MAROKO133 Eksklusif ai: Investors Flocking to Super-Anonymous Cryptocurrency Used for the

📌 MAROKO133 Update ai: Investors Flocking to Super-Anonymous Cryptocurrency Used f

When Bitcoin first launched in 2009, it promised its users a decentralized digital currency that would be free from the prying eyes of financial authorities. That was possible thanks to the blockchain, a digital stone slab where every transaction could be etched into the collective public record.

Of course, Bitcoin and its look-alikes quickly descended into a haven for grifters, drug sales, and money-laundering, enabling anonymous criminals the world over. Then a few confusing things happened. First, it quickly turned out that Bitcoin wasn’t quite as anonymous as initially advertised, meaning crooks had to take extra steps to launder their ill-gotten gains. And worse, at least for crypto’s most dedicated libertarian acolytes: it also became a legitimate financial vehicle, tracked on the Bloomberg terminal, trafficked by publicly-traded corporations, and increasingly scrutinized by government watchdogs.

Bitcoin going legitimate was evidently a bridge too far. According to new reporting by the Wall Street Journal, cryptobros are now ditching the old currencies in favor of an even more anonymous vehicle: Zcash.

Zcash isn’t new, but the fervor around it is. Launched in October of 2016, the Zcash blockchain protocol was designed by researchers at institutions like Johns Hopkins University and Tel Aviv University to basically add another layer of privacy to the Bitcoin protocol, theoretically making transactions “untraceable.” It’s one of numerous privacy tokens in use these days, meaning cryptocurrencies with relatively small market caps that use advanced cryptography to almost completely obscure their users from sight.

Zcash’s closest rival, Monero, launched in 2014 and has become the preferred privacy token among both privacy-obsessed cryptobros and, increasingly, criminals seeking to conceal everything from ransomware payments and international transactions to drug deals on darknet markets.

That’s not stopping some household name fintech guys from getting in on the fun. According to the WSJ, two of Bitcoin’s biggest original backers, Tyler and Cameron Winklevoss, are investing $50 million into a crypto startup called Cypherpunk Technologies, which will hold gobs of Zcash as part of its portfolio. So far, the Winklevii — who just injected $100 million of their own Bitcoin holdings to prop up their beleaguered legacy crypto company, Gemini — have stockpiled more than 300,000 Zcash tokens that we know of, with a value of over $157 million at the time of writing.

Accordingly, the price of Zcash soared over the past month, while trading volume is consistently the highest it’s been in three years after a massive spike in transactions late in 2025 — from around 2,000 a day prior to October to over 70,000 in mid-November.

As cofounder of venture capital firm Multicoin Capital Tushar Jain told the WSJ, “Zcash is what Bitcoin should be. It’s what Bitcoin was originally meant to be.”

Zcash comes with plenty of downsides. There’s the crime thing, naturally; some cybercriminals are already accepting Zcash as payment options, while rival privacy token Monero has become the new bullion in those circles. This is probably why at least 10 nations have heavily restricted privacy tokens on regulated exchanges, with countries like Japan, South Korea, and India deploying outright bans.

Beyond cybercrime, the influx of institutional capital into privacy tokens also points to a telling shift among certain players in the finance industry.

The more legitimate money that flows through firms like Cypherpunk into privacy tokens, the more capital overall can hide from tax authorities and regulators, at a time when the ultra-rich are racing to secure their fortunes before the proverbial music stops. Similar to the blockchain currencies that came before, tokens like Monero and Zcash represent a far-right political project which asserts that freedom from democratic financial scrutiny is somehow a human right — a step backward from the already de facto unregulated system of corporate banking.

By design, these same currencies have become indispensable tools for organized crime — revealing that the libertarian fantasy of “financial freedom” amounts to freedom for the wealthy to hide their assets, and freedom for criminals to operate without accountability. Indeed, it’s becoming nearly impossible to tell them apart.

More on crypto: Bitcoin Developers Are Debating a Move That Could Send Crypto Markets Into a Tailspin

The post Investors Flocking to Super-Anonymous Cryptocurrency Used for the Sketchiest Stuff Imaginable appeared first on Futurism.

🔗 Sumber: futurism.com

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

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