Nurlan Ibraev

276 posts

Nurlan Ibraev banner
Nurlan Ibraev

Nurlan Ibraev

@Nolikzero

Building @shipaitoday to help people start SaaS. Sharing OpenClaw automations, AI, and building in public.

Katılım Şubat 2018
291 Takip Edilen60 Takipçiler
Sabitlenmiş Tweet
Nurlan Ibraev
Nurlan Ibraev@Nolikzero·
Introducing ShipAI - Best Way to Start SaaS without paying to 500$+ every month for other services Its' included all (i mean really all) that you need start any digital product - Landing, App, Admin all in one - Seo Foundation - Internalization 7 languages - Billing - AI layer - Affiliate Program - Authentation - Search server - Mail Sender - Storage - Database - Monitoring, tracing - Background jobs - Analytics - Vector DB - AI Friendly - Well structured docs about from start to launch One time offer to independent from other SaaS monthy subscriptions
Nurlan Ibraev tweet media
English
2
0
2
769
Nurlan Ibraev
Nurlan Ibraev@Nolikzero·
@Rahll Just high iq people brings more to company worth with AI, than others
English
0
0
0
8
Reid Southen
Reid Southen@Rahll·
Wait wait wait waaait a second.... He's saying he wants his $500,000 engineer to actually cost him $750,000 by using $250,000 worth of AI tokens. Wasn't AI supposed to make things cheaper, not cost 50% more?
TFTC@TFTC21

Jensen Huang: "If that $500,000 engineer did not consume at least $250,000 worth of tokens, I am going to be deeply alarmed. This is no different than a chip designer who says 'I'm just going to use paper and pencil. I don't think I'm going to need any CAD tools.'"

English
371
404
5.8K
235.6K
Nurlan Ibraev retweetledi
TFTC
TFTC@TFTC21·
Jensen Huang: "If that $500,000 engineer did not consume at least $250,000 worth of tokens, I am going to be deeply alarmed. This is no different than a chip designer who says 'I'm just going to use paper and pencil. I don't think I'm going to need any CAD tools.'"
English
436
568
7.5K
2.4M
Sterling Crispin 🕊️
Sterling Crispin 🕊️@sterlingcrispin·
@Nolikzero Yeah brother I've exhibited my work all over the world and it's in the permanent collection of several major museums cope harder
English
1
0
0
37
Sterling Crispin 🕊️
Sterling Crispin 🕊️@sterlingcrispin·
Claude 4.6 is a good programmer but writes insanely severe bugs constantly, it won't catch them all in audits, nor will other claudes You need codex 5.4 auditing every commit 4+ times. If you don't believe me, try it. I have an /auditcodex skill for it github.com/sterlingcrispi…
Sterling Crispin 🕊️ tweet media
English
76
28
689
72K
Nurlan Ibraev
Nurlan Ibraev@Nolikzero·
My biggest regret in life is that when I was 16, I didn’t have Claude Code and Claude Opus 4.6.
English
0
0
1
62
corbin
corbin@corbin_braun·
composer 2 I am liking it.
English
9
0
61
4.4K
Nurlan Ibraev
Nurlan Ibraev@Nolikzero·
@SIGKITTEN China company uses own money to train model = ban, Cursor use their users data = great company, I am right?
English
0
0
0
24
Nurlan Ibraev
Nurlan Ibraev@Nolikzero·
@sawyerhood No other code agents uses user data to train "their" models like Cursor
English
0
0
0
120
Nurlan Ibraev
Nurlan Ibraev@Nolikzero·
@leerob Pretraining compute budget typically accounting for 80-95%+ of total FLOPs spent on developing the final model
English
0
0
1
113
Lee Robinson
Lee Robinson@leerob·
Yep, Composer 2 started from an open-source base! We will do full pretraining in the future. Only ~1/4 of the compute spent on the final model came from the base, the rest is from our training. This is why evals are very different. And yes, we are following the license through our inference partner terms.
Fynn@fynnso

was messing with the OpenAI base URL in Cursor and caught this accounts/anysphere/models/kimi-k2p5-rl-0317-s515-fast so composer 2 is just Kimi K2.5 with RL at least rename the model ID

