Serious Meester

@mitsuhiko Tbh: we already missed the train, the bus, the plane and now missing the spaceship.

Europeans please wake up.

Can’t wait to try it out. If there would be an European provider which serves open models.

Big LOL — it’s too dangerous, but if you are a US citizen everything is fine. Open-weight models are the future; it has never been more true than today.

@dargelito @twitchspion Er redet es ja noch gut, passt also zu der ganzen Truppe.

@twitchspion Also jemand der schlechte Bewertungen seines Restaurants löschen lässt um besser dazustehen macht bei Influenzerkollegen Undercovertest und prangert an...genau mein Humor....

Bernd Zehner hat einen Undercover-Test bei „Honest Kebab“ von Holle gemacht. >> Def Döner scheint nicht sehr honest gefüllt zu sein:

Das neue Apple Siri AI wird nicht in der EU verfügbar sein Wir haben so fertig.

@HamburgSent @mayukh_panja Beg you pardon. Doctors are less of a problem? 😂 you have the same issue in whole NRW.

@mayukh_panja Berlin does not represent Germany. Doctors appointment are less of a problem elsewhere. Regarding the passport you could always opt for the express option an pay 15€ more 🤷🏼‍♂️ The slowness in everydaylife annoys me sometimes but I take my time at the supermarket too 🤣

I sometimes really really hate how slow everything in Germany is. I am stuck here in Berlin because it took 6 weeks to just get an appointment at the Foreigner's Office. And now it's another 8 weeks to just print a new card. I have a Permanent Residency so its not like documents are being verified. Just changing the date and the passport number on the card. Why does it take 8 weeks? If you need to see a doctor, it can be up to a month before you get an appointment. And then to go to a specialist doctor you first may have to go to a general physician who recommends you to a specialist and that is another 4 weeks. Every single expat I know, from Russians, to Mexicans to Indians makes yearly trips home and get all their health work done. There is also a general slowness in the air. People take their sweet time in the supermarket. Even if there is a really long queue, people don't seem to have the courtesy to be a bit quicker. It is absolutely insane to me. I know this is supposed to be the European way of life but sometimes it can be really really frustrating. At some point I will leave, but I don't simply get this style of living.

@enjojoyy 95% of the vibe bros are just producing nothing. Same as the claw bros etc. The same guys also moan about prices and limits. Just ignore them.

People that run 24+ hours Codex tasks Can you share what you’re running exactly? Everyone is sharing the hours but not the task itself, I feel that most of them are just engagement baits

@justjayvi @theZachChen Or you do it the old way... w/o AI and code by yourself.

@theZachChen welcome to the Era where you pay to do work.

bruhhhhh

@TommyJgoat @CookingHub_ And they think they are doing someone a favor by giving tips.

@CookingHub_ Fucking 2 inches deep oil layer and burn the shit out of the eggs. These people are retarded

A pro cooking tip from a restaurant.

🚨 Ongoing supply chain attack on Composer packages! We just found multiple laravel-lang/* packages compromised on Packagist (lang, http-statuses, attributes). Payload runs at autoload time. At least 50 package versions were compromised. If you installed a compromised version, the malware already executed. Pin to a clean COMMIT (not version) and rotate secrets immediately. If your lockfile already had an older commit from before today, you are safe. But you should not update at the moment.

@fitchmultz Just curious: Cursor CLI is superb. Why pi? Just habbit? Or your default choice for a harness?

Composer 2.5 Fast inside Pi is superb. It’s genuinely blowing me away with how fast and high-quality it feels. Big update coming today for pi-cursor-sdk: Cursor models now have full native capabilities and seamless access to all Pi extensions & tools via the new MCP bridge. This is a serious upgrade. Stay tuned.

@cursor_ai criminally underrated. Their harness + Composer 2.5 is just on another level. Came back after trying Claude, Codex & BYOK for months… and it still feels so smooth. Who else refuses to code without Cursor now? 🔥

@badlogicgames This made me laugh more than it should, I guess.

flicker company taketh, pi.dev giveth. sound on for euromaxxing vibes.

@0xSero How kind of coder it is?

I'm stunned by how good Gemma is, I like it more than Qwen. Has it's flaws but feels at least 5x it's size. Cost to run at home: 1000$ - Macbook or Nvidia/AMD - Vision/Audio enabled - Okay agent (Qwen better) - Doesn't overthink - The MoE works with CPU offloading 8gb vram

@xiong_hui_chen Estimate for it to arrive in Token Plan?

our qwen 3.7 max is released, try our best toward agentic frontier🚀 qwen.ai/blog?id=qwen3.7 #qwen

Anyone else on the new Alibaba Token Plan and getting constantly quota/rate limit errors on Qwen3.6-Plus in pi? Not happening in droid as example.

@birdup000 @nahcrof Go then. Man tries to keep up financially.

@nahcrof Don't make me leave your service altogether you are thinking of switching things again?

a bunch of people said that keeping the $20 plan with a better deal than the larger plans would just be a bad idea so I figure I should probably make a poll about it should intermediate be 2,500 daily requests or should all the plans stay at a consistent $1/100 daily requests?

@nahcrof Doesn’t sound a good deal to get the higher plans. Then remove them if you don’t want them to be around. This will lead to people creating multiple accounts. Which is quite normal practice in OpenAI

y'know what, I'll make a compromise for the intermediate users since we just hit 2,500 followers, it feels fitting to keep the 2,500 request daily limit of the $20 plan max and scale are still being lowered to 10,000 and 5,000 respectively to be clear though

@aiCrof Any plans to introduce plan(s) where the "precision" models don't count with a mulitplier? I'm more and more settling with the "precision" variants tbh

@0xSero you as a droid user: which models you prefer for the missions: planning, worker and validation. From your perspective which are best suited for the specific roles?