MAROKO133 Update ai: The Most Popular Streamer on Twitch Is Now an AI Construct Terbaru 20

📌 MAROKO133 Breaking ai: The Most Popular Streamer on Twitch Is Now an AI Construc

Amidst mostly unfounded fears that artificial intelligence is coming for our jobs, it turns out there’s one gig AI is pretty excellent at: streaming on Twitch.

Meet Neuro-sama, a cutesy AI-powered character which is now the top streamer on Twitch by active subscriber-count, according to Twitch Tracker. At the time of writing, Nuero-sama — streaming via the Twitch channel Vedal987 — has 165,268 paid active subscribers, well above the second-highest channel, the human-ran Jynxzi.

On Twitch, each paid subscriber is valued at $5 a month, some of which is split with Twitch. According to an estimate by Dextero, that means Neuro-sama is raking in at least $400,000 a month just from subscriptions — and that’s on top of random viewer donations, sponsorships, and ad revenue.

Over the past few years, content creators known as VTubers have taken Twitch by storm. VTubers are essentially like regular streamers, except they use digital motion-capture effects to alter their voice and appearance, often taking on cartoon-like personas to stand out and appeal to viewers looking for something beyond the physical realm.

Neuro-sama takes the concept one step beyond. Despite appearances, the avatar isn’t merely puppeted by a human streamer, but is entirely AI-generated. The channel’s owner, identified only by the handle Vedal, has been running the channel as his full-time job since at least 2023, according to Bloomberg.

Per the publication’s reporting, the avatar’s speech is first generated by a large language model (LLM) before being converted to audio via another text-to-speech AI application.

Though the AI avatar was originally designed to live-stream the rhythm game Osu, it has since developed a life of its own, discussing random topics with various subscribers who communicate with the LLM via Twitch’s baked-in live chat features.

“Part of it is definitely the novelty,” Vedal told Bloomberg of the character’s popularity at the time. “Part of it is some of the things she says. She can be quite chaotic, saying things humans wouldn’t really say.”

Though Neuro-sama sits atop the Twitch throne at the moment, there is still a ways to go if the channel is ever to become the all-time subscription record holder: it sits at number eight on that list — still an accomplishment that was once unthinkable.

More on AI: Doctors Say AI Use Is Almost Certainly Linked to Developing Psychosis

The post The Most Popular Streamer on Twitch Is Now an AI Construct appeared first on Futurism.

🔗 Sumber: futurism.com

📌 MAROKO133 Update ai: MIT Researchers Unveil “SEAL”: A New Step Towards Self-Impr

The concept of AI self-improvement has been a hot topic in recent research circles, with a flurry of papers emerging and prominent figures like OpenAI CEO Sam Altman weighing in on the future of self-evolving intelligent systems. Now, a new paper from MIT, titled “Self-Adapting Language Models,” introduces SEAL (Self-Adapting LLMs), a novel framework that allows large language models (LLMs) to update their own weights. This development is seen as another significant step towards the realization of truly self-evolving AI.

The research paper, published yesterday, has already ignited considerable discussion, including on Hacker News. SEAL proposes a method where an LLM can generate its own training data through “self-editing” and subsequently update its weights based on new inputs. Crucially, this self-editing process is learned via reinforcement learning, with the reward mechanism tied to the updated model’s downstream performance.

The timing of this paper is particularly notable given the recent surge in interest surrounding AI self-evolution. Earlier this month, several other research efforts garnered attention, including Sakana AI and the University of British Columbia’s “Darwin-Gödel Machine (DGM),” CMU’s “Self-Rewarding Training (SRT),” Shanghai Jiao Tong University’s “MM-UPT” framework for continuous self-improvement in multimodal large models, and the “UI-Genie” self-improvement framework from The Chinese University of Hong Kong in collaboration with vivo.

