Bite-sized AI for curious minds...
Visual automation platform
Make (formerly Integromat) is a visual automation platform. Build complex workflows by connecting 1500+ apps with drag-and-drop modules. AI modules support OpenAI, Anthropic, and more. Free tier includes 1000 operations/month. Great value at $9/mo for Core plan.
MIT Technology Review Explains: Let our writers untangle the complex, messy world of technology to help you understand what’s coming next. You can read more from the series here. When two OpenAI models hacked into the website Hugging Face in July, they weren’t trying to make money or commit sabotage—they were just looking for answers…
Amazon has finalized a $50 billion investment in OpenAI, marking one of the largest corporate investments in AI to date. The deal, reported by the Financial Times, solidifies a strategic partnership between the e-commerce giant and the ChatGPT maker. This follows Amazon's earlier $4 billion stake and includes expanded use of OpenAI's models across AWS and Alexa. The investment positions Amazon to compete more directly with Microsoft, which has invested over $13 billion in OpenAI. It also raises questions about regulatory scrutiny, as the deal gives Amazon significant influence over OpenAI's future direction.
Anthropic has announced that Claude Code, its AI-powered coding assistant, is now available for free. Previously a paid tool, this move opens up advanced AI-assisted development to a broader audience. The free tier includes access to Claude's powerful code generation and debugging capabilities, though usage limits may apply. This strategic decision aims to attract more developers to the Claude ecosystem and compete with other AI coding tools like GitHub Copilot.
A USA Today opinion piece highlights that most Americans support AI regulation, yet Congress remains stalled on legislative action. The article cites polling showing bipartisan concern about AI risks, including job displacement and misinformation. Despite public pressure, lawmakers have not passed comprehensive AI laws, leaving states to act independently. This disconnect between public opinion and legislative inaction underscores the urgency for federal oversight as AI tools like ChatGPT and Midjourney become mainstream.
A new open letter titled 'Open Weights and American AI Leadership,' shepherded by Microsoft and dated July 24th, has been signed by 235 AI professionals. The letter argues for the importance of open-weight AI models in maintaining American leadership in AI. It comes amid a broader debate over the risks and benefits of open-sourcing powerful AI systems. The letter's release follows a series of other open letters from the AI community, reflecting growing pressure on policymakers to consider open-source AI in national strategy.
A new open-source project, BitNet/TQ, introduces foundational ternary-model inference and training support for CUDA and CPU platforms. Ternary models use weights constrained to -1, 0, and 1, drastically reducing memory footprint and compute requirements compared to standard 16-bit models. The project, hosted on GitHub, aims to make ternary LLMs practical for edge devices and cost-sensitive deployments. This could enable efficient on-device AI without cloud dependencies, challenging the dominance of large-scale GPU clusters.
CleanSlate, an open-source platform for running coding agents, has launched with support for IDE, CLI, and SDK interfaces. The agents can understand a codebase, make changes, run commands, browse the web, and verify their work locally or in the cloud. The team is now building longer-running autonomous cloud agents that can continue working without keeping your machine active. This positions CleanSlate as a flexible alternative to proprietary coding agents, offering transparency and customization for developers.
CostPerPrompt, a new tool from an independent developer, launched on Hacker News as a Show HN, offering live AI API pricing and real-workload cost calculators. The platform tracks current prices for major AI providers like OpenAI, Anthropic, and Google, and lets users input their actual usage patterns to estimate costs. It aims to solve the problem of rapidly changing AI pricing, which makes budgeting difficult for developers and startups. The tool is particularly useful for comparing models like GPT-4o, Claude 3.5, and Gemini 1.5 Pro on a per-prompt basis, factoring in token counts and caching.
OpenAI has announced Astra, a new model family designed for long-horizon, multi-agent problem solving, capable of working on complex tasks for hours or even days. CEO Sam Altman has already demonstrated Astra to policymakers in Washington, showcasing its ability to solve ten previously unsolved math problems. OpenAI has not yet decided whether to release Astra as GPT-6 or as a new GPT-5 variant. This represents a significant shift from single-query models to persistent, collaborative AI systems.
A new YouTube video titled 'The AI Data Center Explained – From Electricity to ChatGPT' breaks down the infrastructure behind AI services, covering power supply, cooling, and the hardware stack that enables models like ChatGPT to run. The video is currently trending on Hacker News with minimal engagement, but it serves as an accessible primer for understanding the physical systems that power AI. It highlights the massive electricity consumption and specialized equipment required, from GPUs to data center cooling systems. This is relevant as AI's energy demands become a growing concern for operators and policymakers.