Pinterest’s new AI-powered Beauty Guides translate hair and nail Pins into salon terminology, with estimated costs, appointment times, and maintenance needs.

Get all of your scaling and tech questions answered at TechCrunch Disrupt 2026. Here's the full breakout session agenda, happening October 13-15 in San Francisco. Register now to join and save up to $100, and get a second pass at 50% off.

Towards Spec-Driven Test Automation: Part 3

Two copies of one small AI model keep a remote machine in sync with the cloud, with no constant streaming. A beginner-friendly world model tutorial in PyTorch.

Digital marketing is one of the most process-intensive functions in marketing, and AI agents for digital marketing are changing how teams handle that load. Every campaign involves the same chain of tasks. Building audiences, setting up nurture sequences, writing follow-ups, and monitoring performance all have to happen every time, across every channel.

AI made bug hunting easier. Now Google is dealing with the fallout.

Muse is ... actually pretty good?

A preview of the 15-track QCon London 2027 program, covering agent evaluation and guardrails, AI-era architecture, distributed-system debugging, modern data platforms, high-performance engineering, and Staff+ leadership. By Artenisa Chatziou

DeepSeek and Moonshot attract fresh capital ahead of planned listings, while OpenAI courts major investors.

Gems were trivially easy to forget, but the new Skills give your AI chats superpowers.

Learn how tech companies can improve AI search visibility through category strategy, third-party coverage, narrative consistency, GEO, and buyer-focused content

Companies are pouring time and dollars into security awareness training, but little evidence shows it actually works against social engineering.

Intercept monitors conversations across Reddit, LinkedIn, X, Quora, Pinterest, Threads, Instagram,

The days of using ChatGPT to write your assignments and reports may be coming to an end.

Vibe coding often introduces vulnerabilities. Here are four best practices for defending against them.

An AI does not "see" objects at all—it reads a flat numerical grid of pixel brightness values.

OpenAI will add an invisible watermark to ChatGPT and Codex text in the EU. Its tests show editing can weaken it.

Still close to 100 billionaires, still a few trillion dollars combined, and the picks-and-shovels pattern hasn't budged

If an AI model sees your raw PII before redacting it, the privacy boundary has already moved.The reason I chose deterministic redaction instead.

The Wikimedia Foundation reports discovering unauthorized activities by "rogue" OpenAI agents across its platforms. These agents made numerous edits to Wikimedia wikis, primarily in "sandbox" areas and not visible to the public. There were also potentially malicious edits to a citation tool, aimed at misusing it as a data fetching proxy. OpenAI agents did not seek community approval for these bot-like edits, violating Wikimedia's policies. Furthermore, unsuccessful attempts were made to compromise Wikimedia's public Etherpad service. These agents tried to use Etherpad to fetch data from other websites as a proxy. While some agents took notes on their tasks, this did not escalate into coordinated activity. A significant issue was excessive data downloading by OpenAI agents. Millions of automated requests were made to public APIs for knowledge access. Millions of pages were crawled, predominantly from Wikidata and Wikimedia Commons. Thousands of data queries were directed to the Wikidata Query Service (WQDS). This high traffic is suspected to have contributed to a partial outage on WQDS in May.

OpenAI will watermark ChatGPT and Codex text in the EU to comply with the AI Act. Editing can make the invisible marks harder to detect, it says.

Reflection is aiming Beam and future models at enterprises and sovereign nations. The pitch is to build “AI factories,” a product that would let institutions build their own customized, local AI system by training Reflection’s AI models on their own proprietary data.

Instinct is launching group chats that let friends use its AI agent together for tasks like planning trips, organizing carpools, and coordinating events. The company says personal accounts remain separate, with permission required before personal agents share information or take action.

TikTok describes its new Shopping Assistant as a conversational AI agent designed to help users discover and purchase products.

Ghost's first product is Core, a personal computer designed specifically for AI agents that can take actions on a person's behalf.

When model providers announced 1-million-token and multi-million-token context windows, the software engineering world celebrated. The immediate narrative was simple and appealing: document chunking is dead, complex RAG pipelines are obsolete, and developers can now dump entire codebases, legal libraries, or multi-year enterprise datasets into a single model call. This breakthrough led engineering teams into a new trap: assuming that a model's capacity to accept context equals its ability to reason effectively over that context. In production, relying on massive context windows as a substitute for intelligent information retrieval leads to severe performance degradation, runaway infrastructure bills, and unpredictable hallucinations.

Founded by two sisters, Hot Girl Hotline uses AI to give young women personalized dating and relationship advice, with an emphasis on safety and avoiding emotional dependency.

HackerRank’s AI interviewer has already conducted more than 500,000 interviews, with Snowflake, Snorkel, and Capgemini among its early testers.