Adding to the buzz, OpenAI CEO Sam Altman recently shared his vision of a future with self-improving AI and robots in his blog post, “The Gentle Singularity.” He posited that while the initial millions of humanoid robots would need traditional manufacturing, they would then be able to “operate the entire supply chain to build more robots, which can in turn build more chip fabrication facilities, data centers, and so on.” This was quickly followed by a tweet from @VraserX, claiming an OpenAI insider revealed the company was already running recursively self-improving AI internally, a claim that sparked widespread debate about its veracity.

Regardless of the specifics of internal OpenAI developments, the MIT paper on SEAL provides concrete evidence of AI’s progression towards self-evolution.

Understanding SEAL: Self-Adapting Language Models

The core idea behind SEAL is to enable language models to improve themselves when encountering new data by generating their own synthetic data and optimizing their parameters through self-editing. The model’s training objective is to directly generate these self-edits (SEs) using data provided within the model’s context.

The generation of these self-edits is learned through reinforcement learning. The model is rewarded when the generated self-edits, once applied, lead to improved performance on the target task. Therefore, SEAL can be conceptualized as an algorithm with two nested loops: an outer reinforcement learning (RL) loop that optimizes the generation of self-edits, and an inner update loop that uses the generated self-edits to update the model via gradient descent.

This method can be viewed as an instance of meta-learning, where the focus is on how to generate effective self-edits in a meta-learning fashion.

A General Framework

SEAL operates on a single task instance (C,τ), where C is context information relevant to the task, and τ defines the downstream evaluation for assessing the model’s adaptation. For example, in a knowledge integration task, C might be a passage to be integrated into the model’s internal knowledge, and τ a set of questions about that passage.

Given C, the model generates a self-edit SE, which then updates its parameters through supervised fine-tuning: θ′←SFT(θ,SE). Reinforcement learning is used to optimize this self-edit generation: the model performs an action (generates SE), receives a reward r based on LMθ′’s performance on τ, and updates its policy to maximize the expected reward.

The researchers found that traditional online policy methods like GRPO and PPO led to unstable training. They ultimately opted for ReST^EM, a simpler, filtering-based behavioral cloning approach from a DeepMind paper. This method can be viewed as an Expectation-Maximization (EM) process, where the E-step samples candidate outputs from the current model policy, and the M-step reinforces only those samples that yield a positive reward through supervised fine-tuning.

The paper also notes that while the current implementation uses a single model to generate and learn from self-edits, these roles could be separated in a “teacher-student” setup.

Instantiating SEAL in Specific Domains

The MIT team instantiated SEAL in two specific domains: knowledge integration and few-shot learning.

• Knowledge Integration: The goal here is to effectively integrate information from articles into the model’s weights.
• Few-Shot Learning: This involves the model adapting to new tasks with very few examples.

Experimental Results

The experimental results for both few-shot learning and knowledge integration demonstrate the effectiveness of the SEAL framework.

In few-shot learning, using a Llama-3.2-1B-Instruct model, SEAL significantly improved adaptation success rates, achieving 72.5% compared to 20% for models using basic self-edits without RL training, and 0% without adaptation. While still below “Oracle TTT” (an idealized baseline), this indicates substantial progress.

For knowledge integration, using a larger Qwen2.5-7B model to integrate new facts from SQuAD articles, SEAL consistently outperformed baseline methods. Training with synthetically generated data from the base Qwen-2.5-7B model already showed notable improvements, and subsequent reinforcement learning further boosted performance. The accuracy also showed rapid improvement over external RL iterations, often surpassing setups using GPT-4.1 generated data within just two iterations.

Qualitative examples from the paper illustrate how reinforcement learning leads to the generation of more detailed self-edits, resulting in improved performance.

While promising, the researchers also acknowledge some limitations of the SEAL framework, including aspects related to catastrophic forgetting, computational overhead, and context-dependent evaluation. These are discussed in detail in the original paper.

Original Paper: https://arxiv.org/pdf/2506.10943

Project Site: https://jyopari.github.io/posts/seal

Github Repo: https://github.com/Continual-Intelligence/SEAL

The post MIT Researchers Unveil “SEAL”: A New Step Towards Self-Improving AI first appeared on Synced.

🔗 Sumber: syncedreview.com