Marouane Lamharzi Alaoui (@marouane53) 's Twitter Profile
Marouane Lamharzi Alaoui

@marouane53

Twitch : twitch.tv/marouane53 🇲🇦🇺🇸
Pro Poker Player & Streamer & Tech Entrepreneur

ID: 60883067

linkhttp://firassa.ai calendar_today28-07-2009 11:52:09

4,4K Tweet

54,54K Followers

2,2K Following

Marouane Lamharzi Alaoui (@marouane53) 's Twitter Profile Photo

It's crazy how fast AI is advancing. Just a few months ago, Claude was my daily driver. I chatted with it every single day. But now, whenever I need anything that involves reasoning or requires higher intelligence, I turn to o3. And I don't even use Claude for programming anymore

Marouane Lamharzi Alaoui (@marouane53) 's Twitter Profile Photo

نهار كتبت هاد البوست بيتكوين كان كايسوا تقريبا 500 دولار. ماعرف شحال تكون ساوية دابا ديك الارض الفلاحية في الريف

نهار كتبت هاد البوست بيتكوين كان كايسوا تقريبا 500 دولار.

ماعرف شحال تكون ساوية دابا ديك الارض الفلاحية في الريف
Marouane Lamharzi Alaoui (@marouane53) 's Twitter Profile Photo

The fact that this was ever legal baffles me. Thank you, Trump, for signing this into law. Hopefully, more countries will follow your lead.

Marouane Lamharzi Alaoui (@marouane53) 's Twitter Profile Photo

How crazy is it that this song is 100% AI-generated? Just a year ago AI songs sounded so fake you could instantly tell. Now Suno 4.5 makes songs with realistic Moroccan Derija lyrics and genuine Moroccan instruments. I'm honestly blown away by how fast AI music improved

Marouane Lamharzi Alaoui (@marouane53) 's Twitter Profile Photo

I genuinely doubt that Claude 4 could outperform Gemini 2.5 Pro, DeepThink, or even OpenAI o3, but honestly, I'd love to be proven wrong.

Marouane Lamharzi Alaoui (@marouane53) 's Twitter Profile Photo

After 3 prompts using Claude Opus 4 I gave it 75000 tokens in and I got 19300 tokens out. I still need to test the code and see if it's as good as Gemini 2.5 pro but it looks promising

After 3 prompts using Claude Opus 4 

I gave it 75000 tokens in and I got 19300 tokens out. 

I still need to test the code and see if it's as good as Gemini 2.5 pro but it looks promising
Marouane Lamharzi Alaoui (@marouane53) 's Twitter Profile Photo

I tested Claude Opus 4 with a particularly complex feature that I plan to integrate into my production code. It performed reasonably well, but the initial implementation wasn't complete. It couldn't get it done in one go. Gemini 2.5 Pro, on the other hand, handled the task more

Marouane Lamharzi Alaoui (@marouane53) 's Twitter Profile Photo

عندي واحد ال Benchmark ديما كنديرو فاش كيخرج شي AI جديد وهو كانقول له يكتب لي distrack و Ego Trip بالدارجة، وأغلب المرات كيخرج دكشي فيه بزاف ديال التخربيق وبلا معنى. لكن هاد Claude 4 Opus صراحة مزكني والكتابة ديالو عندها معنى وكيبان شوية ديال الإبداع.