Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Cursor Introduces Composer 2.5

cursor.com

Post reply on HN
Print view View original

211–220 of 238 posts

Re: Cursor Introduces Composer 2.5

#211

Post by WhitneyLand »

Aurornis wrote: post #166
WhitneyLand wrote: post #155

Say what you want about Cursor but they don’t lack for ambition. Forking VS Code, going big on bleeding edge features like cloud agents, and now they’ve thrown down the gauntlet directly challenging frontier labs by training their own model (“much larger” than Kimi 2.5’s 1T parameters) from scratch. They’ve been highly successful so far. Raised $50B, $2B in revenue, forecast to end 2026 above $6B. But even at these h…

EDIT: As others have pointed out, the comment above contains hallucinations (Like the $50 billion number) and a lot of AI tells. The account doesn’t have a history of AI-like comments but the hallucinations and structure in this one are suspicious. If anything, don’t trust the numbers it cites because they’re made up. Cursor is a team that I want to see succeed. They have stacked their company with very smart people…

So, you’re wrong on two counts.

1. Evidently you’re no longer able to distinguish AI from people as the whole comment was written by a human off the cuff.

2. The numbers are not hallucinations. It’s word on the street reporting, so yes it’s speculative, but a model did not make up it up unless that’s where TechCrunch got it which is not on me.

https://techcrunch.com/2026/04/17/sources-cursor-in-talks-to...

Re: Cursor Introduces Composer 2.5

#212

Post by WhitneyLand »

Aurornis wrote: post #191
pdq wrote: post #178

Earlier quoted context omitted.

Why is this comment upvoted? It is most likely AI generated with a nice "Raised $50B" hallucination and filled with cliches ("thrown down the gauntlet", "mountain you don’t climb just once", "not for the faint of heart").

Good catch. I didn’t even notice it at first, but the hallucinations on top of cliches gives it away. The account doesn’t have a history of other comments that have too much of an AI vibe, but this one does. Even if it wasn’t AI, it’s misinformation.

Please see reply to your other comment on this thread.

Re: Cursor Introduces Composer 2.5

#213

Post by Bolwin »

scosman wrote: post #133
throwaw12 wrote: post #78

> Composer 2.5 is built on the same open-source checkpoint as Composer 2, Moonshot's Kimi K2.5. Really nice to see they're giving credit to the company and I am optimistic Kimi K open models soon will outperform Opus models

> I am optimistic Kimi K open models soon will outperform Opus models Hard to outperform the model you distill...

They're not distilled. Stop spreading anthropics misuse of the term.

They do use it for synthetic data/judging though, so yes, hard to outperform.

Not that they need to. If they can basically match it for a fifth of the price.

Re: Cursor Introduces Composer 2.5

#214

Post by ryanshrott »

The cost claim is the easy part to sell. The real test is whether it stays useful in ugly codebases, long files, and repos with a bunch of half-broken conventions. That’s where these assistants usually fall apart, even when the benchmark numbers look great.

Re: Cursor Introduces Composer 2.5

#215

Post by vessenes »

bfeynman wrote: post #193
vessenes wrote: post #121

Earlier quoted context omitted.

Sounds like it's the last Kimi-line model at Cursor? As expected they say they'll be training a larger model on the SpaceX infrastructure, or have already started most likely. I'm very curious to read about the Composer 3 architecture when it comes out. More frontier coding models are a good thing, especially if they diversify into different strengths/weaknesses.

That only seems plausible if whatever corpse of xAI is around is giving them engineering time. I don't know if they hired a bunch of ex frontier lab staff but its unlikely they have the technical capability to train their own frontier models especially the pretraining. Because the thing is if its not competitive with claude/codex it will be panned.

Hmm, I read the situation a little differently. Grok is not a slouchy model. It’s not the best, but it’s not the worst. X currently has one source of proprietary data, Twitter, and grok is by far the best at all the things you might imagine there - today’s zeitgeist, who’s saying what, current news, etc.

Cursor adds in a large corpus of proprietary coding data — I think this is actually fairly hard to acquire right now, because claude and codex are so good.

I bet there’s enough talent at the Grok team to work with the cursor team and data to get something good out the door.

That said, I don’t track Grok’s engineering leads — I’m not sure who’s currently around, and who is not.

Re: Cursor Introduces Composer 2.5

#216

Post by Aurornis »

WhitneyLand wrote: post #211
Aurornis wrote: post #166

Earlier quoted context omitted.

