Iosif Chatzimichail retweetledi
Iosif Chatzimichail
372 posts

Iosif Chatzimichail
@afonic
Founder @ https://t.co/fdZ6n3OeQS Web developer and digital marketing specialist.
Kos island, Greece Katılım Mart 2009
282 Takip Edilen116 Takipçiler

@ns123abc That Cursor valuation is a joke. Also it was obvious there were using another model, they are not a lab and frankly not disclosing this is closer to fraud that "bad vibes".
English

🚨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.



Elon Musk@elonmusk
@fynnso Yeah, it’s Kimi 2.5
English

@theo If it has to work for over an hour in a single feature you're doing something wrong.
English

@pushpak1300 Thanks for running all those benchmarks! What effort level was used?
English
Iosif Chatzimichail retweetledi

@PovilasKorop Did you use Laravel Boost and a tool like Context7 for AlpineJS docs ?
English

I tried to "blindly" vibe code the app with Laravel/Filament.
With the goal of TOTALLY not looking at the code and seeing what happens.
First 5 simple features went ok.
But then smth didn't work and Claude didn't know how to fix Filament/Livewire/Alpine issue.
But worst part is that I didn't understand the code at that point, I lost control.
And then, even as a developer, I felt lazy to dig and understand why Claude did some things the way it did.
So, imagine a non-dev vibe-coder. Clueless.
Not everything can be solved with (re-)prompting.
When you get out of CRUD-like apps universe, the real life is much more tricky, with many more edge cases that you wouldn't even know how to prompt for.
So yeah, real devs on serious projects shouldn't lose their jobs because of AI.
Unless companies they work for will struggle with their business model because of AI.
English

And because a demo is a thousand posts, here you go: demo.easy-forms.dev
English

Easy Forms v2 has been released fully supporting the new and shiny @statamic v6!
We also added Grid field support in the frontend so now you can have repeater-like functionality fully supported throughout Statamic!

English

@MediaKing Wordpress yes. Building a custom CMS is not cost efficient though. A platform like @statamic with easy customization, config and content in flat files is the way to go.
English

@JackEllis Exactly, most of the posts are engagement bait. Both models work well, if it's working for you no need to switch at the moment. Most of the "oh wow it's so stupid / intelligent" stuff also are down to tooling and workflow. 99% of "it one-shotted this" posts are clickbait.
English

Be careful with tweets like this. My guess is that this is comparing Claude code to codex, not the actual models. We have used both models via Cursor and Opus wins for us. I also prefer 5.2 non codex to codex. Make up your own mind and ignore us all. Thats all, enjoy your day.
BridgeMind@bridgemindai
I had a 4.5 hour vibe coding session today Claude Opus 4.6 was failing to complete a lot of tasks GPT 5.3 Codex was one shotting the tasks that Claude Opus 4.6 failed at Take note of this
English


Following the release of @statamic v6 we've been busy updating all our addons. Livewire Filters is now v6 compatible Oh and as a bonus it's also Livewire 4 compatible! Check it out, you might not even know you need it!

English

Statamic 6 is out! Phenomenal work by everyone involved to release this beauty!
statamic.com/blog/statamic-6
English
Iosif Chatzimichail retweetledi
Iosif Chatzimichail retweetledi

@AnthropicAI could price Opus at $1000 / month and it would still be worth it imo.
English

@taylorotwell I've been using Laravel since v5. Opus 4.5 with Laravel Boost is amazing. Well done Taylor and team!
English

This is why Laravel + AI agents just works.
The opinionated structure means agents know where to put things and write clean, organized, human-readable code. It scales to large apps even when vibe coding.
New doc 👉 laravel.com/docs/ai
English
Iosif Chatzimichail retweetledi








