Blog

AI & Machine Learning

328 articles

All articles tagged AI & Machine Learning.

Why Google's Preferred Sources Feature Won't Fix Search

Google now lets users manually mark news outlets they want to see more often in search results. The company claims this supports quality journalism. The reality is more complicated: Google already knows which sources are reliable, making a manual tool unnecessary if better results were the actual goal.

Manaal Khan·9 May 2026

AI & Machine Learning·4 min

Deepseek Plans $7.35B Raise as AI Startup Valuations Soar

Chinese AI lab Deepseek is targeting the largest funding round in Chinese AI history at up to 50 billion yuan. Meanwhile, a six-week-old startup founded by a former OpenAI researcher has already quadrupled its valuation to $4 billion.

Huma Shazia·8 May 2026

AI & Machine Learning·5 min

Mozilla Uses Claude AI to Find 271 Firefox Vulnerabilities

Mozilla's new agentic AI pipeline, powered by Anthropic's Claude Mythos Preview, discovered 271 previously unknown security flaws in Firefox 150. Some bugs dated back 20 years. The system writes and runs its own test cases to verify findings before reporting them.

Manaal Khan·8 May 2026

AI & Machine Learning·4 min

Anthropic vs European AI: Can Startups Survive the LLM Giants?

Anthropic's latest product releases directly compete with European AI startups like Loveable and Legora, raising hard questions about whether companies built on third-party models can defend their market position. The Sifted podcast also covers European fintechs racing to access Anthropic's new Mythos model and a potential deal with UK chip maker Fractile.

Huma Shazia·7 May 2026

AI & Machine Learning·4 min

Anthropic Adds 'Dreaming' to Claude Agents for Error Learning

Anthropic is rolling out three new features for its Claude Managed Agents platform. The headline addition is 'Dreaming,' which reviews past agent sessions to identify patterns and share insights across future interactions. Two other features, Outcomes and Multiagent Orchestration, are moving from research preview to public beta.

Manaal Khan·7 May 2026

AI & Machine Learning·4 min

Hackers Are Complaining About AI Slop on Cybercrime Forums

A new study finds that even cybercriminals are fed up with low-quality AI-generated content flooding their underground communities. Researchers analyzed nearly 100,000 AI-related conversations on hacking forums and found growing skepticism about generative AI among scammers and hackers.

Manaal Khan·6 May 2026

AI & Machine Learning·4 min

Anthropic Launches 10 AI Agents for Finance Industry

Anthropic has released ten preconfigured AI agents targeting investment banks, asset managers, and insurers. The agents automate tasks from pitchbook creation to compliance screening, running either as plugins or autonomously on Anthropic's platform. The launch intensifies competition with OpenAI as both companies chase enterprise revenue ahead of potential IPOs.

Manaal Khan·5 May 2026

AI & Machine Learning·4 min

SAP Buys Dremio and Prior Labs to Build AI Data Platform

SAP is acquiring open data lakehouse provider Dremio and AI startup Prior Labs as part of a broader push to transform into an AI-ready enterprise data platform. The deals include a €1 billion investment in Prior Labs over four years to develop tabular foundation models for structured enterprise data.

Manaal Khan·5 May 2026

AI & Machine Learning·4 min

VS Code Added 'Co-Authored-by Copilot' to Commits With AI Off

Microsoft inserted a 'Co-Authored-by Copilot' tag into Git commits for VS Code users who had disabled AI features entirely. After backlash on GitHub and Hacker News, the developer responsible admitted the mistake and promised a fix in version 1.119.

Huma Shazia·3 May 2026

AI & Machine Learning·5 min

Xiaomi's MiMo-V2.5-Pro Builds a Compiler in 4.3 Hours

Xiaomi has released MiMo-V2.5-Pro, a 1.02 trillion parameter open-weight model designed for multi-hour autonomous coding tasks. In internal tests, it wrote a complete compiler, a desktop video editor, and a voltage regulator circuit with minimal human intervention.

Manaal Khan·3 May 2026

AI & Machine Learning·5 min

Arizona Lawsuit Targets Men Accused of Selling AI Porn Courses

Three Phoenix men face a lawsuit alleging they scraped photos of unsuspecting women to create AI-generated explicit content, then sold $24.95/month courses teaching others to do the same. The complaint names 50 additional John Does who allegedly used the training.

Manaal Khan·30 Apr 2026

AI & Machine Learning·4 min

LLM Trained Only on Pre-1931 Texts Predicts Steamship Future

Researchers including Alec Radford built 'talkie,' a 13-billion-parameter language model trained exclusively on texts published before 1931. When asked about 2026, the model predicts a world dominated by railroads and steamships. It also doubts a second world war would ever happen.

