Open-Source AI

Today's Stories

1. 1

   AI coding tools spawn 'slopsquatting' supply chain attacks

   Slopsquatting is a new supply chain attack that exploits AI language models' tendency to generate fictitious software package names. Attackers register those fake package names and populate them with malicious code, which developers unknowingly install when they use AI coding assistants. As developers increasingly rely on AI coding assistants, they may unknowingly grant cybercriminals access to their software from day one. The attack combines AI hallucinations—where models invent nonexistent packages—with traditional typosquatting tactics, creating a threat developers cannot easily spot.

   The body does not specify availability timelines, remediation steps, or quantified impact; readers should monitor how development teams and tool vendors respond to this emerging attack vector.

2. 2

   Claude Code's web research hits a wall—token costs soar on JavaScript-heavy sites

   A developer measured token costs for Claude Code (Anthropic's AI coding agent) retrieving web pages and found wide variation: an average Wikipedia article costs 68,240 tokens of raw HTML, while Nike's homepage costs 353,000. The tool's built-in web fetch works well on simple sites like Wikipedia (summarizing to about 950 tokens) but fails entirely on JavaScript-rendered pages and anti-bot sites, returning no data or a 403 error. For developers using AI agents to do research, this reveals a hidden cost: when the standard fetch fails, the agent dumps raw HTML back into context—still failing to extract information—which wastes tokens and processing time. The developer also notes cases where the agent appeared to fall back on training data or cached information, raising questions about reliability when live data retrieval is needed.

   The developer has built an open-source stealth browser (a recompiled Chromium instance) that runs as an MCP (Model Context Protocol) for Claude Code, Cursor, and Claude Desktop, designed to bypass these failures by returning cleaned-up HTML instead of raw markup.

3. 3

   Open source AI now used by roughly half Fortune 500, Hugging Face CEO says

   Hugging Face CEO Clem Delangue stated that open source AI has grown significantly, with the company now used by roughly half the Fortune 500. He described Hugging Face as functioning like a GitHub for AI, where builders share and download open models and datasets. Companies typically start with frontier APIs but shift toward open source models as they scale due to cost pressures. Delangue expressed concern that a handful of big companies could end up controlling everything, and highlighted that Chinese labs are producing the majority of open models being downloaded in the U.S.

   Delangue sees robotics as an especially urgent case for open, transparent AI because of how much a robot can see in a home and family life. He has also turned down a large investment from Nvidia, choosing capital efficiency over traditional Silicon Valley fundraising.

4. 4

   FORT Robotics brings outside-in safety to robots with Nvidia Halos

   FORT Robotics joined Nvidia's Halos for Robotics ecosystem and is demonstrating an agentic safety application using Nvidia's Outside-In Safety Blueprint this week at the Automate conference in Chicago. The solution combines external infrastructure sensors and visual AI agents with onboard robot perception to deliver real-time functional safety. Traditional robot safety systems rely only on onboard sensors and force robots to operate conservatively, slowing them down in dynamic warehouse and factory environments. Outside-In Safety automatically adjusts robot efficiency across changing environments, which means warehouses and factories can run robots faster while keeping workers safe—unlocking cost savings from processes like trailer unloading, inventory replenishment, and product assembly.

   FORT is a member of Nvidia's AI Systems Inspection Lab, the world's first ANSI National Accreditation Board (ANAB)-accredited inspection lab designed specifically for physical AI and autonomous systems. The lab verifies functional safety, cybersecurity, and AI compliance for autonomous vehicles, robotics, and sensor technologies.

5. 5

   Hugging Face CEO: companies are abandoning costly AI APIs for open-source models

   Hugging Face CEO Clem Delangue told TechCrunch's Equity podcast that companies consistently follow a pattern—they start with frontier APIs, but as they scale, costs push them toward open-source models. Hugging Face, described as a GitHub for AI where builders share models and datasets, is now used by roughly half the Fortune 500. The shift reflects a real economic pressure on businesses; paying per-API call becomes unsustainable once usage grows. Delangue is concerned that a handful of big companies could end up controlling everything, suggesting the open vs. closed-source fight shapes who gets access to AI tools and at what cost.

   Delangue's comments come in the wake of Anthropic's halted Fable release, which appears to have prompted renewed discussion about whether the industry's future belongs to open or proprietary models.

6. 6

   Ollama raises $65M to expand open-model AI platform

   Ollama Inc. announced a $65 million（約100億円） Series B funding round led by Theory Ventures, with participation from Benchmark, 8VC, Y Combinator, Pace Capital, 49 Palms, GTMFund, and other investors and angels. Ollama positions itself as the largest platform connecting developers to open-source AI models. The funding signals investor confidence in the open-model developer ecosystem, which competes with proprietary AI platforms.

   The company plans to use the capital to grow its platform, though specific product roadmap details or geographic expansion plans are not detailed in the announcement.