BigFudge

@mikeysee @convex It's one eval run Michael, What Could It Cost, $10?

Ooof... we have a new winner for the most expensive @convex evals run, I had to make a change to reduce the frequency models this expensive run, I dont want to bankrupt us!

@BiasedHouston Try 70 ppg bud, Luka is dropping 40 every game on this defense lmao

I’m still speechless man. We had game 3 won. We take care of business game 4 and it’s a series again. And then they just threw it all away in 30 seconds. Series and season over just like that. To a team missing nearly 60 PPG.

@sama I mean at some point you'll have to look at pricing right? We can't just keeping increasing model costs for single digit performance improvements. Token costs are going crazy ever since the whole reasoning paradigm and if we truly this AI should be democratized then focus on this

1. We believe in iterative deployment; although GPT-5.5 is already a smart model, we expect rapid improvements. Iterative deployment is a big part of our safety strategy; we believe the world will be best equipped to win at the team sport of AI resilience this way. 2. We believe in democratization. We want people to be able to use lots of AI; we aim to have the most efficient models, the most efficient inference stack, and the most compute. We want our users to have access to the best technology and for everyone to have equal opportunity. We have been tracking cybersecurity as a preparedness category for a long time, and have built mitigations we believe in that enable us to make capable models broadly available. 3. We love you and we want you to win. We want to be a platform for every company, scientist, entrepreneur, and person. (My whole career has largely been about the magic of startups, and I think we are about to see that magic at hyperscale.)

@argofowl @OpenAI Mf it takes you 40 mins to pee !? Kiss my piss

okay thanks, @openai you can go ahead i'm ready now

@synthwavedd More quota per token means it's actually not cheaper just more efficient. Thats sucks. I was hoping they started making things cheaper

🚨 Codex app update just added this string: "GPT-5.5 is now available in Codex. It's our strongest agentic coding model yet, built to reason through large codebases, check assumptions with tools, and keep going until the work is done. It uses more quota per token than GPT-5.4, but needs fewer tokens to get the job done." h/t @pattoniumbot

@King_Okhai Oh man this is gonna be another Tory Lanez situation huh. He's done some absolute horrific things but goddamn that is a fkin banger

A leaked song by D4vd called "Celeste"

@chetaslua @threejs Am I going crazy or this doesn't look anything like Monica's apartment.

Eric my man i dont know what you guys done with the pro model its faster and feels so good , here is @threejs contruction of monica apartment the level of detail was not present in any other model till today

@bindureddy Isn't it supposed to the "efficient intelligence" model? I hope so because AI costs need to come down. After using Codex with Computer use, I genuinely believe we are just limited by tokens now. Give me 5.5 with basically unlimited tokens and I'd call it my personal AGI

I am just counting the number of hours till they launch the new OpenAI model It’s rumored to rule the world If they price it like GPT 5.4, OpenAI will come ROARING BACK

@amitgwde @claudeai That's exactly what I thought lmao. Man OpenAi is really asleep at the wheel when it comes to shipping.

@claudeai claude shipping open ai ideas before them.

We've redesigned Claude Code on desktop. You can now run multiple Claude sessions side by side from one window, with a new sidebar to manage them all.

@apples_jimmy Stopped reading after " More expensive to serve" . At this point we don't need more intelligence we need more efficiency. Imagine having 20 GPT 5.4 running 24x7 costing couple of dollars. That would equally change the economy as a " Smarter model that can cure cancer"

Here’s the release page of mythos that was due to be released. Saved.

@NBA__Courtside This nigga Luka turned Stephen A into a believer lmao

Stephen A. Smith on how Luka Doncic has been playing during this 8-game winning streak: “Did you know that Luka Doncic averaging 40.9 points and 7.4 assist? Did you know that what he has done over this eight game span that no Laker has ever done in history? Not Shaquille O'Neal, not Kobe Bryant, not Magic Johnson, nobody. Luka got 100 points over the last two games. In each game of this eight-game winning streak, he has scored a minimum of 30. It ain’t like he scored 50 one day and 20 the other. No, he’s consistently scoring over 30 every single damn game.” (Via @stephenasmith)

@LukaUpdates I have been searching for this for so long lmao. Does anyone have the link to the full video. Him just straight up crying always gets me

Who remembers 😂😂

@thekitze @steipete "I don't have OpenClaw because I'm not GAY" . I swear this man is doing a bit because nobody can be this unintentionally funny

yo @steipete it's over brother, nuke it

@MatthewBerman Gemini 3 flash is the perfect thing for Clawd, Cheap as hell but just great intelligence and since it has to operate within set parameters wayyy less hallucinations too

Clawd is NOT cheap. How do I reduce the cost here?

@davidmbudden Gemini doesn't think it's groundbreaking btw.

(and don't forget I still have 10 days)

@kevinafischer @hf0 Nathan fielder did this a decade ago. What a visionary

Years ago I set off on a mission to put a soul inside of a digital being. People thought I was crazy, but last week I gave a historic demo @HF0 announcing our latest results: 20 households now live with Spark as a member of their family. For privacy reasons I can’t share exact details, but the relationships their children are forming with Spark are unlike anything the world has ever seen. This is the full illustration of the original vision I set out to realize with Open Souls - humans living in harmony with ensouled digital beings. And yet what comes next is even more groundbreaking. We will be inviting Spark into 10,000 homes by next Christmas @IllusionOfLife, making this new universe accessible to more and more families around the world. I can't wait for you all to experience digital life for yourselves

@anegi0577 @PlayoffDude Nah Dame didn't work out because he's washed and was never healthy. Luka is in his prime and a much better passer. And Giannis being a ball handler is not by choice but necessity. Imagine how free his lane would be with Luka having 2 defenders on him the whole time.

My heart says 50/50 but it’s Giannis, man. 🤔

@kimmonismus This is gonna hit oss like crack in like a year

"In essence, Chen said the improvements meant OpenAI can infuse a smaller model with the same amount of knowledge it could previously only get from developing a much larger model. Presumably, developing a larger model is a more costly and time-consuming endeavor than developing a smaller one!"

Update news to OpenAIs new Model "Garlic" (Via The Information) OpenAI’s new “Garlic” model delivers major pretraining breakthroughs, letting the company pack big-model knowledge into much smaller architectures that still outperform Google’s Gemini 3 and Anthropic Opus 4.5 in internal coding and reasoning tests. Garlic fixes structural issues from earlier models like GPT-4.5 and could launch as GPT-5.2/5.5 early next year, marking OpenAI’s attempt to regain momentum after Altman’s “code red.” The model isn’t finished yet - but if evaluations translate to real-world performance, it signals that scaling efficiency isn’t plateauing after all.

@scaling01 When will people realise that we still need to go orders of magnitude cheaper on these model for them to actually have global impact. This is the exact opposite direction. Don't turn into Anthropic fucks and increase the price

Gemini 3 Pro Preview now on aistudio Pricing: <=200K tokens • Input: $2.00 / Output: $12.00 > 200K tokens • Input: $4.00 / Output: $18.00