
Google DeepMind
Google DeepMind is an AI research and development company based in London.
- 3k

Google DeepMind is an AI research and development company based in London.
deepmind.googleCrawled Oct 15, 2026
Researchers are designing more effective drugs with AlphaFold. Karen Akinsanya is President of R&D, Therapeutics, at Schrödinger in New York City.
The page states that the world produces about 400 million tonnes of plastic waste each year, much of which ends up in landfills, while a significant portion pollutes the oceans.
The company is beta launching SynthID, a tool developed with Google Cloud to watermark and identify AI-generated images, to a limited number of Vertex AI customers.
Collaborate with leading thinkers at Google DeepMind. Build AI that benefits humanity.
The page says AI research and technology are advancing exponentially.
The company researches and builds safe artificial intelligence systems.
Google DeepMind’s **Gemini Omni 1.2 Flash** now generates AI video clips up to **15 seconds long** (up from 10 seconds) and reduces pricing by **30%** to **$0.07 per second** for 720p. The model introduces **Conversational Editing**, allowing incremental adjustments like background or lighting without full regeneration. It also improves **reference adherence**, maintaining consistency in faces, clothing, and brand colors using input images or motion clips. Visual stability and natural dialogue—including lip-sync and multi-speaker distinction—are enhanced, alongside support for complex camera movements like drone footage.
Nov 2026The DecoderNews article
Google DeepMind employees protest a contract clause imposing a lifetime ban on criticizing the company. Workers argue the policy discourages open discussion about AI risks and ethical concerns. The restrictions reportedly apply to all staff, limiting transparency around technology development.
Nov 2026Washington PostNews article
The first TPU in space is running a Google DeepMind Gemma 4 model and writing cheesy poetry.
Oct 2026TechCrunchNews article
Google’s **Gemini 4 Carbon** outperforms competitors on most enterprise AI benchmarks, including a **9-point lead** on the Artificial Analysis Intelligence Index (59 vs. 58 for Claude Opus 5.5). It excels in **cybersecurity vulnerability detection** (+12.8% on ExploitGym) and **complex software engineering** (+5.3% on SWE-sweep), though it trails OpenAI’s GPT-6 Astra in some coding and computer-use tests. Access remains restricted to Google’s **Fairwind Program** for cybersecurity defenders, with broader release pending feedback. Pricing matches Argon’s introductory rates ($2M input, $10M output tokens), but lacks a defined end date for discounts. Argon remains the publicly available model for now.
Oct 2026VentureBeatNews article