Latest news with #decentralized
Forbes
3 days ago
- Business
- Forbes
How Direct Preference Optimization Can Bring User‑Driven Agility To AI
Ashutosh Synghal, VP of Engineering at Midcentury Labs, pioneers of a decentralized AI & secure data exchange. Imagine training a voice‑recognition system without hand‑transcribing thousands of hours of audio. Traditional supervised learning demands that developers label every snippet with exact text—a costly, error‑prone bottleneck. Now flip the script: Present two candidate transcriptions and ask a reviewer which sounds closer to reality. That quick 'A or B?' encodes far more than it seems. Multiply it across samples, and you obtain a rich dataset of human judgment. This shift—teaching AI through preference rather than perfection—is powering a new training method called direct preference optimization (DPO). From Labels To Choices The recent boom in generative AI has exposed the pain of manual labeling. For tasks such as captioning images or refining a chatbot's tone, there is rarely a single 'correct' answer—only a spectrum of better or worse ones. DPO exploits that truth by optimizing models directly on comparisons: Which output did people prefer? The idea builds on reinforcement learning from human feedback (RLHF). RLHF asks humans to rank outputs, trains a separate reward model and then fine‑tunes the base model via reinforcement learning. It works, but the pipeline is heavy: three models, delicate reward tuning and weeks of compute. Stanford researchers showed you can drop the middleman. With DPO, you skip the reward model and the reinforcement loop. You simply fine‑tune the base model so that preferred answers become more probable and rejected ones less so. Alignment becomes a straightforward classification‑style loss, reducing complexity and instability. Faster, Cheaper, Often Better Because it eliminates reward modeling and iterative rollouts, DPO can reduce training time and compute budgets. Teams iterate in days, not weeks, and early studies find quality equal to—or slightly better than—classic RLHF on tasks such as sentiment control and summarization. In one benchmark, a language model tuned with DPO outranked its RLHF counterpart in human preferences while using a fraction of the resources. Why Preference Learning Wins Humans are far better at choosing between options than at crafting flawless answers from scratch. Pairwise votes or thumbs‑up/thumbs‑down signals capture that intuitive skill, bypassing the need for exhaustive gold‑standard datasets. A customer‑service bot, for instance, can launch with a starter model and rely on user clicks or ratings to improve continuously—no massive annotation campaign required. Organizations already sit on mountains of implicit preference data: A/B tests, search click‑through rates, star reviews. DPO transforms that by‑product into training fuel. Microsoft's Azure OpenAI team notes that customers 'often have preference data already collected' and can reach RLHF‑level quality with a far simpler workflow. The method also shines in subjective domains—speech, translation, multimodal generation—where 'correctness' is nuanced. Whether a voice assistant sounds friendly or an image caption feels apt is ultimately a matter of taste. By directly optimizing for those tastes, DPO teaches models tone, style and context in ways rigid labels cannot. Momentum In The Market Open‑source communities quickly embraced DPO as an accessible alignment strategy, and enterprise platforms are following. Azure OpenAI now offers DPO fine‑tuning in preview, citing equal effectiveness to RLHF with faster turnaround. Intel's NeuralChat and several startups report similar gains. The technique is moving from research curiosity to industry standard. Real‑World Impact For product builders, the benefits are tangible: • Speed: Preference loops compress iteration cycles, letting a two‑person team ship and refine a niche speech application in weeks. • Cost: Cutting out reward models can save compute and reduce carbon footprints. • Safety: Because humans review outputs, harmful or biased generations are spotted earlier. DPO's direct link to user sentiment can also curb model drift toward unwanted behaviors. Caveats To Address DPO isn't magic. Biased or low‑quality feedback will poison results, and narrow demographic sampling can overfit the model to a single user group. Teams must curate diverse, representative comparisons and periodically audit outcomes. The good news? The efficiency gains free up time and budget to do exactly that. Still, swapping labels for likes doesn't magically wash away bias. If the judgments you feed a model are lopsided or sloppy, DPO will learn those flaws just as efficiently. The fix isn't glamorous, but it's straightforward: Be explicit about what "better" means, draw signals from a genuinely mixed crowd, and watch how that crowd behaves over time. I start with a one‑page cheat sheet for reviewers—clarity, safety, usefulness, tone—so 'prefer A' isn't just a gut reaction, but a choice grounded in shared criteria. Diversity beats sheer volume. A thousand comparisons from one demographic tells you what that group likes, not what your market needs. I've had more success with smaller, stratified batches—different regions, expertise levels and even devices—than with massive but skewed logs. And because not all clicks carry the same signal, I quietly