🚨 Tibo STOP hyping, we know OpenAI's masterplan. Step 1: Find what's working well for other companies. Step 2: Make your version even better. Step 3: Rename it. Step 4: Ship it. Step 5: “You don’t know what we’re releasing next week.” FINAL STEP: Release it. (Grok Bot → ChatGPT Dots )

I apologize in advance for this crash out, but holy shit I'm so tired. I'm trying to figure out what is even being measured here and it's nearly impossible. I read the post. It provided no insight whatsoever into how these "tests" work. Things that were not mentioned: - Harnesses used - Tasks used - How many times tasks are run - What is done to identify variance in daily runs - What analysis is done on "bad" runs to detect root causes for failures - What APIs are being used (matters a LOT) - How the +/- 10% "variance" was selected - Why tokens and costs are "weighted" the same, and combined rank roughly as high as intelligence - Why tokens are included at all when costs are the metric that matters I'm sorry @bridgemindai - am I missing something here? I just want to make sure I understand fully before building my own alternative bench. I DM'd you asking for traces from your runs, would be super helpful as I start digging deeper here.

I'm down to put my money where my mouth is. BridgeMind put $10,000 towards benching, I'll do the same. If my benches show I'm wrong, and that there is meaningful model "nerfing" aligned with his viral post, I will donate another $10,000 to a charity of @bridgemindai's choice. If my benches show I'm right, I want Bridgemind to take down "nerf bench" and replace it with a page apologizing, listing all the reasons the bench was flawed (written by me). I'm even down to have a qualified third party come in and audit both of our work. Deal?

All right, we’re locking in. Only things being worked on are simplifications, more efficiency for more usage, groundbreaking features or new models. Sometimes you have to invest ahead of the curve, but feedback is clear that you all want things to get simpler. On it.

Opus 5.5 has been nerfed already!?? Looks like we’re about to see the worst version of an otherwise incredible model over the next few days

Anthropic has a usage limit problem, and I finally have receipts. I forensically audited Claude $100 Max’s 5 hour meter using Claude’s own UI history. Here’s what I found: Yesterday 4h38m into the window: 67% used 33% still available Today 2h38m into the window: 100% used nothing left Same Claude Max plan. Same Opus 5.5. Same workload on my end. And this isn’t “Claude feels like it runs out faster now.” My Codex Computer History feature captured Claude’s actual interface in real time, including the usage percentage and exact reset countdown. I had Codex tear through the raw event history and verify it. It isolated the 5-hour meter from the weekly/context limits, normalized timestamps, caught stale UI readings, cross-checked Claude’s own saved usage history and transcripts, deduplicated the records, and verified 1,669 distinct response entries against their original source lines. How does the same “5-hour” allowance go from having 33% left near the END of one window to being completely exhausted near the MIDDLE of another? Scamthropic is at it again.

🚨 Antigravity just got a SERIOUS model upgrade. Google just added Anthropic’s latest models: → Claude Opus 5.5 → Claude Sonnet 5.5 → Thinking modes → Low / Medium / High reasoning And GPT-OSS 120B is being phased out. But here’s what caught my attention. Google is putting Claude’s latest models inside its own platform while its own Gemini 4 Argon is sitting on the sidelines. If Argon gets added too… Antigravity could become the place where you can throw the best models against each other. Now THAT would be really interesting. 👀🔥

Yeah😂 Google hosting Anthropic models is definitely a plot twist.

Wtf 👀 that's so bad. if this is what we are getting on pro plans it's feels more like marketing gimik to get people on antigravity

anthropic is going hard before the ipo. fable 5.5 is better than gpt-6.1 astra, apparently. i don't think that comes as a surprise tbh. opus 5.5 and sonnet 5.5 were already big jumps. fable 5.5 is the big model they've been cooking for months.

Someone just burned through 70% of their $500/month ChatGPT plan in ONE HOUR because they accidentally left Ultrafast Mode on $500/month and you can still accidentally speedrun your entire usage?? This doesn't make sense to me... Right??

I'd never expect this after paying 500$. Such a scam

Show me a bigger LLM failure than this. I’ll wait.

🚨 Anthropic isn't trying to beat OpenAI anymore. It's trying to make OpenAI irrelevant. And the lineup tells you exactly what Anthropic is building: → Opus 5.5 for the hardest work → Sonnet 5.5 for everyday coding → Haiku 5.5 for speed and cost → Fable 5.5 could be the wildcard sitting above them all An entire lineup covering almost every kind of AI workload. And that’s when this stops looking like a model race… and starts looking like a full-stack attack on OpenAI. 👀🔥

Anthropic isn’t innocent either. Fable 5.5 is probably going to be something only insanely wealthy users can properly access, and you don’t even get a preview of it on the $20 plan. At least OpenAI lets Plus users try Astra, even if only briefly. So yeah, somehow Anthropic manages to be even worse on this front, and I absolutely hate that.

🚨 For the first time in a long time, Anthropic doesn’t have to worry about OpenAI. It has to WORRY about itself. Opus 5.5 is already insanely capable. And it costs $4/$20 per million tokens. Fable 5.1 costs $10/$50. So when Fable 5.5 arrives, Anthropic has a VERY interesting problem. It either needs to match Opus 5.5’s price… or be so much better that developers actually WANT to pay more. Because if Fable 5.5 is only incrementally better than Opus… Why would anyone pay the Fable premium? For once, Anthropic isn't just trying to beat OpenAI. It's trying to beat its own best model. 👀🔥

Paying this 9th month in a raw Still making $0 What about you guys?

Why does a “free reset” also reset the timer for my next weekly reset?LOL I assumed it would refill my quota while keeping my normal reset date. Instead, it basically starts a brand new 7 day cycle. And yep, OpenAI confirms that’s how full banked resets currently work. So it feels less like extra usage and more like pulling your next week’s usage forward. Kinda defeats the whole “free reset” feeling IMO.

pov when you hire vibe coder for rejecting an email 😭

openai support is the worst shit i have ever seen. they banned my account 7 days into a 229€ month. no reason given. just "recent activities". so 175€ gone for days i cant use, wanted a refund. things i tried: -appeal (sep 22, not even a confirmation) -second appeal (confirmed, 0 answer since) -refund request (closed by an ai after 39 seconds. "we will not respond to additional support requests") -gdpr data request -any mail from my account email (auto rejected. all of them) so i wrote from another email and got "escalated to a support specialist". Hrhrhehhehehe one day later the specialist answers: "im an ai support assistant". next line: "hi, im ai-assisted support". Lil bitch luna introduced itself twice. what the "specialist" did: -asked for the order number i sent one mail before -asked when i canceled. you canceled it. by banning me -told me to submit an appeal (did it twice) -answered my gdpr request with "log into your account and export your data". THE ACCOUNT IS BANNED -i answer "give me the specialist" and repeat everything to show some sort of complexity. THE AI SUPPORT GHOSTS ME NOW are they stupid? and it cant read one email. pretty sure keeping money for a service you took away is not even legal in the eu. @OpenAI give me a human or my 175€

Has anyone actually lost their job because of AI? I haven’t met anyone yet.