@MinLiBuilds: A genius team won't be buried; Musk is a true talent scout. In early 2025, Cursor and Claude complemented each other, one as the programming harness, the other as the model. In early 2025, Altman and Liang Wenfeng both disliked the coding scenario; one worked on Sora, the other on competitive low-cost inference. ...

A genius team will never be buried—Musk is a true talent scout.

At the start of 2025, Cursor and Claude complemented each other perfectly: one handled the programming harness, the other the model.

At the start of 2025, Altman and Liang Wenfeng both disliked the coding scenario—one worked on Sora, the other on competitive low-cost inference.

At the start of 2025, Cursor charged $20 per month, Copilot $10. Back then, Cursor paired with Opus 3.5–3.7, and users raved about how great it was—it was the hot new thing in Silicon Valley, once valued at $7 billion.

The turning point came in the middle of Q1: Claude heavily used Cursor’s business data to distill 3.7 and later versions, launching Claude Code.

Cursor’s business kept declining, its growth derivative turned negative, while Claude’s growth derivative was exponential.

2025 was truly a magical year. In the second half, Claude, Gemini, and Codex all recognized the programming and long agentic scenarios, with Grok following closely.

Cursor was no longer on the list of frontier AI companies. They didn’t have their own model, no computing power—people mocked them for making a wedding dress for Claude. Even fine-tuning a Kimi would get them roasted.

The great, brilliant, human prophet @elonmusk stepped in and provided Cursor with full support, just like he did for OpenAI back then.

The genius Cursor team also delivered a perfect answer. Congrats Cursor, congrats SpaceXAI.

Michael Truell (@mntruell): Excited to release Grok 4.5 with @SpaceXAI.

It’s an Opus-class model that’s fast and low cost. It’s a significant step up over any model we’ve developed so far, including Composer 2.5, and has become the daily driver for many on our team.

First of many releases. More soon.

Now that would be quite the comedy and drama.

The Cursor team originally trained large models for CAD, but found it didn’t work, so they pivoted to coding harness.

They are truly brilliant. Actually, their approach is a bit like Manus, right?

Just like your name suggests, in the Journey to the West, those who survive all have a background.

I’m an early practitioner of AI-assisted coding, so I have a clear picture of the early information.

The process is the result. As long as you hold on, the outcome won’t be bad.

I hope Grok can get better—they aren’t anti-China.

Cursor and SpaceX merged—a win-win, unstoppable.

No one saw this script coming at the start of 2025.

What other talent-spotting moves has he made? Haha.

But he really knows how to manage. When everyone wasn’t optimistic about Grok, he made a masterstroke.

Grok is the world’s Grok—the least anti-China of them all.

Got it? Musk could write a second biography now.

Musk made a great buy here—looking forward to SpaceX’s performance.

Makes a lot of sense. In this chaotic AI era, several heroes have emerged.

Musk is incredible. Truly impressive. Well done, Musk—I’m your fan.

Cursor’s story is too dramatic. Even Wu Cheng’en couldn’t write this.

OpenClaw proved itself effective, so all the companies followed.

A classmate of mine was a product manager at Lanhu at the time. He was doing prompt engineering on 3.5, and the product was very good. When 3.7 came out, they realized all their work was in vain.

Too amazing—is Musk a time traveler?

Install Grok CLI and try it out. If it truly matches the level of Opus 4.8, Musk will become a god.

Indeed, Grok is on the right track now—a 2T model.