hodler

15 May 2026, 12:45

Osaurus brings both local and cloud AI models to your Mac

BitcoinWorld Osaurus brings both local and cloud AI models to your Mac As AI models become increasingly commoditized, startups are racing to build the software layer that sits on top of them. One interesting entrant into this space is Osaurus, an open-source, Apple-only LLM server that lets users move between different local AI models — either running on-device or in the cloud — while keeping their files and tools on their own hardware. From AI companion to local LLM server Osaurus evolved out of the idea for a desktop AI companion called Dinoki, which co-founder Terence Pae described as a sort of “AI-powered Clippy.” Dinoki’s customers had asked him why they should buy the app if they still had to pay for tokens — the usage units AI companies charge for processing prompts and generating responses. That got Pae thinking more deeply about running AI locally. “That’s how Osaurus started,” Pae, previously a software engineer at Tesla and Netflix, told Bitcoin World over a call. The idea, he explained, was to try to run an AI assistant locally. “You can do pretty much everything on your Mac locally, like browsing your files, accessing your browser, accessing your system configurations. I figured this would be a great way to position Osaurus as a personal AI for individuals.” Pae began building the tool in public as an open-source project, adding features and fixing bugs along the way. How Osaurus works: a harness for AI models Today, Osaurus can flexibly connect with locally hosted AI models or cloud providers like OpenAI and Anthropic. Users can freely choose which AI models they’re using, while keeping other aspects of the AI experience on their own hardware — like the models’ own memory, files, and tools. Given that different AI models have different strengths, the advantage of this system is that users can switch to the AI model that best fits their needs. Such a structure makes Osaurus what’s called a “harness” — a control layer that connects different AI models, tools, and workflows through a single interface, similar to tools like OpenClaw or Hermes. However, those tools are often aimed at developers who know their way around a terminal. And sometimes, like in the case of OpenClaw, they may pose security issues. Osaurus, meanwhile, presents an easy-to-use interface for consumers and addresses security concerns by running things in a hardware-isolated, virtual sandbox. This limits the AI to a certain scope, keeping your computer and data safe. Hardware requirements and the future of local AI Running AI models on your machine is still in its early days, given that it’s heavily resource-intensive and hardware-dependent. To run local models, your system will need at least 64 GB of RAM. For running larger models, like DeepSeek v4, Pae recommends systems with about 128 GB of RAM. But Pae believes local AI’s needs will come down in time. “I can see the potential of it, because the intelligence per wattage — which is like the metric for local AI — has been going up significantly. It’s on its own curve of innovation. Last year, local AI could barely finish sentences, but today it can actually run tools, write code, access your browser, and order stuff from Amazon. It’s just getting better and better,” he said. Supported models and plugins Osaurus today can run MiniMax M2.5, Gemma 4, Qwen3.6, GPT-OSS, Llama, DeepSeek V4, and other models. It also supports Apple’s on-device foundation models, Liquid AI’s LFM family of on-device models, and in the cloud, it can connect to OpenAI, Anthropic, Gemini, xAI/Grok, Venice AI, OpenRouter, Ollama, and LM Studio. As a full MCP (Model Context Protocol) server, you can give any MCP-compatible client access to your tools as well. Plus, it ships with over 20 native plugins for Mail, Calendar, Vision, macOS Use, XLSX, PPTX, Browser, Music, Git, Filesystem, Search, Fetch, and more. More recently, Osaurus was updated to include voice capabilities. Adoption and next steps Since the project went live nearly a year ago, it has been downloaded north of 112,000 times, according to its website. Currently, Osaurus’ founders (who include co-founder Sam Yoo) are participating in the New York-based startup accelerator Alliance. They’re also thinking about next steps, which could see Osaurus being offered to businesses — like those in the legal space or in healthcare — where running local LLMs could address privacy concerns. As the power of local AI models grows, the team believes it could lower the demand for AI data centers. “We’re seeing this explosive growth in the AI space where [cloud AI providers] have to scale up using data centers and infrastructure, but we feel like people haven’t really seen the value of the local AI yet,” Pae said. “Instead of relying on the cloud, they can actually deploy a Mac Studio on-prem, and it should use substantially less power. You still have the capabilities of the cloud, but you will not be dependent on a data center to be able to run that AI.” Conclusion Osaurus represents a notable step in making local AI more accessible and practical for everyday Mac users. By offering a flexible harness that connects both local and cloud models with a strong emphasis on privacy and security, it addresses key concerns around data control and cost. As local AI models continue to improve in capability and efficiency, tools like Osaurus could help shift the balance away from cloud-dependent AI toward more decentralized, on-device solutions. FAQs Q1: What is Osaurus? Osaurus is an open-source, Apple-only LLM server that lets users run and switch between local and cloud AI models while keeping files and tools on their own hardware. Q2: What are the hardware requirements for running local models with Osaurus? For local models, a Mac with at least 64 GB of RAM is recommended. For larger models like DeepSeek v4, about 128 GB of RAM is advised. Q3: Which AI models and cloud providers does Osaurus support? It supports local models like MiniMax M2.5, Gemma 4, Llama, DeepSeek V4, and Apple’s on-device models. Cloud providers include OpenAI, Anthropic, Gemini, xAI/Grok, and others. This post Osaurus brings both local and cloud AI models to your Mac first appeared on BitcoinWorld .

