Workflow
Opus 4.6%
icon
Search documents
Speed is ALL the matters for AI
Matthew Berman· 2026-02-13 19:14
vindicated. That is how I feel. I have been a speed maxi when it comes to AI models for a very long time.But no, no. Everybody said quality is by far the most important. Well, no.It turns out speed is just as important. When you are giving agents the ability to go out and get realworld work done for you, and the output of these models aren't limited by how fast a human can read them, then speed is everything. And it feels so good to realize many other people are coming to the same conclusion now.In fact, An ...
Everything you need to know about ChatGPT 5.3 Codex
Matthew Berman· 2026-02-06 14:00
Open AAI and Anthropic are going head-to-head within minutes of each other. Opus 4.6% was dropped and now GPT 5.3% Codeex. They are both going so hard into a Gentic coding.That is where the industry is headed. That is where all of these frontier labs are investing their time. Long horizon task, agents, sub aents, agent teams.One of the biggest complaints about codec has been how slow it is. A lot of people say it is the best coding model out there, but it is so brutally slow as compared to Opus and other co ...