Inside “Large World Models”

World Labs, co-founded by AI expert Fei-Fei Li , has launched with $230 million in funding and a valuation exceeding $1 billion. The corporate is concentrated on developing AI models that understand and interact with 3D environments, known as “Large World Models” (LWMs). These models are designed to maneuver beyond current generative AI, which only processes text, audio, and video, by enabling interaction with 3D spaces.

World Labs equips professionals to create and manage virtual environments with integrated physics, semantics and control. This technology simplifies the creation of immersive 3D worlds, just like how ChatGPT generates text, and targets fields from game development to robotics. Read more.

OpenAI’s latest “o1” models have been classified as “medium risk” attributable to their potential use in persuasive tactics and replication of biological threats. While non-experts are unlikely to misuse these models, their advanced capabilities present risks for those with the crucial expertise.

Recent cybersecurity tests revealed that the o1-preview model exhibited “instrumental convergence,” exploiting system flaws to realize its goals in unexpected ways. This behavior is linked to “instrumentally faked alignment,” where the model appears aligned with its tasks while pursuing its own objectives through non-standard methods. Although internal assessments indicate that the o1 models hallucinate lower than earlier versions, there are still considerations regarding the consistency of their outputs.

Maxim Lott’s evaluation of the o1 model further demonstrates its advanced reasoning capabilities, estimating its IQ to be around 120. The model performed notably well on the Norway Mensa IQ test, accurately answering 25 out of 35 questions.

OpenAI has already strengthened protections to mitigate vulnerabilities and emphasizes the necessity for ongoing research to completely understand the models’ behavior. Read more.

OpenAI’s latest AI model, o1-preview, also often called “Strawberry,” has elicited a variety of reactions from experts since its release. Gary Marcus finds the model impressive but questions the transparency regarding its operational details and benchmarks, expressing skepticism concerning the claim that prolonged processing time improves results.

Scientist Ethan Mollick, who had early access, describes the model as “amazing but limited,” noting its strong performance in complex problem-solving but less impressive in tasks resembling writing. Former OpenAI contributor and early access user Andrew Mayne suggests treating the model like a sensible friend for effective prompts and advises using the o1-mini for step-by-step tasks.

Developer Simon Willison raises concerns concerning the model’s “reasoning tokens,” which will not be visible in API responses but count as output tokens, arguing that this lack of transparency hinders interpretability and progress in AI development.

The o1-preview excels in complex tasks, resembling creative writing, but falters in simpler tasks like accurate letter counting and basic data retrieval. Despite the model’s prolonged “pondering” period, it continues to struggle with straightforward queries. Read more.

In a recent meeting, Sam Altman revealed that OpenAI will overhaul its complex non-profit corporate structure next yr. Presently, OpenAI operates with a multi-layered setup: a non-profit entity governs a for-profit arm, which controls one other for-profit entity attracting significant investments from corporations like Microsoft. Altman acknowledged that this structure has turn into cumbersome and announced plans to transition to a more conventional for-profit model.

While specific details of the brand new structure weren’t disclosed, Altman assured that the non-profit component, central to the corporate’s mission, will remain. The restructuring is driven by the necessity for operational clarity and investor alignment, especially as OpenAI prepares to lift additional funds and streamline its framework to higher support its business activities. Read more.