15 May 2026, 10:23

OpenAI braces for fresh legal bout with Apple as Elon Musk trial nears end

OpenAI is getting ready to take Apple to court over their ChatGPT deal that went bad. The AI firm brought in an outside law firm to figure out what they can do. They might send Apple a letter saying the company broke their contract, though a full court case might not happen right away, according to a Bloomberg report. OpenAI probably would get done with Elon Musk’s trial mess before making any legal moves against Apple. The partnership with Apple was announced at its big conference back in June 2024. Apple was to integrate ChatGPT into iPhones, where users could directly inquire through Siri and also use the camera to send pictures to ChatGPT for queries. OpenAI wanted millions of new paying customers through the deal. But this partnership could not deliver what Altman expected. OpenAI blamed Apple for the failure because it hid the ChatGPT features and made them too hard for people to find. Money from the deal came nowhere close. One OpenAI manager said Apple asked them to “take a leap of faith and trust us” but “it didn’t work out well.” OpenAI is not the only one with concerns. Apple also worries how OpenAI deals with privacy. Apple’s long history of dumping partners This mess follows a pattern at Apple. The company has a long track record of working with big software firms and then pushing them out. Google Maps was a key part of the first iPhone, but Apple dumped it in 2012 for its own maps app. That went so badly that CEO Tim Cook had to say sorry in public. Google and Apple started having problems after Google made Android phones in 2008, one year after the iPhone came out. Eric Schmidt, who ran Google back then, left Apple’s board in 2009 as the companies became rivals. Steve Jobs refused to let Adobe’s Flash work on iPhones and iPads. He wrote a letter in 2010 explaining why, which basically killed Flash on phones. Spotify spent years saying Apple used the App Store to hurt other music apps after Apple Music started in 2015. European officials agreed and made Apple pay almost 1.8 billion euros in March 2024. But sometimes, Apple can patch things up when there’s money involved. Google is now helping Apple with AI after signing a deal in January. Google’s Gemini models will run the next version of Apple’s smart features. Apple is paying Google about $1 billion each year for this. OpenAI restructures Microsoft deal Just a month before, OpenAI changed its deal with Microsoft in April. The new setup puts a limit on how much money OpenAI has to share with Microsoft. OpenAI will keep paying Microsoft 20% of what it makes through 2030, but there’s now a cap on total payments. Reports say that cap is $38 billion. Microsoft put more than $13 billion into OpenAI starting in 2019. The new deal lets OpenAI work with other cloud companies like Amazon and Google, not just Microsoft. Microsoft can still use OpenAI’s technology through 2032, but other companies can use it too now. OpenAI signed a huge deal with Amazon in February. Amazon said it would put up to $50 billion into OpenAI and let the company use Amazon’s cloud services. OpenAI is also making hardware The company bought Jony Ive’s firm last May for $6.5 billion. Chris Lehane, who handles global affairs for OpenAI, said at a meeting in Davos that the first device should come out in the second half of 2026, directly entering Apple’s market. He wouldn’t say what it looks like, but reports say it might be small with no screen, maybe something you wear. Sam Altman said it will be simpler and calmer than a phone. As reported by Cryptopolitan , OpenAI is facing another legal battle with an apparently unserious Musk, who did not show up in the closing arguments because of his visit to China. OpenAI has even said that Musk had been looking at memes once during a meeting. If you're reading this, you’re already ahead. Stay there with our newsletter .

