltraFast Inference
19:14
The Rapid Story Behind Decisions API
23:21
What Decisions API Is and How It Works
25:32
What OpenAI Is Building With the New APIs
30:24
Prompt Caching, Pre-Warming, and API Performance
32:23
Context Compaction for Long-Running Agents
35:20
Memory, Higher-Level APIs, and the AI Cloud
37:13

Transcript

Speaker 4: Okay, we're very excited to be here. Today is OpenAI Dev Day. Speaker 3: We're the first podcast after the livestream. Speaker 4: We have Ari here, who leads the product and engineering team for computer use agents. Before we kick in and dive ...