Studies find that around 85 - 90% of TON Coin holders are at a loss. 

Recent data and market analysis indicate that a significant portion, around 85-90%, of TON coin holders are currently experiencing losses.

This situation is primarily attributed to a combination of factors including a 20% price drop following the arrest of Telegram's CEO Pavel Durov, which led to a massive sell-off, and broader market sentiment influenced by declining activity on the network, a death cross formation in price charts, and increasing inflation due to no cap on TON's supply.

Additionally, researchers suggest that whale investors have been reducing their holdings, contributing to bearish pressure on the coin's price.

@webthreeth

SUI Blockchain gets added to Phantom Wallet - Expanding Trading into Memecoins created on Sui Blockchain

SUI Blockchain is a layer 1 blockchain designed to offer high performance and scalability for decentralized applications (dApps). Developed by the team at Mysten Labs, which includes former engineers from Meta's Novi Research, SUI aims to provide fast transaction processing and low latency through its unique architecture.

SUI has recently been integrated into the Phantom Wallet, expanding the wallet's support beyond Solana, Bitcoin, and Ethereum. This integration allows users to manage Sui assets directly within the Phantom Wallet, which is celebrated for its user-friendly interface and secure features like staking and in-app token swaps. The addition of Sui to Phantom Wallet is a significant step towards enhancing interoperability among different blockchain ecosystems, providing users with a more seamless multichain experience.

@webthreeth

The methodology of DeepSeek created it's model from OpenAI data : Knowledge Distillation 

Knowledge distillation in artificial intelligence is a technique where a smaller, less complex "student" model learns to mimic the behavior of a larger, more sophisticated "teacher" model. The teacher model, which has been extensively trained on a dataset to achieve high accuracy, serves as a guide for the student.

This dual learning approach not only makes the student mimic the teacher's decision-making but also helps in better generalization. The benefits of this method are substantial: it enables the deployment of AI models on resource-constrained environments like mobile phones or IoT devices where the computational capacity is limited. Furthermore, it can lead to models that are not only faster and smaller but in some cases, even more accurate at making predictions due to the regularization effect of learning from soft targets.

@webthreeth

OpenAI started a new Deep Research version of Large Language Model (LLM) in it's pro version 

OpenAI has introduced "Deep Research," a new feature in ChatGPT that functions as an autonomous AI agent capable of conducting complex, multi-step research tasks. Powered by a specialized version of the upcoming OpenAI o3 model, Deep Research leverages advanced reasoning to browse, interpret, and analyze vast quantities of online content, including text, images, and PDFs.

This agent can synthesize information from hundreds of sources to generate comprehensive reports that would typically take humans many hours to compile. It aims to serve professionals in knowledge-intensive fields like finance, science, policy, and engineering, offering a significant time-saving advantage by performing tasks in just tens of minutes.

Deep Research is designed to handle complex queries and provide detailed, well-cited outputs, making it particularly valuable for tasks requiring thorough investigation and analysis. Initial access is limited to ChatGPT Pro subscribers, with plans to extend it to Plus, Team, and Enterprise users.

The feature includes capabilities such as embedding images and data visualizations in reports, and it offers detailed summaries of the research process for transparency and verification. However, OpenAI acknowledges potential limitations, including the risk of inaccuracies or misinterpretations, emphasizing that while Deep Research is a powerful tool, it should complement rather than replace human expertise and critical thinking in research.

https://openai.com/index/introducing-deep-research/

@webthreeth

Cryptocurrency Market Overview : February 10

The cryptocurrency market is experiencing a downturn today, with Bitcoin (BTC) dropping to $96,060.87, down 1.02%, and Ethereum (ETH) falling to $2,600.30, a decrease of 1.76%. Among the notable losers are Solana (SOL) at $199.35 (-1.39%), XRP at $2.3568 (-3.79%), and Cardano (ADA) at $0.669662 (-4.55%).

However, some assets are gaining; TON has increased to $3.7792 with a 0.41% gain, LEO is up to $9.8561 with a 0.13% rise, and Litecoin (LTC) has climbed to $109.267 with a significant 3.07% increase.

@webthreeth

ByteDance created an AI Video creator, designed to help creators and businesses 

ByteDance, the powerhouse behind TikTok, has ventured into AI video generation with their models 'Goku' and 'Goku+'. Goku is a flow-based generative foundation model specifically designed for the advertising domain, capable of producing ultrarealistic video content with impressive physics and movement. Building on this, Goku+ likely represents an advanced iteration, enhancing the capabilities of its predecessor by potentially offering even more sophisticated video generation techniques, aiming to revolutionize digital advertising through AI-driven, highly realistic video content.

I have seen the videos, and my God they are super impressive.
https://x.com/SebJefferies/status/1888917551825993900

@webthreeth

Anthrophic Launches It's Economic Index, which aims to showcase how AI is impacting the Economy across the Globe.

The Anthropic Economic Index is an initiative by Anthropic, an AI company, aimed at tracking and analyzing the impact of artificial intelligence on labor markets and the broader economy over time.

According to Anthrophic, the tasks and occupations with by far the largest adoption of AI on Anthropic dataset were those in the “computer and mathematical” category, which in large part covers software engineering roles. 37.2% of queries sent to Claude were in this category, covering tasks like software modification, code debugging, and network troubleshooting. These were followed by Arts & Media, and Education and Library sectors.

@webthreeth

Can AI start thinking by itself : A recent research thinks so

A recent research paper which reveals that large language models can now think in a latent space.

This means that these models can perform reasoning internally without needing to display every step in a way that's visible to users. This advancement suggests that even smaller models can achieve impressive results by using smarter design rather than just being bigger, making AI more efficient and possibly more widespread.

@webthreeth

Altcoin Market Plummets by $234 Billion: Bitcoin's Performance Creates Growing Divergence

According to Glassnode, the altcoin market has just experienced one of its most significant devaluations on record, losing $234 billion in market value.

This sharp decline underscores a widening gap between altcoins and Bitcoin, which continues to outperform its counterparts.

This divergence highlights the increasing volatility and risk within the altcoin sector, while Bitcoin maintains its relative stability and investor confidence.

@webthreeth

The second president to adopt Bitcoin, launched the first memecoin - Central African Republic

In an astonishing 48-hour period since its launch, $CAR has skyrocketed from an $80K market cap to an all-time high of $900M in just four hours, with nearly $1B in trading volume on decentralized exchanges.

The memecoin has been listed on several major exchanges including @MEXC_Official, @BingXOfficial, @BitMartExchange, @LBank_Exchange, and @XTexchange, among others, and has become the trending #1 on Solana, amassing over 27,000 holders.

Alongside its market success, $CAR has launched new initiatives in the Central African Republic, endorsed by President Faustin-Archange Touadéra, making it the first-ever country memecoin.

Looking ahead, the team plans to expand exchange listings, secure major partnerships, initiate a global marketing campaign, and collaborate with large institutions to boost credibility and impact, all while continuing to update on development initiatives in the country.

@webthreeth

CZ's Dog Named Broccoli Sparks Meme Coin Frenzy Across Blockchains

Binance co-founder and former CEO Changpeng Zhao, known as "CZ," sparked a meme coin frenzy when he revealed his dog's name, Broccoli, on X, leading to the creation and speculation around numerous "CZ's Dog" tokens on BNB Chain. Tokens like CLEO, BROWNIE, and PERRY saw multi-million dollar market caps as traders tried to predict the name. However, these tokens plummeted after the reveal.

The actual naming of his 18-month-old Belgian Malinois as Broccoli triggered a rush on BNB Chain's Four.meme launchpad, with one Broccoli token momentarily reaching a $400 million market cap.

The frenzy around Broccoli-related tokens on BNB Chain was so intense that it reportedly knocked the Four.meme platform offline. Despite the chaos, one version of the BROCCOLI token settled at around a $74 million market cap, generating over $220 million in trading volume within two hours and creating at least 11 wallets with more than $1 million profits.

@webthreeth

Perplexity AI Unveils Deep Research: A Game-Changer in AI-Powered Inquiry

Perplexity AI launched its Deep Research tool, an innovative feature aimed at transforming how users conduct research. This new tool allows users to generate expert-level insights on various topics quickly, with free access up to 5 queries per day, and up to 500 queries per day for Pro users.

Deep Research performs multiple searches, reviews hundreds of sources, and compiles comprehensive reports, scoring 93.9% accuracy on the SimpleQA benchmark and 21.1% on Humanity's Last Exam, outperforming leading models like Gemini Thinking and DeepSeek-R1. This launch positions Perplexity AI as a formidable competitor in the AI research tool market, offering a solution that's both accessible and powerful.

@webthreeth

The Libra Rug Pull: From Political Promise to Crypto Catastrophe

Libra token was launched with much fanfare, promoted by none other than the President of Argentina, Javier Milei, as a means to stimulate the country's economy.