15 May 2026, 08:05

Trump discloses $220M in trades tied to U.S. companies in Q1

More on U.S. government Trump Brought An Army Of CEOs To Beijing For A Reason Calif. Gov. proposes new tax targeting cloud-based software sales Republican-led Senate Banking Committee advances crypto bill Trump touts China's order for 200 Boeing jetliners; shares fall

15 May 2026, 07:35

Tech and AI Leadership Propel Global Equities to Record Highs, Deutsche Bank Reports

BitcoinWorld Tech and AI Leadership Propel Global Equities to Record Highs, Deutsche Bank Reports Global equity markets have surged to unprecedented levels, driven overwhelmingly by the technology and artificial intelligence sectors, according to a new analysis from Deutsche Bank. The report highlights a concentrated rally, with a handful of mega-cap tech stocks accounting for a disproportionate share of the market’s gains, raising both opportunities and risks for investors. Market Drivers and Sector Performance Deutsche Bank’s analysis points to a sustained appetite for AI-related infrastructure, software, and services as the primary catalyst. Companies leading in generative AI, cloud computing, and semiconductor manufacturing have seen their valuations climb sharply, pulling major indices like the S&P 500 and Nasdaq to new all-time highs. The report notes that this rally is distinct from broader market advances, as gains remain heavily concentrated in a narrow set of high-growth names. Implications for Investors and the Broader Market While the tech-led surge has created substantial wealth for shareholders, Deutsche Bank cautions that the narrow breadth of the rally could signal underlying fragility. Historically, markets that rely on a small number of stocks for momentum are more vulnerable to sharp corrections if sentiment shifts. The report advises investors to consider diversification and to monitor valuation metrics closely, particularly in the AI and tech sectors where price-to-earnings ratios have expanded significantly. Geographic and Sectoral Spread The rally is not confined to the United States. European and Asian markets with significant exposure to AI and tech have also recorded gains, though at a more moderate pace. Deutsche Bank notes that the performance gap between tech-heavy indices and those weighted toward traditional sectors like energy, utilities, and financials has widened, reflecting a market that is increasingly pricing in a future shaped by AI-driven productivity gains. Conclusion Deutsche Bank’s report confirms that the current equity market record highs are a direct reflection of investor confidence in the transformative potential of technology and artificial intelligence. However, the concentration of gains in a few key players warrants careful attention from market participants. The sustainability of this rally will depend on continued innovation, earnings delivery, and the broader economic environment. For now, the market’s direction remains firmly tied to the fortunes of the tech and AI sectors. FAQs Q1: Which sectors are driving the current equity market record highs? The technology and artificial intelligence sectors are the primary drivers, with companies in AI infrastructure, cloud computing, and semiconductors leading the rally. Q2: What risks does Deutsche Bank highlight in its report? Deutsche Bank warns that the rally is narrowly concentrated in a few mega-cap tech stocks, which could make the market more vulnerable to a correction if sentiment changes or valuations become stretched. Q3: Is the rally limited to U.S. markets? No, European and Asian markets with significant tech and AI exposure have also seen gains, though the performance is most pronounced in the United States. This post Tech and AI Leadership Propel Global Equities to Record Highs, Deutsche Bank Reports first appeared on BitcoinWorld .

15 May 2026, 05:50

South Korea to Unveil Detailed Security Token Rules in July, Allowing Pooled Asset Issuance