weigh feedback: Quick, low‑effort taps matter less than consistent raters whose judgments line up with peers. Maintenance is constant but light. I seed each batch with a few 'gold' pairs I already know the answer to; if accuracy on those slips, something's off—fatigue, fuzzy instructions or a pipeline bug. I also schedule periodic red‑team passes around sensitive topics. Those exercises surface blind spots and generate fresh comparison pairs that keep the model honest. The upside of DPO's efficiency is that you can afford this hygiene. When you're not burning weeks on reward tuning, you can spend that time auditing feedback, tightening guidelines and collecting smarter comparisons. In my own projects, that trade—less GPU thrash, more human rigor—has been the difference between a model that merely ships and one that actually feels aligned with its users. Listen To Your Users The future of AI training is looking a lot less like drudgery and a lot more like collaboration. By embracing preference‑based learning, we can reduce manual grind and gain a direct line to what people actually expect. In my own projects, models trained through the lens of human preference not only ship faster—they feel more attuned to users. In the high‑velocity AI market of 2025 and beyond, that alignment will be decisive. DPO proves that training smarter with feedback humans naturally provide unlocks better AI sooner—and it is fast becoming a cornerstone of modern development. Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
Crypto Insight
3 days ago
- Business
- Crypto Insight
Hyperliquid revenue surges as it takes users from Solana: VanEck
Massive adoption of decentralized derivatives exchange Hyperliquid led to a huge increase in Hyperliquid network revenue in July, largely at the expense of Solana, according to VanEck. In July, Hyperliquid earned 35% of all blockchain revenue, with growth coming specifically at the expense of Solana, as well as Ethereum and BNB Chain, VanEck researchers wrote in a monthly crypto recap report. 'Hyperliquid was able to capture much of Solana's momentum, and likely Solana's market capitalization, because it offers a simple, highly functional product,' VanEck head of digital assets research, Matthew Sigel, and fellow analysts Patrick Bush and Nathan Frankovitz said, adding: 'Hyperliquid has poached high-value users from Solana and has retained them.' While Solana has struggled with reliability issues and failed to meet production deadlines for core software upgrades, Hyperliquid has capitalized by providing a superior user experience in derivatives trading, they said. 'Solana has not delivered meaningful improvements to boost its user experience, specifically in perpetual futures (perps) trading, and Hyperliquid stepped up with a better product.' Hyperliquid open interest surges 'Hyperliquid is emerging as the leading onchain perps venue,' reported Our Network in a newsletter seen by Cointelegraph. Open interest reached $15.3 billion in July and is up 369% year-to-date, with more than $5.1 billion USDC having been bridged in, it added. Phantom Wallet integration, which offers in-app perps, drove $2.66 billion in volume, $1.3 million in fees, and 20,900 new users to Hyperliquid in July. Crypto perpetual futures are derivative contracts that let traders speculate on cryptocurrency prices without expiration dates. HYPE prices hit July all-time high The platform's native token has also rallied, hitting an all-time high of $49.75 on July 14 from a low of just over $10 in early April. Solana's native token has lost 44% since its January all-time high, which was primarily driven by the memecoin frenzy. HYPE was trading down 3% on the day at $37.38 in a broader market retreat, at the time of writing. Source:
Associated Press
4 days ago
- Business
- Associated Press
Imagen Network (IMAGE) Introduces Smart Community Hubs for Decentralized Social Collaboration
New AI-driven community spaces empower users to co-create, engage, and govern with full ownership in Web3 environments. Singapore, Singapore--(Newsfile Corp. - August 5, 2025) - Imagen Network (IMAGE), the decentralized AI social platform, has launched Smart Community Hubs, an innovative feature designed to enable seamless collaboration, engagement, and governance for social users in Web3. These hubs provide an environment where members can co-create content, manage decentralized discussions, and drive projects without relying on centralized platforms. [ This image cannot be displayed. Please visit the source: ] Empowering decentralized collaboration with intelligent, community-driven hubs. To view an enhanced version of this graphic, please visit: Smart Community Hubs integrate Imagen's AI-powered feed curation, adaptive moderation, and peer engagement tools, allowing communities to self-organize and thrive. Each hub can be customized to reflect its members' values, engagement style, and creative goals. Token-enabled features further allow governance, reward distribution, and direct creator monetization, empowering members to fully own their contributions. By combining AI personalization with blockchain-backed community tools, Imagen Network's Smart Community Hubs pave the way for next-generation social experiences. Users retain full control of their identities and data, while communities can evolve dynamically, scaling with the needs of participants and creators alike. About Imagen Network Imagen Network is a decentralized social platform that blends AI content generation with blockchain infrastructure to give users creative control and data ownership. Through tools like adaptive filters and tokenized engagement, Imagen fosters a new paradigm of secure, expressive, and community-driven networking. Media Contact Dorothy Marley KaJ Labs +1 707-622-6168 [email protected] Social Media Twitter Instagram To view the source version of this press release, please visit