The token's market cap soared to an astonishing $4.5 billion within hours, fueled by speculative buying and promises of economic rejuvenation. However, the excitement was short-lived as insiders, including the development team, executed what's now infamously known as the "Libra Rug Pull." They manipulated liquidity pools, adding only Libra tokens without any backing assets like USDC, which allowed them to siphon off over $100 million in mere minutes, leaving retail investors with nearly worthless tokens.

The aftermath of the Libra Rug Pull was swift and severe. The token's value plummeted by 94%, erasing billions from the market cap and causing widespread panic selling among investors.

@webthreeth

Grok III : 'The Smartest AI' expected to be released on Monday

Grok 3, the latest iteration of xAI's AI model, is poised to redefine the standards of artificial intelligence with its advanced capabilities in context understanding, nuanced response generation, and insightful assistance across various knowledge areas.

With enhancements in natural language processing and machine learning, Grok 3 is set to offer improved complex reasoning and a more intuitive user interaction. This eagerly awaited update is expected to be released on Monday.

@webthreeth

OPENAI Roadmap : A single unified GPT-5 Model

The OpenAI roadmap outlines a progression towards a single unified GPT-5 model, as depicted in the image.

Starting with GPT-4.0, the roadmap shows an incremental increase in performance over time. GPT-4.5 is marked as the last non-reasoning model, indicating a shift in focus towards more advanced capabilities. The roadmap then progresses to O1 and O3, with O3 noted as not being released as a standalone model. The culmination is the GPT-5 model, which comes in three versions: Free, Plus, and Pro, integrating all tools including voice, search, canvas, and deep research.

This roadmap suggests a strategic development path aimed at enhancing performance and integrating comprehensive functionalities into a unified model.

@webthreeth

The high level of competition in AI is increasing the level of technology and disruption in AI.
https://www.linkedin.com/pulse/high-level-competition-ai-increasing-technology-tomas-gebremariam-cxchc?utm_source=share&utm_medium=member_android&utm_campaign=share_via

Deep Research to expand it's capabilities over the coming weeks?

OpenAI has recently introduced "Deep Research," an AI-powered tool within ChatGPT designed to enhance research capabilities by synthesizing information from multiple online sources.

According to recent announcements, over the coming weeks, OpenAI plans to expand Deep Research's features to include data visualizations and graphs. This update aims to provide users with a more comprehensive analysis, moving beyond text-only outputs to incorporate visual elements that can enhance understanding and presentation of complex data.

The integration of data and graphs is expected to make the tool even more useful for generating in-depth research reports, offering users not only textual analysis but also visual representations of data trends and patterns.

@webthreeth

Grok 3 : Comparison in Science and Math to other GPT Models

At the launch of Grok 3 Event, a detailed benchmark comparison was presented, showcasing the performance of various AI models across different domains.

The chart highlighted the capabilities of Grok-3, Grok-3 mini, Gemini Pro, DeepSeek-V3, Claude 3.5 Sonnet, and GPT-4o in Math (AIME '24), Science (GPQA), and Coding (LCB Oct-Feb). Grok-3 demonstrated a strong performance, particularly in Science with a score of 75, while Grok-3 mini showed a balanced performance across all three categories.

This visual representation provided a clear insight into how Grok-3 stacks up against its competitors in these specialized areas, emphasizing its advanced capabilities in the AI landscape.

@webthreeth

Reasoning + Test-Time Compute : Grok-3 vs Other Models

During the launch of Grok 3 Event, a comprehensive comparison of various AI models' reasoning capabilities was showcased through the "Reasoning + Test-Time Compute" benchmark chart.

This chart illustrated the performance of Grok-3 Reasoning Beta, Grok-3 mini Reasoning, o3mini(high), o1, Deepseek-R1, and Gemini-2 Flash Thinking across three domains: Math (AIME'24), Science (GPQA), and Coding (LCB Oct-Feb).

Grok-3 Reasoning Beta led with impressive scores of 96 in Math, 85 in Science, and 79 in Coding, demonstrating its superior reasoning capabilities. The comparison highlighted the advancements in AI reasoning and computational efficiency, setting a new benchmark in the field.

@webthreeth

SKYworkAI released the best video generation AI model to date.

SkyReels V1 540p, developed by SkyworkAI, stands out from other video generation models due to its superior performance across various metrics.

According to the comparison data, SkyReels V1 achieves the highest overall score of 82.43, surpassing models like VideoCrafter-2.0 VEnhancer and CogVideoX1.5-5B.

Finetuned on HunyuanVideo with 10M film and TV clips, the T2V model delivers advanced facial animation, capturing 33 distinct expressions with over 400 natural movement combinations.

@webthreeth