Claude Code Replaced Cursor for Me — Here's Why
A 37-minute conversation between Riley Brown and Ras Mic (Michael Shimeles) on why the model provider building its own tools is changing the AI coding wars.
June 25th 2025A 20-minute breakdown of the $60B SpaceX–Cursor deal, a hands-on tour of the platform, and a one-prompt playbook for migrating your entire Codex or Claude skill library.
SpaceX bought Cursor to run a controlled experiment: Cursor's developer-trace corpus is the variable, Colossus compute is the lab, and a $60B exit was always the price of finding out whether that data closes the frontier coding model gap.
SpaceX exercised a pre-negotiated $60B option to buy Cursor after jointly training Composer 2.5 on Colossus compute. The real asset was Cursor's developer-trace corpus — the largest in the world — which xAI needed to close the gap on Codex and Claude Code. Cursor's messaging dropped 'for developers' entirely, signaling a pivot toward a general-purpose super app. The platform is already nearly feature-identical to Codex in layout and agent capabilities; its only missing pieces are native document rendering and computer use, both of which appear imminent. For existing Codex or Claude users, a single export prompt moves all your skills and memory into Cursor globally.
Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.
Create a free account →
SpaceX acquired Cursor; preview of acquisition analysis, platform tour, and skills migration.

Deal structure: $60B option or $10B experiment. SpaceX gets developer-trace data; Cursor gets Colossus compute and a guaranteed exit. The subsidy gap between Cursor and Codex/Claude Code explained. Cursor's messaging signals pivot away from developer-only positioning.

UI walkthrough comparing Cursor and Codex side-by-side. Model selection, in-app browser advantage, design mode demo. Vibe coding a personal site with Vercel deploy. Building a database-connected notes app with Convex. Things Cursor will add: Composer 3.0, document rendering, computer use, mobile app.

Single-prompt export from Codex creates a codex-import folder. Cursor ingests it globally. 73 skills imported with clean metadata. Live test of YouTube researcher skill and global memory recall. Works identically from Claude.
The SpaceX–Cursor deal is not a valuation story — it is a data-for-compute swap that restructures the competitive dynamics of the entire AI super-app category.
“Notice how Cursor isn't saying 'for developers' — just 'useful AI.'”
“You're gonna create software not just for you to use but for your AI agent to use as well.”
“With your $200 subscription, someone actually did the math — you're able to use $14,000 worth of compute if you absolutely max it out.”
See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.
SpaceX exercised a $60B option and walked away with the fastest-growing AI coding tool in the world — and the developer-trace corpus that might be the missing ingredient in the frontier model race. Riley Brown spent eight hours inside Cursor after the news dropped and came back with a verdict: it's two features away from being a genuine Codex and Claude Desktop competitor for everyone, not just developers.
Three platforms competing to become the single AI-powered workspace for all knowledge work and coding — agents, browser, documents, automation in one app.
SpaceX structured the deal so they couldn't lose — either they get a $60B company or they buy the world's best coding training data for $10B.
“I will be making many videos talking about super apps, how to become agent native going forward. So please hit the like button, subscribe button.”
Standard end-card verbal CTA, low pressure, no visual overlay shown in frames
00:02
00:17
00:38
00:54
01:09
01:32
01:43
01:56
02:11
02:27
02:42
02:57
03:13
03:28
03:44
04:04
04:15
04:30
04:46
05:01
05:17
05:32
05:48
06:03
06:19
06:34
06:44
07:03
07:24
07:39
07:52
08:01
08:18
08:38
08:49
09:07
09:21
09:40
09:55
10:11
10:26
10:42
10:54
11:13
11:28
11:44
11:59
12:15
12:30
12:46
13:01
13:16
13:32
13:47
14:03
14:18
14:34
14:49
15:12
15:20
15:36
15:51
16:07
16:22
16:35
16:53
17:09
17:24
17:40
17:55
18:10
18:26
18:41
18:57
19:12
19:28
19:43
19:59
20:14
20:30A 37-minute conversation between Riley Brown and Ras Mic (Michael Shimeles) on why the model provider building its own tools is changing the AI coding wars.
June 25th 2025A 22-minute breakdown of Anthropic's Mythos-class Fable 5 model — live demos, a structural argument about building blocks, and a pricing window closing June 22.
June 12thA 26-minute breakdown of OpenAI’s Intelligence at Work event: Codex merges with ChatGPT, role-specific agent plugins become startup killers, and Sites turns vibe coding into hosted apps.
June 3rdA 27-minute breakdown of why Opus 4.8 barely moved the needle and why Codex platform updates mattered far more.
May 31stClaude, Codex, Google, and Cursor are all racing to build the same thing — and only two of them know what they are building.
May 23rdRiley Brown and Ras Mic spend 83 minutes mapping the entire 2026 AI super-app war: Codex pulling ahead, Anthropic spreading too thin, SpaceX semi-acquiring Cursor, OpenClaw eating mundane work, and why a great coding model is now the only model that matters.
May 2nd