AI news for builders and product teamsUpdated Nov 18, 2026, 07:00 UTC
AI news, straight from the source.
Releases, research, and ideas for developers and product teams. Updated every hour.
Across the feeds
Newest first
Google announced Gemini 4 Argon, a frontier model with a 1-million-token output limit that is rolling out first to trusted cyber defenders through the Fairwind Program. The September 2026 roundup also includes Gemini 3.8 Flash, new voice and music models, Connected Apps, and science projects.

Google Research announced a next-generation Federated Learning system that uses Trusted Execution Environments (TEEs) to provide externally verifiable privacy guarantees while moving computation to the server for faster training and better accuracy. Gboard has adopted the system for English and Japanese next-word prediction models.

Nvidia has raised the US price of its Shield TV Pro streaming box by $100 to $299.99, effective October 2, citing substantially higher component costs including memory. The most recent version launched in 2019 at $199.99.

Clay co-founder and CEO Kareem Amin will speak on the AI Stage at TechCrunch Disrupt 2026 about the rise of the GTM engineer, a role Clay coined in 2023. Clay says roughly 100 GTM engineering job listings now appear monthly.

Uber has rebuilt major parts of the Uber Eats search pipeline and reports a 50% reduction in end-to-end search latency. The changes span retrieval, feature hydration, ranking, advertising, presentation and infrastructure, and an agentic coding workflow was used to identify and validate optimizations.

Airbnb CTO Ahmad Al-Dahle, formerly head of generative AI at Meta and lead of the Llama launches, says 60% of Airbnb's code is now AI-authored, its support agents resolve roughly half of tickets, and internal context graph Everest helped ship airport pickups in about six weeks.

Exhibit bookings for TechCrunch Disrupt 2026 close tonight, October 2, at 11:59 p.m. PT, or when tables sell out. The event runs October 13–15 at San Francisco's Moscone West, with 10,000+ attendees expected.

Cloudflare launched a Web Search API through its AI Gateway, letting agents retrieve live web results from providers Ceramic.ai, Exa, and Linkup instead of guessing URLs. Access is available via a REST endpoint, Workers bindings, or the AI Playground, at provider list pricing without markup.

NVIDIA announced a 64GB unified memory configuration of DGX Spark, available October 23 from Acer, ASUS, Dell, Gigabyte, HP and MSI starting at $4,999. Two units can cluster via NVIDIA Sync Cluster Assistant to pool memory to 128GB and support up to 200-billion-parameter models.

Cloudflare has selected 30 nonprofits for its first non-profit startup cohort, awarding over $7.5 million in credits to civil society organizations building AI and automation tools on its developer services, and is now opening applications for a second cohort.

OpenAI has parted ways with three researchers who allegedly leaked confidential information to an outside AI safety organization, according to the Wall Street Journal. A fourth safety researcher reportedly left shortly afterward.

Robert O'Callahan resigned from Google's GDM team over its work on a new generation of AI chips, the latest departure in what Zvi Mowshowitz describes as an accelerating preference cascade. A Senate Homeland Security subcommittee hearing on rogue AI and new polls showing broad support for slowing development followed.

San Antonio City Councilmember Ric Galvan says data centers in his district have grown from unassuming office buildings into massive hyperscale facilities, with more than a dozen now present and about 20 expected if proposed projects are completed. They cluster along two main roads, concentrating nuisances in the same neighborhoods.

A New York City server says some diners with food allergies have disputed her warnings about dishes, citing ChatGPT instead. She described close calls involving shellfish allergies.

Amazon Web Services CEO Matt Garman published a blog post urging support for AI data center projects, warning that blocking them could harm the US economy and national security. He addressed public concerns about jobs, power demands, and the environment.

A Mercor study found AI models now outperform licensed CPAs on structured accounting tasks in speed and accuracy, while no model fully completes tasks on the larger APEX Benchmark without human oversight.

Databricks published a post titled "Real-Time Retail Intelligence: Building E-Commerce Recommendations with Lakebase and AI Search on Databricks," about building e-commerce recommendation systems on its platform. Only the headline and a brief excerpt about personalization as a revenue engine were available.

OpenAI announced product and developer updates at DevDay 2026, including GPT-6.1 Sol, computer use for the Agents API, cloud-based Codex environments, a Decisions API in limited preview, and new ChatGPT plugin capabilities.

QCon San Francisco 2026 will run November 16–20 at the Hyatt Regency San Francisco, featuring practitioners from Airbnb, OpenAI, Netflix, and Honeycomb discussing production systems as AI agents take on larger roles. Sessions cover Airbnb's AI customer support safeguards, an OpenAI ads case study, Netflix adaptive compression, and Honeycomb's observability MCP.

Suno has launched a speech feature in public beta on its web and mobile platforms that generates spoken voices from scripts or prompted descriptions, and can produce voiceovers and background music at the same time. Chief product officer Jack Brody said music remains central while the company's vision extends to other forms of expression.