Globe and Mail
5 days ago
- Business
- Globe and Mail
Imagen Network (IMAGE) Introduces Smart Community Hubs for Decentralized Social Collaboration
New AI-driven community spaces empower users to co-create, engage, and govern with full ownership in Web3 environments. Singapore, Singapore--(Newsfile Corp. - August 5, 2025) - Imagen Network (IMAGE), the decentralized AI social platform, has launched Smart Community Hubs, an innovative feature designed to enable seamless collaboration, engagement, and governance for social users in Web3. These hubs provide an environment where members can co-create content, manage decentralized discussions, and drive projects without relying on centralized platforms. Empowering decentralized collaboration with intelligent, community-driven hubs. To view an enhanced version of this graphic, please visit: Smart Community Hubs integrate Imagen's AI-powered feed curation, adaptive moderation, and peer engagement tools, allowing communities to self-organize and thrive. Each hub can be customized to reflect its members' values, engagement style, and creative goals. Token-enabled features further allow governance, reward distribution, and direct creator monetization, empowering members to fully own their contributions. By combining AI personalization with blockchain-backed community tools, Imagen Network's Smart Community Hubs pave the way for next-generation social experiences. Users retain full control of their identities and data, while communities can evolve dynamically, scaling with the needs of participants and creators alike. About Imagen Network Imagen Network is a decentralized social platform that blends AI content generation with blockchain infrastructure to give users creative control and data ownership. Through tools like adaptive filters and tokenized engagement, Imagen fosters a new paradigm of secure, expressive, and community-driven networking. Media Contact Dorothy Marley KaJ Labs +1 707-622-6168 media@ Social Media Twitter Instagram
Associated Press
5 days ago
- Business
- Associated Press
EU Awards €1.5M Grant to AIxBlock, With €61.5M More in the Pipeline
EU backs AIxBlock's mission to decentralize AI development and automate infrastructure workflows—starting with a €1.5M grant and €61.5M in non-dilutive funding pipeline CALIFORNIA, Aug. 04, 2025 (GLOBE NEWSWIRE) -- In a significant show of institutional support, AIxBlock, a decentralized AI tech startup in the Bay Area, has secured a €1.5 million innovation grant from the European Union, alongside with up to €61.5 million in 2 additional grants pre-approved to support its long-term expansion across Europe. The backing positions AIxBlock as a frontrunner in a fast-emerging category: decentralized, automation-first AI infrastructure. The move signals growing European interest in infrastructure that empowers open, decentralized, community-driven platforms—not just hyperscalers—to play a leading role in the AI ecosystem. A Regulatory-Ready Alternative to Centralized AI Infrastructure AIxBlock's €1.5M secured grant comes with zero dilution (with onboarding (KYC) underway prior to fund disbursement), allowing the company to deepen its technology footprint in the EU while preserving ownership and control. The €61.5M in pre-approved funding of 2 other grants will include support for building a physical decentralized GPU network, leveraging underutilized data centers in the region to offer scalable, eco-aligned compute infrastructure in future phases. What AIxBlock Builds AIxBlock is a full-stack platform for end-to-end AI development and workflow automation. It allows startups, agencies, and enterprises to: The platform is designed to give builders full ownership of their infrastructure stack—without vendor lock-in—and to make AI development and automation more affordable, modular, and decentralized. Open Source with Long-Term Contributor Incentives AIxBlock is built with an open-source foundation and invites the global developer and AI community to contribute. Those who participate in improving the platform—whether through code, infrastructure, validation, or ecosystem support—will be eligible for long-term revenue and benefit sharing, proportionate to the value and level of their contribution. This model ensures that core contributors, infrastructure providers, and validators are aligned with the platform's long-term success, with transparent tracking of input and rewards over time. Blockchain as a Coordination and Incentive Layer Blockchain technology plays a critical role in AIxBlock's architecture, serving as the underlying layer for transparency, incentive distribution, and coordination across its decentralized ecosystem. It powers: Scaling Through EU Infrastructure, Responsibly A key part of AIxBlock's roadmap involves tapping into the EU's large base of underutilized data centers to deploy its decentralized GPU network. This