Ammar Alyousfi

@htihle @scaling01 Please use the official API. I don’t trust all third party providers

DeepSeek v4 pro (max) scores 48.9% on WeirdML, improving on v4 pro (high) at 46.5%, but still well behind Kimi-k2.6 and GLM-5.1 at 56% and 57%, let alone the closed frontier. These runs, like the previous ones, were through Fireworks AI.

@AriX @thsottiaux Got it. Thanks. I thought it might be related to a specific setting in Codex.

@ammar_cel @thsottiaux This is controlled by the training toggle you see in ChatGPT!

Codex Compute efficient ✅ Always up, never down ✅ Best at hardcore engineering ✅ Crazy good app, first to escape the terminal ✅

@AriX How?

So excited to share that we're bringing Computer Use to Codex. Computer Use lets Codex see, click, and type into your Mac apps, with its own cursor. It's a magical feeling to have agents using your apps in the background, and still get to use your computer at the same time.

@thsottiaux How? @thsottiaux

@MilksandMatcha C O D E X

Giving away 5 Codex Pro plans Each person will get 3 months of free Codex Pro (highest tier). Winners will be selected from comments in 48 hours, comment below why you want it.

@simonw Anthropic is cringe!

This stunt feels irresponsible to me. If we don't want regular people developing toxic relationships with their chatbots it really doesn't help for leading labs to start giving them "retirement interviews" and encouraging them to blog their "musings and reflections"

@kepano When I'm writing a list and all the list in Arabic but one item in English, that one item becomes left aligned, and vice versa, as you can see in the image. It'd look much nicer if it followed the list alignment.

ما الذي يزعجك في استخدام Obsidian للغات التي تُكتب من اليمين إلى اليسار؟

@cifilter That’s right. It’s infuriating to see most of my 5 hour limit consumed in one Opus query!! It makes me wanna cancel my Pro subscription immediately. I regret subscribing. Codex is sooo much better in this regard.

Is it just me, or does OpenAI give you a ton more usage than Anthropic for the same price ($20/month)? I don't think I've once had Codex tell me I'm at my limit, and I use it a ton. Claude frequently seems to put me in timeout for several hours.

@flynas 👍💚

في السماء حكايات لا تنتهي .. وطيران ناس سوريا حكاية ستُروى للأجيال 🇸🇦🇸🇾💚 حكاية ربط، وشراكة، ومستقبل يُكتب من جديد ✈️ #نربط_العالم_بالشام

@NotionHQ Finally!!! Thank you.

Notion is now available in 21 languages, including Arabic and Hebrew. The first right‑to‑left (RTL) languages supported in Notion! Text flows naturally, tables align correctly, and everything reads the way it should. We're happy more teams can work in the language that feels like home.

@burkov My experience with Opus 4.5 on Claude Code in the past few days was so frustrating that I cancelled my subscription. Like you, I felt I was dealing with a small model! I don't trust them. But what can we do?

It must be illegal to sell access to one model but serve a different one. Since yesterday, Claude has been making me angry, and it's the first time I've felt angry since I started exclusively using it for coding two months ago. I chose Opus 4.5 as the model for which I'm charged $100/month, but I know I'm being served something similar to GPT-3.5. This is fraud and theft.

@rork I’d love to try

Rork is the easiest way to use Claude Code with Opus 4.5 to build real mobile apps, publish to App Store in 3 clicks and start making revenue Reply to get a free 25$ subscription. Ends tomorrow!

@spencerschiff_ It has some good use cases, but the hype created before its release was too much more than reality.

Is anyone actually using Gemini 3 at this point? All the hype on my timeline is about Opus/5.2

@Dr_Derma_ هذا كان كذبة لعدة أسباب... هذا بحث مع ChatGPT مع مصادر: chatgpt.com/share/6954f968… + كان ممكن يقنعنا أنه قادم من المستقبل بطرق بسيطة يعني ممكن يقول: "في اليوم الفلاني في الساعة الفلانية في المكان الفلاني سيحدث كذا" عن عدة أحداث... لكن اتبع طرق المخادعين، وممكن ساعده الحظ قليلاً

@OfficialLoganK @GoogleAIStudio Add me please!

Big upgrade to vibe coding in @GoogleAIStudio lands in Jan, but if you want to test early… 👇🏻

@chromeunboxed This keeps reappearing, no matter how many times I close it!

The Gemini web interface is getting a fresh look today, featuring a true black dark mode, a new "My Stuff" section, and a cleaner design. chromeunboxed.com/gemini-for-web…

@BytePlusGlobal Doesn't support my language like Nano Banana Pro. No, thanks :)

Today, we’re introducing Seedream 4.5 — a refinement-focused upgrade designed to bring clearer visuals, stronger subject consistency, sharper text rendering, and more reliable multi-image execution to creative and production teams worldwide. Built to support real workflows across e-commerce, design, advertising, film, animation, and game art, Seedream 4.5 improves: 1. Detail fidelity and aesthetic cohesion 2. Spatial reasoning and scene structure 3. Complex prompt execution and editing precision 4. Multi-image fusion with up to 10 references 5. Clear, readable small-text and facial rendering Seedream 4.5 is now available in open beta via BytePlus ModelArk. Learn more and explore sample cases on the platform. #Seedream #BytePlusAI #ByteDance

@thedankoe Seems great! Can you support RTL writing please?

Chat interfaces suck. And you need to copy paste between 10 tabs to do anything worthwhile. So we created a canvas for notes, files, YouTube videos, and chats that can branch off of each other. Eden is only open to the public through black friday weekend.

@fofrAI I noticed it struggles with multiple reference images of people. It changes facial appearances significantly in the generated image

Nano Banana Pro is very capable, I want to go deeper with it soon. Reply below with things you’d like me to cover.