EDIT: As others have pointed out, the comment above contains hallucinations (Like the $50 billion number) and a lot of AI tells. The account doesn’t have a history of AI-like comments but the hallucinations and structure in this one are suspicious. If anything, don’t trust the numbers it cites because they’re made up. Cursor is a team that I want to see succeed. They have stacked their company with very smart people…

So, you’re wrong on two counts. 1. Evidently you’re no longer able to distinguish AI from people as the whole comment was written by a human off the cuff. 2. The numbers are not hallucinations. It’s word on the street reporting, so yes it’s speculative, but a model did not make up it up unless that’s where TechCrunch got it which is not on me. https://techcrunch.com/2026/04/17/sources-cursor-in-talks-to...

Quoting directly from your comment:

> They’ve been highly successful so far. Raised $50B,

They have not raised $50B. The article you linked says they're raising $2B, not $50B.

The valuation is not the amount raised.

Re: Cursor Introduces Composer 2.5

#217

Post by mgambati »

Glohrischi wrote: post #147
radu_floricica wrote: post #143

Earlier quoted context omitted.

Buy "Cursor", not "Cursor's IP". This means brand, users, and a shitton of data. And if you combine a shitton of data with a lot of compute, large userbase and good engineers, you have a pretty good chance of doing something interesting.

Yeah you know how much 10-60 Billion are? You could literaly just give your compute away for free for a year to pull people in. Make an API Endpoint for free with the caviat that they are allowed to use the data for traing, what everyone else does too.

And you still don’t get the quality of data that cursor have which is the best due to being collected pre vibe coding.

Re: Cursor Introduces Composer 2.5

#218

Post by ccimmergreen »

vessenes wrote: post #215
bfeynman wrote: post #193

Earlier quoted context omitted.

That only seems plausible if whatever corpse of xAI is around is giving them engineering time. I don't know if they hired a bunch of ex frontier lab staff but its unlikely they have the technical capability to train their own frontier models especially the pretraining. Because the thing is if its not competitive with claude/codex it will be panned.

Hmm, I read the situation a little differently. Grok is not a slouchy model. It’s not the best, but it’s not the worst. X currently has one source of proprietary data, Twitter, and grok is by far the best at all the things you might imagine there - today’s zeitgeist, who’s saying what, current news, etc. Cursor adds in a large corpus of proprietary coding data — I think this is actually fairly hard to acquire right n…

Unlikely, given that large swathes of talent have already left xAI, ostensibly due to poor leadership management. Simply throwing money in to build the biggest datacenters in the world doesn't do much good without bright minds to back it up. https://www.fastcompany.com/91531084/inside-the-xai-exodus

Re: Cursor Introduces Composer 2.5

#219

Post by WhitneyLand »

Aurornis wrote: post #216
WhitneyLand wrote: post #211

Earlier quoted context omitted.

So, you’re wrong on two counts. 1. Evidently you’re no longer able to distinguish AI from people as the whole comment was written by a human off the cuff. 2. The numbers are not hallucinations. It’s word on the street reporting, so yes it’s speculative, but a model did not make up it up unless that’s where TechCrunch got it which is not on me. https://techcrunch.com/2026/04/17/sources-cursor-in-talks-to...

Quoting directly from your comment: > They’ve been highly successful so far. Raised $50B, They have not raised $50B. The article you linked says they're raising $2B, not $50B. The valuation is not the amount raised.

So I made a mistake reading the article? So what?

The point is you made two brigade style comments about my posts sounding suspiciously like an LLM and having hallucinations.

Neither turned out to be true and I think a better response would concede the point.

It may be more helpful for us to stick together as humans since we can’t always recognize each other so easily anymore.

Re: Cursor Introduces Composer 2.5

#220

Post by maxdo »

ifwinterco wrote: post #198
ta20240528 wrote: post #195

Earlier quoted context omitted.

And when you here unsubstantiated rumours* that ­say Anthropic has been sending exchanges to say Alibaba's Qwen, will you als oconclude the same about the entire US AI industry? I doubt it. * publish the logs.

Even if it's true, it's not like US AI companies can complain, given their entire business is based on ripping off text without attribution

chinese ai is not doing the same? or they don't parse?

they do except they also send thousands of sex-spies to do espionage of this kind on the scale.

Post reply on HN