expansion is planned for future phases supported by the broader grant pipeline, reflecting a strategy rooted in efficiency, sustainability, and digital sovereignty. By revitalizing existing infrastructure rather than building from scratch, AIxBlock aims to deliver scalable compute capacity while minimizing both environmental impact and geopolitical dependence. Quiet Execution, Long-Term Vision Without any paid marketing since MVP, AIxBlock has already onboarded over 12,000 users organically and secured some enterprise clients who are deploying large-scale model development workflows on the platform. The team continues to focus on expanding core capabilities and preparing for the next wave of infrastructure activation across the EU. As regulators, enterprises, and builders begin to question the scalability and transparency of centralized AI pipelines, platforms like AIxBlock offer an emerging alternative—designed for compliance, customization, and control. Strategic Positioning for Market Leadership With EU support secured and a substantial funding pipeline in place, AIxBlock has built both the credibility and resources to scale with confidence. Its ability to win institutional grants—combined with growing demand for decentralized infrastructure—makes it a compelling partner for strategic investors and ecosystem collaborators. The convergence of AI infrastructure demand and decentralized technology adoption presents a generational opportunity. EU backing provides the regulatory confidence that institutional investors and enterprise clients increasingly seek when evaluating infrastructure plays. As the decentralized AI landscape matures, platforms with public-sector backing, proven technology, and execution capital will be best positioned to lead. AIxBlock's current standing—anchored by confirmed funding, a pre-approved pipeline, and a clear expansion roadmap—signals its emergence as a serious contender in the global AI infrastructure race. Building the AI Operating Layer for Europe With institutional trust, a clear multi-stage public funding pipeline, and a differentiated approach to AI infrastructure, AIxBlock is positioning itself as a critical operating layer for the next era of enterprise and open-source AI development. At a time when developers are demanding more freedom, privacy regulations are tightening, and the cost of compute is rising, AIxBlock offers a new model: Decentralize the resources. Automate the workflows. Reward the community. Build locally, scale globally. About AIxBlock AIxBlock is a next-generation AI infrastructure platform built for developers, startups, and enterprises to build, fine-tune, and deploy AI models with unmatched flexibility and cost efficiency. Combining decentralized GPU resources, end-to-end AI development tools, and workflow automation, AIxBlock eliminates vendor lock-in and reduces compute costs by up to 90%. The platform offers a full-stack, automation-first environment—from data collection and labeling to model training, deployment, and integration with external apps—powered by a decentralized marketplace of models and compute providers. With institutional funding from the European Union (€1.5M Secured) and a roadmap to activate €61.5M more in non-dilutive grants, AIxBlock is building the AI operating layer for Europe and beyond. Website | X | Telegram Contact: Lee Ng [email protected] Disclaimer: This content is provided by AIxBlock. The statements, views, and opinions expressed in this content are solely those of the content provider and do not necessarily reflect the views of this media platform or its publisher. We do not endorse, verify, or guarantee the accuracy, completeness, or reliability of any information presented. We do not guarantee any claims, statements, or promises made in this article. This content is for informational purposes only and should not be considered financial, investment, or trading advice. Investing in crypto and mining-related opportunities involves significant risks, including the potential loss of capital. It is possible to lose all your capital. These products may not be suitable for everyone, and you should ensure that you understand the risks involved. Seek independent advice if necessary. Speculate only with funds that you can afford to lose. Readers are strongly encouraged to conduct their own research and consult with a qualified financial advisor before making any investment decisions. Neither the media platform nor the publisher shall be held responsible for any fraudulent activities, misrepresentations, or financial losses arising from the content of this press release. In the event of any legal claims or charges against this article, we accept no liability or responsibility. Globenewswire does not endorse any content on this page. Legal Disclaimer: This media platform provides the content of this article on an 'as-is' basis, without any warranties or representations of any kind, express or implied. We assume no responsibility for any inaccuracies, errors, or omissions. We do not assume any responsibility or liability for the accuracy, content, images, videos, licenses, completeness, legality, or reliability of the information presented herein. Any concerns, complaints, or copyright issues related to this article should be directed to the content provider mentioned above. Photos accompanying this announcement are available at:
