I can officially retire from X now.
Sam Altman just quoted my tweet. 🔥
1407916.7K
Stats Wire✓
@StatsWire
AI Engineer | Working with Agents I build AI explainers with code deep dives on YouTube
194 Following3.4K Followers11.8K Posts
- Gemini 4 Carbon isn't even available to everyone yet, and Gemini 4 Xenon rumors are already circulating. Claims of better performance than Claude Fable 5.5 and cheaper inference. Nothing confirmed yet. Rumor.005161
- I'm starting to wonder why Claude Fable 5.1 costs 2.5x more than Opus 5.5. Fable 5.1: $10/$50 Opus 5.5: $4/$20 Opus is cheaper, leads the coding benchmarks, and I'm seeing people prefer it for larger codebases. What exactly are we paying the extra money for with Fable?100141.3K
- Rody, I just got ANOTHER Antigravity update. Now you're making me even more suspicious. What are you guys cooking?
Rody Davis @rodydavis30271.3K 
- I've seen the “Restart to Update” button in Google Antigravity 3–4 times today alone. Every time I restart, another update seems to be waiting. Google is definitely cooking something behind the scenes.
1531357.5K - Claude Code's usage-limit complaints are still coming in. New GitHub issues today: • Usage draining much faster than last week • Weekly quota showing 96% in one place and 99% in another • Max users questioning why unchanged workloads now hit limits sooner Anthropic still hasn't explained the broader pattern.71342K
- Gemini 4 Carbon is faster than both GPT-6 Astra and GPT-6.1 Sol. Equal to Opus 5.5 at 98 tokens/sec. Output speed: • Carbon: 98 tokens/sec • Opus 5.5: 98 tokens/sec • Sol: 79 tokens/sec • Astra: 78 tokens/sec
62452.7K - Gemini 4 Carbon just became the world's #1 AI model, but its hallucination rate has nearly TRIPLED. Gemini 4 Argon: 4% Gemini 4 Carbon: 11% That's a 2.75x increase in hallucinations despite Carbon being the more intelligent model.
50171.1K - OpenAI was hours away from cutting $200 Pro users in half. Instead they cancelled the cut and called it “doubling.” Here’s the real math: Pro 100: 5x → 10x Pro 200: stays at 20x Pro 500: 25x → 50x Plus: NOTHING They almost took half your quota, then gave it back and asked for applause.123543.9K
- 🚨OpenAI just doubled Codex and Work usage limits for paid Pro tiers. Pro 100, 200, 500 and Business Premium all get 2x usage, effective today. Pro 200 users won't lose those limits later either.Tibo @thsottiauxDay 25/ We are doubling all usage limits for all Pro 100/200/500 tiers and Business Premium, effective today. We heard your kind feedback and won’t be reducing usage limits for Pro 200, so your Work and Codex usage will stay the same. We’re also going to continue offering extended limits for dots for now, and usage of your dot will still continue to be separate from your default usage. And we'll of course also keep shipping better models and other efficiency improvements that increase your effective usage over time.3015906
Day 25/ We are doubling all usage limits for all Pro 100/200/500 tiers and Business Premium, effective today. We heard your kind feedback and won’t be reducing usage limits for Pro 200, so your Work and Codex usage will stay the same. We’re also going to continue offering extended limits for dots for now, and usage of your dot will still continue to be separate from your default usage. And we'll of course also keep shipping better models and other efficiency improvements that increase your effective usage over time.6.8K2.5K36.3K4.6M- Gemini 4 Carbon is at #2 on Agent Arena, ahead of almost every OpenAI and Anthropic model. • Claude Opus 5.5: +21.1% • Gemini 4 Carbon: +16.2% • Claude Sonnet 5.5: +13.9% • GPT-6 Astra: +13.4% • GPT-6.1 Sol: +12.7% Only Opus 5.5 is ahead of Carbon.
71261.8K - Claude Fable 5.5 hasn't even been released yet, and already hearing rumors about Claude Opus 6.31282.8K
- Mistral just dropped Mistral Large 4 ("Le Chonk"), a 1 Trillion parameter open-weight MoE model on Hugging Face. • ~1T total / 52B active parameters (128 routed + 1 shared expert) • 1M token context window • Natively multimodal (vision + text) with configurable reasoning (low/medium/high)
006628 - Wait, OpenAI is now giving ChatGPT subscribers API credits worth their entire monthly subscription? • $100 Pro → $100 API credits • $200 Pro → $200 API credits • $500 Pro → $500 API credits And you still keep your ChatGPT subscription. You have to activate the benefit, and unused credits expire each month.
607945 - GPT Image 2.6 could be arriving sooner than expected. References to the model have surfaced in OpenAI's API specifications, including a new image-editing feature called 'markup'
10201.1K - Gemini 4 Carbon just took the #1 AI intelligence ranking. But Claude Opus 5.5 is still ahead in coding. Intelligence Index: Carbon: 59 Opus: 58 Coding Agent Index: Opus: 62 Carbon: 59 Google has the overall lead. Anthropic still has the coding lead.21241.6K
- ChatGPT's new model picker is confusing, so here's the difference: • GPT-6 + High = Sol • GPT-6 + Extra High = Sol at maximum reasoning • GPT-6 + Pro = Astra-powered GPT-6 Pro Selecting Extra High does NOT give you Astra Pro. To use the Pro model, select Pro.727419.6K
- Tibo launches Lauren: I can't afford Alexandr Wang joins the conversation. Lauren: WHO INVITED YOU? Alexandr: “I thought we were all friends” Meanwhile, Dario Amodei has Anthropic running like a military boot camp. No fun. Just ship Claude.
10524020.7K - Google priced Gemini 4 Carbon at a fraction of what its biggest rivals charge. Carbon: $2/$10 Claude Opus 5.5: $4/$20 GPT-6 Astra: $10/$50 Carbon is 2x cheaper than Opus and 5x cheaper than Astra.171717K
- Show more posts
