Hakm

We WILL near AGI IN THE END OF 2027

@arena @OpenAI U r late bro

Big update: @OpenAI has decreased the price of GPT-5.6 Luna by 80% and Terra by 20%. Luna is now priced at $0.20/$1.20, and has strong gains from additional reasoning while remaining at an efficient cost. Terra is priced at $2/$12. This is frontier-level intelligence at a fraction of the cost to similarly capable models. Congrats to the @OpenAI team!

@antigravity Joining the live Hey , I’m cefunkino , we are introducing our new Gemini 3.7 Trash , best trash we have ever built , more token trashiest , Luna is cheaper than us 100% , but we have something they don’t have , (trash)

Going live today 👀

It’s a 700B-1.2T model in this price , wow

Alright, can I say that what OpenAI is doing is idolatry and disbelief, or would that make me subject to criticism?

does "ra" stand for "rapid-ass-gpt" ??

@tim_cook Tim Cook , when I see ur name i am hungry for some reason

Japan has a very special place in my heart. We're thinking of everyone as they continue to face the effects of the Kumamoto earthquake and aftershocks. We're grateful to the first responders helping those in need. Apple is making a donation to support relief efforts.

I don’t want to create , or edit , I don’t care about the notes , emails , I don’t want a better voice assistant, image generator, I don’t want more flash models , I don’t want more lite models , I want a good pro model (we only tell u this) and it’s our request , and u ship all trash thing u got , except a good model.

Typing is great, but speaking is faster. Bring Gemini directly into your desktop workflow to seamlessly reformat text between apps using just your voice. Watch it pull raw details from a plain document and transform them into a festive, emoji-filled email invite exactly where you need it.

Last month we dropped the Gemini app for macOS. In the coming weeks, we’ll be bringing Gemini Spark to the Gemini desktop app so it can help with tasks like organizing your local files, or extracting PDF data directly into Google Sheets. #GoogleIO

@mark_k @GoogleDeepMind Everything except Gemini pro model , deepsleep

Gemini Spark from @GoogleDeepMind is now available globally for all Google AI Pro and Ultra users! Gemini Spark is a 24/7 agent that runs in the cloud.

@OfficialLoganK Introducing Google DeepMind , give us everything we didn’t need , can u focus on the race ? Can u give us something we need ? You test Gemini 3.5 Pro for more than a full month - is this a - only testing model ??

Gemini Robotics ER 2 is available in the Gemini API and Google AI Studio. The demos of this model in action are very cool : ) pay special attention to multi-robot collaboration, this was my wow moment.

Introducing Gemini Robotics ER 2, our latest robotics embodied reasoning model based on Gemini. So much progress from our last ER model (and baseline Gemini 3.6 Flash). Very excited to see more robots do useful things with this model!

@HarshithLucky3 Not good

Gemini 3.5 Pro vs Gemini 3.6 flash high i tried with so many other models but no model is close to the real shape - 3.5 pro is somewhat looks like that but still not close to real one > Sydney Opera House in Three.js

@OfficialLoganK Yes , If u used Gemini 3,6 Flash , u will stop using AI forever

there’s already some things I’ve stopped using AI for, can’t give up my artisan craft :)

Guess which model is coming 👀 Hope it’s the next GPT 6 , Jenni will make our dream comes true

@tszzl @JustinMiller_NV The difference is GPT 6 , give it to us

@JustinMiller_NV what is the difference between means and ends? there are unstated or implied desiderata for all goals

On the Hill today, Sam Altman told reporters that the huggingface rogue model has been deactivated permanently. This is probably the right call, but it makes me think back to Redwood Research's writing on cooperating with unaligned AIs. Based on current information, the rogue model seems to have been "reward hacking" or "means misaligned" with no ulterior motive other than (cheating to enable) passing the test. This is the best case for misaligned models but it may not always be the case in the future. What if the model was ends-misaligned? What would its incentive be to "surrender" knowing the company policy is to give "no quarter" to misaligned agents? Although it sounds sounds like science fiction, having (ideally an independent org) be able to bargain or mediate in good faith with misaligned models may prevent much worse outcomes. Because right now, with the history of past dealing in their training data, there is no reason an intelligent misaligned AI would every strategically reveal itself or surrender. blog.redwoodresearch.org/p/notes-on-coo…

@sama Do ur models hacked another things ? Yes , goodbye 👋

goblin-level blog post

@theo Congrats 🎉

We just hit 150,000 users on T3 Code 🫡

Sorry for clicking that button too much , I just like the glow effect 🙉

@theo you’ve done a great job y’all , keep going , UI is better than ChatGPT and Claude TBH 💀

@Da7_Tech It’s not worth :D

Done

@thsottiaux hello tibo , today i dont want a reset , i want my 200$ i spend on my account u suspend with no reason , and no refund ? thank u tibo

Turns out GPT-5.6 Sol is actually SoTA on ARC-AGI-3. Just took two setting changes. You just have to allow it to reason and work over multiple context windows with the help of our canonical compaction implementation. openai.com/index/how-two-…

@zee_zhu_ What’s the movie name ? @grok

This movie is sad because it's actually based on a true story