BitcoinWorld South Korea to Unveil Detailed Security Token Rules in July, Allowing Pooled Asset Issuance South Korea’s Financial Services Commission (FSC) is set to announce detailed regulations and operational guidelines for security tokens in July, moving closer to the planned implementation of a formal framework in February next year. The announcement, first reported by local media outlet Money Today, signals a significant step in the country’s efforts to integrate blockchain-based financial instruments into its regulated capital markets. What the New Rules Will Cover Security tokens are digital securities issued and managed using distributed ledger technology, such as blockchain. They digitally record rights including ownership stakes and dividend entitlements, functioning similarly to traditional stocks or bonds but with the efficiency and transparency of blockchain settlement. Under the forthcoming rules, these tokens will be tradable through licensed securities firms, placing them squarely within the existing regulatory perimeter. The FSC’s planned measures include several key provisions. Authorities are pursuing the legalization of fractional investment securities that pool multiple underlying assets, allowing investors to hold small stakes in diversified portfolios through a single token. This approach could lower barriers to entry for real estate, infrastructure, and art investments. The regulator is also developing a roadmap for tokenizing traditional assets such as equities and debt securities, which would bring established financial instruments onto blockchain rails. Additionally, the FSC intends to expand trading limits on over-the-counter (OTC) exchanges, providing more flexibility for institutional and retail participants. Timeline and Implementation The July announcement will provide the industry with a detailed regulatory blueprint, giving market participants several months to prepare systems and compliance frameworks before the February 2026 enforcement date. The FSC has been working on the security token framework since 2023, when it first signaled its intention to create a legal basis for digital securities. The upcoming guidelines are expected to clarify issuance standards, custody requirements, disclosure obligations, and investor protection rules. South Korea joins a growing list of jurisdictions — including Singapore, Japan, and the European Union — that are actively crafting regulatory regimes for tokenized securities. The FSC’s approach emphasizes investor safeguards while encouraging innovation, reflecting a balancing act seen in many developed markets. Why This Matters for Investors and the Market The introduction of a formal security token framework has broad implications. For retail investors, fractional ownership of pooled assets could open access to previously illiquid or high-minimum-investment asset classes. For financial institutions, tokenization promises faster settlement, reduced counterparty risk, and the ability to create new products. For the broader market, it represents a bridge between traditional finance and blockchain technology, potentially attracting new capital and increasing market efficiency. However, challenges remain. The FSC must address questions around cross-border trading, interoperability between different blockchain platforms, and the treatment of tokens under existing securities laws. The July guidelines are expected to provide initial answers, but the full regulatory picture will only emerge as the framework matures. Conclusion South Korea’s FSC is moving decisively to establish a regulatory foundation for security tokens, with detailed rules due in July and enforcement beginning in February 2026. By allowing pooled asset issuance and expanding OTC trading limits, the regulator is laying the groundwork for a more inclusive and technologically advanced capital market. Investors and industry participants should monitor the July announcement closely for specifics on compliance requirements and market access. FAQs Q1: What is a security token under South Korea’s proposed rules? A security token is a digital representation of ownership or rights — such as dividends or voting power — issued and managed using blockchain technology. It functions like a traditional security but is settled and recorded on a distributed ledger. Q2: When will the new security token regulations take effect? The FSC plans to announce detailed guidelines in July 2025, with the formal regulatory framework taking effect in February 2026. Market participants will have several months to prepare for compliance. Q3: What types of assets can be tokenized under the new rules? The rules are expected to allow tokenization of traditional securities like stocks and bonds, as well as fractional investment securities that pool multiple underlying assets such as real estate or infrastructure projects. This post South Korea to Unveil Detailed Security Token Rules in July, Allowing Pooled Asset Issuance first appeared on BitcoinWorld .

15 May 2026, 05:25

PwC and Anthropic target enterprise tech debt with Claude Code rollout

On Thursday, PwC and Anthropic deepened their collaboration, allowing thousands of consultants to leverage the capabilities of Claude AI in revamping old enterprise systems. PwC will first certify 30,000 US employees on Claude Code, Anthropic’s AI coding agent, then extend access to its full workforce of 364,000 across 136 countries. The companies noted that the partnership will aim to tackle more than $2 trillion in enterprise tech debt worldwide, resulting from costly inefficiencies caused by legacy systems, data silos, and manual processes. McKinsey research has found large companies spend 10% to 20% of their technology budgets keeping old systems running rather than building new ones. Insurance underwriting dropped from 10 weeks to 10 days The partnership covers three areas: AI tools for software engineering teams, AI in mergers and acquisitions work, including due diligence, and the replacement of legacy finance, HR, and supply chain systems. PwC is also launching a new Office of the CFO business group built around Claude, focused first on banking, insurance, and healthcare. Anthropic CEO Dario Amodei said some clients are already seeing results in production, not pilots. “Insurance underwriting that took ten weeks now takes ten days. Security work that took hours now takes minutes,” Amodei said. A stalled HR project was restarted with a working prototype in one week. A COBOL mainframe migration is reportedly running on time and under budget despite a larger-than-expected workload. As Cryptopolitan reported in February, Claude Cowork already connects with Microsoft Office, Google Drive, and other enterprise tools, but scaling from pilot projects to company-wide deployment is where most AI initiatives stall. Every major consulting firm now has an AI lab deal Deloitte signed a Claude deployment covering its 470,000 employees last October. IBM embedded Claude models in its developer tools the same month. Goldman Sachs has had Anthropic engineers embedded inside the bank for six months, building AI agents for trade accounting and client onboarding. OpenAI is competing for the same channel. It holds partnerships with Accenture, Capgemini, BCG, McKinsey, and PwC itself. Google recently launched a $750 million program to help consulting firms deploy AI for clients. Anthropic has leaned into compliance and safety as differentiators for regulated industries. Business Insider reported Anthropic reached 34.4% enterprise adoption in April per the Ramp AI Index, edging past OpenAI at 32.3%. Claude Code was cited as the primary driver. Advocate Health, with its workforce of 167,000, is one of the first health-care companies to embrace widespread Claude adoption. Over 5,000 partners and other senior leaders from PwC have reportedly finished AI training sessions. The smartest crypto minds already read our newsletter. Want in? Join them .