@swyx: someone just told me about this take* on CUA this is one of those gell mann moments for me lol. i've been watching comp…
Summary
A tech influencer (@swyx) shares his perspective on the rapid progress of Computer Use Agents (CUA), citing historical milestones and current developments with GPT, Anthropic, and Adept, while warning that underestimating CUA capabilities is a dangerous error for AI decision-makers.
View Cached Full Text
Cached at: 07/16/26, 06:17 PM
someone just told me about this take* on CUA
this is one of those gell mann moments for me lol. i’ve been watching computer use since World of Bits (Shi, fan, karpathy, hernandez & liang 2017). we were the first technical pod to interview @jluan about Adept’s work three years ago, we were there in the @AnthropicAI building when they first launched Computer Use 2 years ago, I fanboyed over Claude Cowork in our @felixrieseberg pod 3 months ago, and we ran our first full computer use track at @aidotengineer ft. @DhruvBatra_ @proceduralia @francedot 3 weeks ago.
GPT 5.6 + Superapp is even better at CUA than everything i just mentioned. excited for our @AriX podcast to discuss the @skybysoftware story and Codex progress.
if you actually use these things as intensely as we do, CUA is progressing so, so incredibly fast. i have asked my nontechnical team to CUA as much as possible, all their knowledge work with signing up for random payment and invoicing portals and speaker and sponsor and attendee and vendor and union data requests. if you found yourself nodding along to this take below, you are so not up to date that you don’t know what you don’t know, and underestimating capabilities is quite a dangerous category error if you are doing any ai decisionmaking.
*i admire dwarkesh alot; only criticizing one single take, not the message nor the overall enterprise, screenshot only to share
follow this track playlist
Similar Articles
Computer-Using Agent
OpenAI introduced the Computer-Using Agent (CUA), a model combining GPT-4o's vision with reinforcement learning to interact with GUIs like a human, powering the new Operator agent. CUA sets new state-of-the-art benchmarks including 38.1% on OSWorld and 58.1% on WebArena, and is available as a research preview for ChatGPT Pro users in the US.
ToolCUA: Towards Optimal GUI-Tool Path Orchestration for Computer Use Agents
ToolCUA is a new agent framework that optimizes GUI-tool path selection for computer use agents through staged training and reinforcement learning. It achieves state-of-the-art performance on OSWorld-MCP by effectively interleaving GUI actions and high-level tool calls.
Operator System Card
OpenAI released the Operator System Card detailing safety evaluations for its Computer-Using Agent (CUA) model, which combines GPT-4o's vision capabilities with reinforcement learning to interact with GUIs and perform web-based tasks on users' behalf. The card outlines risk areas including prompt injections, harmful tasks, and model mistakes, along with multi-layered mitigations based on OpenAI's Preparedness Framework.
@CobusGreylingZA: https://x.com/CobusGreylingZA/status/2066593705906012188
A detailed thread arguing that true universal AI agents must build their own tools and explore environments dynamically, rather than relying on pre-configured integrations like MCP. It positions the terminal/CLI as the universal integration layer and references supporting research from OSExpert and NVIDIA.
Can Agents Use a Computer Yet? We've Got the Data (17 minute read)
a16z examines progress in computer-using AI agents, citing benchmark improvements that now surpass human-level performance on OSWorld-Verified and noting a shift from raw capability to reliable production deployment in enterprise workflows.