English
311
150
2.3K
944.6K
Nurlan Ibraev
Nurlan Ibraev@Nolikzero·
@kvickart @ns123abc It's crazy how much hype there is around this, I'm so glad the X community always knows the truth
English
0
0
2
35
kvick
kvick@kvickart·
@Nolikzero @ns123abc I agree its crazy but its how these guys get to the top and maybe the only way to do so
English
1
0
0
62
NIK
NIK@ns123abc·
🚨NEWS: Cursor’s $50B “in-house model” is literally Kimi K2.5 with RL on top. Got caught in 24 hours >be Moonshot AI >spend hundreds of millions training Kimi K2.5 >1 trillion parameters, 15 trillion tokens, agent swarm architecture >beat GPT-5.2 and Opus 4.5 on real benchmarks >open-source it because you believe in the ecosystem >one condition: display “Kimi K2.5” if you make over $20M/month from it >Cursor takes the model >runs RL on coding tasks >ships it March 19 as “Composer 2” >blog post: “continued pretraining + scaled reinforcement learning” >zero mention of Kimi K2.5 >“our in-house models generate more code than almost any other LLMs in the world” >publishes benchmark chart >Composer 2 against Opus 4.6 and GPT-5.4 >uses the chart to justify raising at $50 billion! >less than 24 hours later >kimi dev intercepts the API response >model ID: kimi-k2p5-rl-0317-s515-fast >they didn’t even rename it >Moonshot head of pretraining runs tokenizer test >confirms: identical to Kimi’s tokenizer >publicly tags Cursor’s co-founder: “why aren’t you respecting our license?” >two more Moonshot employees post confirmations >all three posts deleted within hours >legal is now involved >but it gets worse >Cursor had Kimi K2.5 listed as a FREE model in their UI just weeks ago >users were openly using it >Feb 9: “K2.5 was in my model list. I updated and it vanished” >it vanished because Cursor pulled it from the picker, and relaunched it as their own model >Moonshot valuation: $4.3B >Cursor valuation: $50B Absolute state of Cursor.
NIK tweet mediaNIK tweet mediaNIK tweet media
Elon Musk@elonmusk

@fynnso Yeah, it’s Kimi 2.5

English
185
326
4.5K
513.2K
Nurlan Ibraev
Nurlan Ibraev@Nolikzero·
@katedeyneka Aren’t they, in fact, using their users’ data/tokens to post-train their own model - specifically, the latest models from Claude and GPT?
English
0
0
0
313
Nurlan Ibraev
Nurlan Ibraev@Nolikzero·
They also claim "our model" 😆
English
0
0
0
38
Nurlan Ibraev
Nurlan Ibraev@Nolikzero·
That was i talking for. I’m also certain they have no agreement with Moonshot AI, and they’re simply hoping there won’t be a lawsuit since the company is based in China. This is capitalism in its purest form: the world’s most advanced and fastest-growing company is using the open-source work of leading engineers and scientists without giving them any credit or paying any royalties.
Lee Robinson@leerob

Yep, Composer 2 started from an open-source base! We will do full pretraining in the future. Only ~1/4 of the compute spent on the final model came from the base, the rest is from our training. This is why evals are very different. And yes, we are following the license through our inference partner terms.

English
1
0
0
74
Nurlan Ibraev
Nurlan Ibraev@Nolikzero·
This is peak 2026 AI: Chinese open models (Kimi series, GLM, DeepSeek) are so strong that Western startups fine-tune them → slap on agent/UX moat → raise at insane vals. But license drama + no attribution = bad optics. Cursor silent so far. Tick tock...
English
0
0
0
43
Nurlan Ibraev
Nurlan Ibraev@Nolikzero·
"Cursor: magic RL ✨ Moonshot: that's our Kimi after RL and stolen tokenizer lmao" (@Avimanyu8008) x.com/Avimanyu8008/s… "A Kimi researcher confirmed then deleted. Probably secret deal?" (@shantanugoel) x.com/shantanugoel/s… Chinese media already calling it piracy vibes 🍉 x.com/BlockBeatsAsia…
Shantanu Goel@shantanugoel

This seems to be true! Composer 2 is just RL'ed kimi. A kimi researcher also confirmed this and then deleted their post, probably because there's a secret deal between cursor and moonshot

English
1
0
0
72
Nurlan Ibraev
Nurlan Ibraev@Nolikzero·
🚨 BIG AI DRAMA ALERT 🚨 Cursor just dropped "Composer 2" claiming it's their frontier coding model built with magic RL... But devs caught the receipts: the internal model ID is literally kimi-k2p5-rl-0317-s515-fast → Composer 2 = Kimi K2.5 + RL post-training Moonshot AI is PISSED 😤
English
1
0
0
69

Keşfet

@Rahll @sterlingcrispin @corbin_braun @SIGKITTEN @sawyerhood @leerob @kvickart @ns123abc