ChainCatcher news, according to Cointelegraph, American tech giant Microsoft plans to invest $3 billion over the next two years to expand its cloud computing and artificial intelligence capabilities in India. This investment includes a partnership with the Indian government on an AI-focused initiative called INDIAai, which aims to provide AI training for 500,000 people and establish new data centers.Additionally, Microsoft has signed strategic partnerships with RailTel, Apollo Hospitals, Bajaj Finserv, Mahindra Group, and upGrad. Microsoft will also provide benefits from the Founders Hub program to up to 1,000 AI startups in India, such as Azure credits, business resources, and mentorship.

ChainCatcher news, according to TechCrunch, Elon Musk stated during a live conversation with Stagwell Chairman Mark Penn that the training of current AI models has largely exhausted real-world data, "We have exhausted the cumulative sum of human knowledge, which happened last year." Musk's views align with those of former OpenAI Chief Scientist Ilya Sutskever, who suggested at the NeurIPS machine learning conference that the AI industry has reached a "data peak," and that the way models are developed may need to change in the future.Musk believes that synthetic data will be a way to supplement real data, and AI will achieve self-learning through generating and self-evaluating data. This trend has been adopted by tech giants including Microsoft, Meta, OpenAI, and Anthropic, with models like Microsoft's Phi-4 and Google's Gemma combining real and synthetic data for training. Gartner predicts that by 2024, about 60% of data in AI and analytics projects will be synthetically generated.The advantages of synthetic data include cost savings; for example, the AI startup Writer spent only about $700,000 to develop its nearly entirely synthetic data-based Palmyra X 004 model, whereas the development cost for a similarly scaled OpenAI model is about $4.6 million. However, synthetic data also carries risks, including a decrease in model creativity, increased output bias, and potential model collapse, especially when the training data itself is biased, which can also affect the generated results.

ChainCatcher message, DWF Labs stated that they are training an AI agent and have released a link for public access.

ChainCatcher news, Musk posted on the X platform that Grok 3 is coming soon, the pre-training is complete, and the computational power is 10 times that of Grok 2.

ChainCatcher news, according to Chainwire, the decentralized AI training network Nodepay has announced the completion of its second round of financing, totaling $7 million. The latest round of investors includes IDG Capital, Mythos, Elevate Ventures, IBC, Optic Capital, Funders.VC, Etherscan founder Matthew Tan, and CoinHako co-founder and CEO Yusho Liu. Previous investors in Nodepay's financing include Animoca Brands, Mirana, OKX Ventures, JUMP Crypto, and Tokenbay Capital.According to reports, Nodepay is a decentralized AI training network dedicated to democratizing AI training through real-time data retrieval. By converting idle internet bandwidth into resources, Nodepay powers the next generation of AI models.

ChainCatcher news, the decentralized AI training platform FLock announced the completion of a $3 million strategic round of financing, bringing the total financing to $11 million. This round of financing was led by Grayscale's parent company, Digital Currency Group (DCG), with Lightspeed Faction participating in a follow-on round, along with strategic partners such as Animoca Brands, Fenbushi Capital, GnosisVC, GSR Ventures, OKCoinJapan, Bas1s Ventures, and A41. This round of financing will enable FLock to advance the democratization of AI model development and continue to build a decentralized AI training platform.FLock aims to address issues such as lack of transparency, weak data privacy, and unfair compensation for contributors in traditional centralized AI systems through blockchain technology. With its decentralized collaboration model, data remains local and does not need to be uploaded to third parties, allowing contributors to fairly share the benefits derived from data, computing power, and feedback, giving everyone the opportunity to participate in AI development.

ChainCatcher news, according to Cointelegraph, OpenAI co-founder Ilya Sutskever stated at the NeurIPS 2024 conference in Vancouver that the era of AI pre-training is coming to an end, and the era of AI superintelligence is approaching. He compared data to fossil fuels, pointing out that the growth of internet data has peaked, while the growth rate of computing power has surpassed the total amount of available training data.Sutskever predicts that the next phase of AI development will be dominated by three main directions: autonomous AI agents, synthetic data, and computational reasoning.

ChainCatcher news, Binance co-founder Zhao Changpeng posted on X: "The last use case for humanity is training artificial intelligence."