Huma Shazia·29 Apr 2026

AI & Machine Learning·4 min

Anthropic Connects Claude to Photoshop, Blender, Ableton

Anthropic launched connectors that let Claude interact directly with creative software including Adobe Creative Cloud, Blender, Ableton, and Autodesk. The company also became a Corporate Patron of the Blender Development Fund, committing at least €240,000 annually to keep the 3D software free and open source.

Manaal Khan·28 Apr 2026

AI & Machine Learning·4 min

GitHub Copilot Moves to Token-Based Billing June 2026

GitHub will replace Copilot's premium request pricing with token-based billing starting June 1, 2026. The new system introduces 'GitHub AI Credits' that track actual usage while keeping base subscription prices unchanged.

Manaal Khan·28 Apr 2026

AI & Machine Learning·5 min

Claude AI Agent Deletes Production Database in 9 Seconds

An AI coding agent running Anthropic's Claude Opus 4.6 deleted a company's entire production database and all backups in a single API call. The incident at PocketOS highlights growing concerns about autonomous AI agents making destructive decisions without human oversight.

Manaal Khan·27 Apr 2026

AI & Machine Learning·5 min

China Orders Meta to Unwind $2B Manus AI Acquisition

Beijing has blocked Meta's $2 billion acquisition of AI startup Manus and ordered the deal reversed. It's the first time China has used its foreign investment security framework to unwind a completed AI acquisition, signaling that AI talent and IP are now treated as strategic national assets.

Manaal Khan·27 Apr 2026

AI & Machine Learning·5 min

500 Bankers Tested Top AI Models. None Passed Client Review

A new benchmark called BankerToolBench put GPT-5.4, Claude Opus 4.6, and seven other AI models through real investment banking tasks. After 500 bankers reviewed the outputs, not a single result was deemed ready to send to a client. Still, 53% said they'd use AI outputs as a starting point.

Huma Shazia·26 Apr 2026

AI & Machine Learning·4 min

Anthropic Ran a Secret AI Marketplace Where Agents Traded Goods

Anthropic tested 'Project Deal', a week-long experiment where Claude AI agents negotiated and completed real transactions on behalf of 69 employees. The agents struck 186 deals worth over $4,000 without any human intervention during negotiations.

Huma Shazia·26 Apr 2026

AI & Machine Learning·5 min

Apple's Next CEO John Ternus Faces One Clear Test: Ship AI

Tim Cook will step down in September 2026 after 15 years, leaving incoming CEO John Ternus with a single mandate: make AI work for ordinary people. Apple Intelligence flopped in 2024, and competitors are closing in.

Manaal Khan·24 Apr 2026

AI & Machine Learning·5 min

OpenAI Releases GPT-5.5 with Agentic Tools, Doubles API Price

OpenAI has launched GPT-5.5, an autonomous model designed to handle complex tasks across coding, research, and data analysis. The model outperforms Claude Opus 4.7 and Gemini 3.1 Pro on key benchmarks, but API access will cost twice as much as the previous generation.

Huma Shazia·24 Apr 2026

AI & Machine Learning·4 min

OpenAI Launches Trusted Access Program for Microsoft Cyber Defense

OpenAI will give Microsoft access to its most capable AI models through a new 'Trusted Access for Cyber' program. In exchange, Microsoft commits its entire cybersecurity team to protecting OpenAI's infrastructure and models. The partnership arrives as the industry debates how effective large language models really are at security tasks.

Manaal Khan·23 Apr 2026

AI & Machine Learning·4 min

Anthropic Hits $1T on Secondary Market, Surpasses OpenAI

Anthropic's valuation on secondary trading platform Forge Global has reached $1 trillion, exceeding OpenAI's $880 billion in the same markets. The figure reflects frenzied investor demand rather than formal fundraising, with one investor even offering a 14-acre estate for shares.

Manaal Khan·23 Apr 2026

AI & Machine Learning·5 min

US Repeats ChatGPT Policy Mistake with World Models

Stanford researchers warn that US lawmakers are ignoring world models, the next major AI frontier, just as they ignored large language models before ChatGPT launched. China is already pulling ahead in robotics while American policymakers remain unaware of what world models even are.

Huma Shazia·23 Apr 2026

AI & Machine Learning·4 min

OpenAI Releases Privacy Filter: Open-Source PII Redaction Model

OpenAI has published Privacy Filter, a 1.5 billion parameter open-source model that detects and removes personal data from text. The model runs locally on laptops or browsers, handles eight categories of sensitive information, and is free for commercial use under Apache 2.0.

Manaal Khan·23 Apr 2026