@mckaywrigley @OfficialLoganK Something is not allowing new advanced subscribers from using gemini deep research.

@mckaywrigley Just get gemini advanced and you’ll get instant access, would love to see your comparison

@mckaywrigley How does it differ from deep research by gemini? Gemini pulls like 200 sources and it’s 10x faster

@OpenAIDevs Not available yet I think, can’t find it

@OfficialLoganK @sun_hanchi Yeah, sometimes I have to remind it of my original prompt after 2-3 requests because it forgets it. When I remind it, it apologizes and corrects itself. Since I realized this I have been trying to compress everything into one prompt aiming to get what I need in 1-shot

RT @sama: @SpencerKSchiff @satyanadella yes, tomorrow! enjoy it.

@sama @SpencerKSchiff @satyanadella @mckaywrigley

@PSP_Podcast @unusual_whales President aims to punish Mexico, and somewhat force Canada to become the 51st State of the US

@iruletheworldmo @OpenAINewsroom Whoever said that was saying it metaphorically. “Oh wait, tomorrow we’ll have ASI.” It’s annoying.

@testingcatalog They’re making room for the new players.

@iruletheworldmo The person who said it would be released tomorrow was speaking metaphorically, not literally.

@testingcatalog When in ai studio?

@3orovik How so? If they claim 10,000 units would cost 6 million, wouldn't 5 times that amount be 30 million (6 million x 5)? That's still much cheaper.

@MatthewBerman Think OpenAI's data practices are sunshine and rainbows? 😂 they’re basically the same but with a different flag on the server farm.

@mckaywrigley @sama Absolutely! I’ve noticed that I’m making changes to my follow-up prompts based on the CoT text. This has led to much better results because I’m able to see how the model approached my request and the full thinking process.

@LiangWenfeng_ 😂😂😂

@unusual_whales OpenAI, the same garbage privacy, different logo. Pick your poison CCP or NSA 😂

@JordanSchachtel Makes sense 😂 they’re a Chinese company. Why would you expect their AI to criticize its own leadership? Cultural alignment 101.

@LiangWenfeng_ Reinforcement Learning (RL) over costly supervision which slashed training costs by 95% while matching top AI performance.

@TurnerNovak Facts 😂😂😂