ChainCatcher news, according to The Block, early-stage investment firm Portal Ventures has announced its support for Stacks' 1:1 Bitcoin programmable asset sBTC. The firm plans to launch a series of seminars to explain the role of sBTC in enhancing the capital efficiency of Bitcoin for institutional investors.Portal Ventures recently completed a $75 million second-phase fundraise, with investors including Chris Dixon, Marc Andreessen, Henry Kravis, and well-known institutions like Accolade and Theta. Portal Ventures is one of the earliest supporters of the Stacks ecosystem, and Portal Ventures partner Catrina Wang stated that sBTC will become an important bridge connecting Bitcoin and the DeFi ecosystem.

ChainCatcher news, the decentralized AI training network Nodepay has completed a new round of financing, with participation from Animoca Brands, Jump Crypto, Elevate Ventures, OKX Blockdream, Mirana Ventures, NGC Ventures, and others. The specific amount of financing and valuation information has not yet been disclosed.

ChainCatcher news, Meta will begin using content publicly shared by adult users in the UK on Facebook and Instagram to train its artificial intelligence models. The company will use public information, such as posts, comments, photos, and captions from adult users on both platforms.Meta stated, "We are building AI at Meta to reflect the diverse communities around the world, and we look forward to launching it in more countries and languages later this year." Meta indicated that it will not use account information from individuals under 18 in the UK.In July, the Irish Data Protection Commission ordered Meta to suspend the launch of its AI assistant in the EU due to data privacy concerns, and Meta subsequently paused its AI rollout in the EU.

ChainCatcher news, the El Salvador National Bitcoin Office (ONBTC) announced that it will provide Bitcoin training and certification for 80,000 government employees. This training program, named "Public Administration Certification 1," has a total duration of 160 hours and is divided into 7 modules, covering various aspects such as strategic management of Bitcoin as legal tender and public policy.ONBTC Director Stacy Herbert stated that training civil servants will have a "compound effect" on El Salvador's Bitcoin-driven economy and plans to announce more educational programs. The certification is implemented by the Higher School of Public Administration Innovation (ESIAP) and aims to enhance the standards of excellence in governance and public administration in El Salvador.

ChainCatcher news, according to Cointelegraph, after a court hearing in Ireland, the European Data Protection Board (DPC) stated that Elon Musk's social media platform X has agreed to stop collecting and processing user data in the EU.It is reported that the X platform has agreed to stop processing personal data contained in public posts of its EU/EEA users from May 7, 2024, to August 1, 2024, for the purpose of training its artificial intelligence Grok.

ChainCatcher news, Musk posted on X that Grok 3 will be trained using greater computing power and is expected to be released by the end of this year.Previous news, Musk stated on the X platform that a lot of work is needed to clean LLM from internet training data, and GROK 2, which will be launched in August, will make significant progress in this regard.

ChainCatcher news, decentralized GPU cloud infrastructure Aethir has partnered with generative AI application platform company TensorOpera. TensorOpera will accelerate the deployment, fine-tuning, and full training of its latest open-source language model TensorOpera Fox-1 by integrating Aethir's enterprise-grade GPU resource pool. This is Aethir's first case of large-scale LLM training on a distributed physical infrastructure network.

ChainCatcher news, according to an official announcement from Binance, Binance has recently been invited to collaborate with the Macau Judicial Police to hold its first face-to-face law enforcement training course in Macau. This training was led by Binance's investigation team, with over 40 representatives from the Macau Judicial Police's Criminal Investigation Department and the Economic Crime Investigation Department related to gambling participating. The training held on May 21 is an important step in strengthening cooperation with the law enforcement community in the Macau region.In 2023 alone, Binance has hosted over 120 online and offline law enforcement training sessions or seminars globally. Last month, Binance held its first-ever online Law Enforcement Training Day, with over 1,300 law enforcement officers from around the world participating in the training.

ChainCatcher news, according to The Block, FLock.io has launched its AI training platform's web application train.flock.io. By integrating private data and on-chain rewards, FLock ensures fair incentives and promotes open collaboration. It meets the demand for customized AI models from both Web3 and Web2 projects while protecting data privacy by training models without exposing source data.It is reported that this beta version brings Web3 closer to providing traceable contributions and on-chain incentives for data owners, model developers, and computing providers, creating dedicated models for different communities.