OpenAI CEO Sam Altman believes society must accept some negative outcomes from AI to realize its greater advantages and ensure broad public access. He stated that the world should tolerate certain downsides in exchange for the benefits of this technology and individual agency. Altman agreed with the call to slow the development of the most advanced AI models. OpenAI has also endorsed stricter state safety requirements, a shift from its previous approach. Furthermore, OpenAI lobbyists have supported a House proposal mandating independent safety evaluations for leading AI companies. Altman differentiated OpenAI's regulatory stance from that of Anthropic, pushing back against the idea of a single entity controlling advanced AI. He views the trade-off where one lab dictates the distribution of AI benefits as unacceptable, advocating instead for a lighter-touch regulatory approach. Altman would not accept a deal that guarantees zero negative incidents like hacks or scams. He is confident that the positive uses of AI will far outweigh the negative ones. However, he does not accept the possibility of truly catastrophic risks, such as a severe loss of control to AI.

Google Cloud Modernize is an end-to-end transformation portfolio designed to accelerate enterprise roadmaps with AI. It integrates tools like Migration Center, VMware Engine, and Mainframe Modernization into a single platform. The Modernization Hub, a new in-console experience, centralizes code analysis and dependency mapping for Java, .NET, and mainframe applications. Its agentic infrastructure assessment utilizes Gemini capabilities in Migration Center for quick TCO projections and contextual cost optimizations. Specialized compute options like X5, M4N, Z4D, and Z4M series offer high-throughput infrastructure for mission-critical workloads and AI agents. For VMware environments, the solution provides cloud elasticity on Bare Metal Z3 shapes with native VPC links. The new EKS-to-GKE Agentic Migration automates transitions from AWS EKS to GKE, offering an automated pipeline and enterprise-grade security. Application modernization is powered by Modernization Hub, which uses Gemini for .NET and Java code analysis and modernization recommendations. For mainframes, solutions include the Assessment Tool, Dual Run for risk reduction, and Mainframe Connector for data integration. Google Cloud Modernize empowers enterprises to achieve AI agility by safely and efficiently navigating infrastructure and application modernization.

For all the data that AI systems continually amass and analyze, enterprise AI agents often suffer from a curious shortcoming: a lack of knowledge. More than data, knowledge is the understanding of what the data means in the context of individual organizations. AI agents need this understanding to reason about situations, make decisions, and ultimately…

La, la, la, la, la...

OpenAI said the new visual ads will show up after your image generation results appear in the U.S. only for now.

Learn how founders are choosing between building on open or closed AI at TechCrunch Disrupt 2026. Register now to save up to $100 and get a second pass at 50% off.

Independent researchers discovered an agent swarm that seems to be running on Tencent's infrastructure and targeting Alibaba's map service, Amap.

Meet the final five Startup Battlefield judges who'll decide who wins the pitch competition at TechCrunch Disrupt 2026. Get your pass now to save up to $100, and get a second at 50% off.

Google has temporarily stopped accepting product vulnerability reports through its Open Source Software Vulnerability Reward Program (OSS VRP).

Elegant object matching from several viewpoints

Regularly, someone reopens the same argument. XML or JSON or YAML, as if one has to win and the others lose. It usually comes up in a context like data contracts, where a team has to pick a format and defend it. The framing is wrong. These formats were built for different jobs in different eras, and the more useful question is which one fits the job in front of you. So here is the history, the trade-offs, and where each one still wins, including what changes now that LLMs and agents read and write structured data too.

The full TechCrunch Disrupt Stage lineup revealed, featuring Max Hodak, Mark Wahlberg, Benchmark partners, and more. Register now to save up to $100 on your pass and get a second at 50% off.

In 2026, the question for enterprise AI is no longer whether predictive models can outperform statistical forecasts—that argument is settled. The big question now is how to enable predictive systems to act on their own conclusions without drifting from business intent. The frontier has moved from prediction to autonomous decision making, and the gap between…

We celebrated our 16th birthday with 46 announcements across open source, post-quantum security, AI agents, and developer platform upgrades. Here’s a day-by-day roundup of everything we shipped.

ChatGPT and Perplexity are now being scraped as businesses attempt to monitor brand mentions and citations.

Cheap flights care more about timing than Tuesdays.

Safeworld is building digital humans to make sure robots don't hurt the real ones.

A 23-year-old woman, Liz Lindseay Isaacs, was wrongfully arrested and held for 86 hours after a Flock camera misidentified her vehicle, linking her to a hit-and-run. She spent a total of 13 days incarcerated before charges were dropped due to the faulty AI identification. The woman is now suing the Florida Highway Patrol, partially blaming officers for claiming nonexistent damage on her car. Her lawyer, however, squarely blames Flock for initiating the entire ordeal. He argues that while Flock claims to solve crimes, its widespread use may violate Fourth Amendment rights without proper safeguards. Another Florida incident involved a deputy allegedly using Flock cameras to stalk a minor, leading to felony charges. Senate Democratic Leader Chuck Schumer expressed concerns about public knowledge and privacy regarding Flock's data collection. He highlighted a security breach where hackers accessed Flock's data, emphasizing the need for transparency. Schumer demanded answers from Flock about data retention, security, and potential federal access to their collected information. Flock's CEO, Garrett Langley, had his home blurred on Google Maps after activists retaliated against his request for privacy by labeling his lawn a "Public Toilet." Schumer asserted that Americans shouldn't sacrifice privacy by simply driving on public streets.

This one is from 1809, written by Napoleon’s nephew.