ro

@beckposting tell me why the sky is blue but as if you were an elephant

@SSHTheDev Idk I feel like this has a clean lighting effect which is hard to do. It’s not 100% realistic but imo it’s right behind fable which I think has the best

Does it beat gpt 5.6 sol? Kinda... personally I still like sol's better. #viewer" target="_blank" rel="nofollow noopener">bhbench.ssh.codes/model/qwen38#v…

@GuntherEagleman lol imagine. Imagine gargling one man’s balls so much you believe that ONLY 2 senators were illegitimate but the other 30 or so were fine. Declare them as such and also say republicans are also illegitimate or your just a pretentious fuckwad deepthroating trump hoping he’ll bust.

🚨 BOOM: President Trump is set to declare BOTH Georgia Democrat Senators ILLEGITIMATE over massive election fraud, Ossoff and Warnock’s seats are on thin ice! This drops right after FBI raids hit Georgia election offices hard. The walls are closing in. Time to EXPOSE IT ALL and drain the swamp for good! LFG!

This video was paid for by a benefactor.

@bridgemindai isn’t credible. But it’s good to know he too is high off of fable *now. Who knows what his position will be in 2 hours lmao.

@MatthewBerman OpenAI has lined the pockets of the Trump Administration so they wouldn’t face the same scrutiny. Considering it’s been widely reported that 5.5 can do a lot of the same exploits as Fable, and they haven’t gotten banned, it seems this corruption works in their favor.

why would openai release their next model right now? they've been claiming their next model (and current models) are as good as fable at cyber. so if they release their model while fable is offline, all attention will be on them and people will ask why their model is allowed to be used by foreign nationals. so either the model isn't as good, or fable should be public right now.

@OpenAI @sama will you take down GPT 5.5 because it too can hack into things and pose a natsec threat are you guys too busy deep throating Pete kegseths schlong?

Me and fable were about to get married and have children and now I can’t gargle Claude’s balls and it’s all ur fault. Bring me my claudipoo. Give me my pookie back Also fuck you.

@PeteHegseth @DeptofWar you guys gargle more balls than my gay cousin and are such incompetent cucumbers and sorry excuses for crackers that you had to take away Claude from me.

@davis7 I think the slowness is temporary as people are rushing to use the latest model. I’ve had a few little bursts where my TPS estimators clocks it at 60TPS but today it’s been averaging 30-40TPS

Testing Fable, it's clearly a great model Quality of code on a couple effect v4 snippets is out of the gate really solid. Also having it do a bunch of stuff involving deployments that touch convex, cloudflare, depot, clerk, and it's figuring out all of it rly well Also credit where it's due, Claude Code is better than I tend to give it credit for - workflows continue to be amazing - auto mode is great - spinning up sub agents for things like research is great The way it fans in/out of workflows is great. Test I'm doing rn is cleaning up old effect code written by a model that's full of dumb promise syntax where it shouldn't be. Model's doing an amazing job An interesting note on auto mode is that the model is clearly fully aware of whats happening. It noticed that it was blocked from something, and told me to respond with an explicit approval so that the check would then let it do what it wants to... Biggest issues are definitely speed and price. Even on low reasoning, this thing takes several minutes at a minimum for even a simple "push this change"

@anshuc You’re building a very cool thing I can’t wait to see it go GA.

Thought I’d introduce myself since I got some new followers from my launch post (welcome to this crazy journey!) Would love to connect with folks interested in local, personal, proactive AI, or startup life Who I am: I led teams at Apple working on AI for future products. Then did a voice AI startup with friends (YC F24). Now I’m building Marina, my vision for the ultimate proactive assistant, using all of my learnings from those experiences

What if your Mac could read your mind? Here’s an early live demo of Marina: personal, proactive intelligence for macOS. There’s no prompting. No chat box. It just understands what you’re working on, and helps you do it faster. Details + beta invites below.

@anshuc Yeah to be honest, I haven’t tried this out at all but it was just something I know overlay apps tend to struggle to implement. Perhaps in full screen mode you could make use of a dynamic notch? Or instead of fully transparent glass a more frosted glass?

@ro_dawoof good question, I hadn't tested this. it does work, but the UX is kinda weird as it ends up overlapping the window. this'll be a fun one to figure out… thanks for the bug report haha

oh, and if you know any extremely cracked and highly opinionated people who want to build something that seems impossible, my DMs are open 👀

that's the update next one in june, allegedly i have been a person for somewhat longer than any of this none of it is real all of it is news i definitely did not miss anything

the same department of war then asked anthropic for permission to use claude for, quote, "all lawful purposes" without limitation which apparently meant mass domestic surveillance and fully autonomous weapons anthropic said no surprise surpris

news observations: feb through may 2026, since i forgot to do one in feb march and april