file
md5: 1953bdf987809282a2cfe3e22ddeada6
🔍
>GPT-5
>Claude 4.1
>Gemini 3
>Gemini Diffusion
Which upcoming model are you most excited about?
For me it has to be GPT-5. It will be a game changer.
2 more weeks until the singularity, right?
>>106141222 (OP)>Gemini Diffusionthis one because I'm curious about it
Alexandr Wang, the former CEO of Scale AI, will serve as Meta’s new chief A.I. officer and lead the MSL team. The 28-year-old recently joined Zuckerberg’s efforts after Meta invested more than $14 billion in Scale AI, which specializes in labelling data for A.I. systems. Wang will co-lead the new group with Nat Friedman, the former head of Github, said Wang in a post on X yesterday (June 30), which also included a roster of other new hires joining the team.
“Towards superintelligence,” noted Wang, who said he’s “thrilled to be accompanied by an incredible group of people.”
The division will include former Google DeepMind researchers Jack Rae and Pei Sun, alongside Anthropic’s Joel Pobar and Sesame AI’s Johan Schalkwyk. One notable absence from the list of hires shared by Wang is Daniel Gross, a co-founder of startup Safe Superintelligence and an investment partner of Friedman. He’s also expected to join Meta’s new A.I. efforts, according to CNBC.
The largest tranche of Meta’s new hires hail from OpenAI. In addition to nabbing researchers like Ji Lin, Hongyu Ren and Jiahui Yu, who contributed to releases like o3 and GPT-4o, MSL includes Shengjia Zhao, a co-creator of ChatGPT, and Trapit Bansal, a top reinforcement learning researcher. Other notable additions include Suchao Bi and Huiwen Chang, who helped build GPT-4o’s voice mode and image generation features, respectively.
>>106141222 (OP)Literally a Black Mirror episode plot
>>106141222 (OP)>Which upcoming model are you most excited about?>For me it has to be GPT-5. It will be a game changer.
I am not an AI illiterate retarded.
While normal-fags are focused on bubble AI we are working on next-generation AI
https://github.com/sapientinc/HRM
and there are around many other projects and papers that could burst the bubble. Sadly investors are to imbecile to read them and be aware.
>>106142506That kind of argument is stupid it’s just argument from ignorance mixed with anecdotal reasoning and confirmation bias because neither side explains it using any X or Y model, and it’s easy to imagine nonsense.
Autoregressive LLMs based on transformers will NEVER reach AGI because the architecture is not designed for that. They scale in their ability to integrate data over semantically compatible patterns, but it is impossible for them to solve logical problems without the aid of other models and tools that are NOT LLMs.
Next models won't even use tokenization neither meme-<think> approach.
