Services & Product Information– category –

On October 21, WingArc1st began rolling out DataTetra, a data platform designed for enterprise AI use. This article examines its design philosophy and pricing structure to explore how companies can address costs that grow as they use more AI and dependence on specific services.

The 117-minute feature-length anime The Song of the Stars Can Be Heard, completed by a three-person team in about three months of hands-on work and comprising around 1,500 cuts, is available for free on YouTube. The numbers offer a glimpse into the role people played in making a feature-length work in the generative AI era.

Three models, including OpenAI’s GPT-6 Astra, have been added to Google Cloud’s Model Garden. Here’s a look at how they differ from API access and what companies need to know about data handling.

AI startup Inception has announced Mercury 3, a diffusion language model that it says can generate 1,068 tokens per second and ranks first in 8 of 12 categories. Results vary by use case. Here's a look at how speed can change the user experience.

Claude’s official account announced a research preview of Claude Duet, an agent app designed to act proactively. This article examines what the next generation of AI apps might look like through four elements of its design: surfacing matters requiring attention, carrying out requested tasks, continuing work while users are away, and seeking confirmation before acting on their behalf.

Cloudflare has previewed Clairic, a document-extraction model, on Workers AI. It returns specified fields such as order numbers and invoice amounts in a single pass. We unpack the design philosophy behind its 19.3-second processing time when working alongside an LLM.

The anonymous AI model “Space Bunny,” which drew attention on leaderboards, has been revealed as MiniMax’s new LLM, M3.1 Flash. The company is emphasizing its low price and speed, while its benchmark rankings vary by area. Here are the key points to keep in mind ahead of the October 23 release of the weights.

AWS has announced a new feature that lets Node.js apps such as Next.js and Hono run directly on Lambda. We look at what doing away with the adapter means, the execution environment specifications and the impact on serverless options.

YouTube Music is introducing labels for songs made using generative AI. They will appear on the playback screen and extend to the main YouTube app. The labels will show AI’s involvement in three categories, setting YouTube’s emphasis on disclosure over restrictions apart from services such as Deezer.

AI-written code is not necessarily something you can fix yourself, even if it passes tests. In the first experiment, a developer who rated their understanding a 4 out of 5 stopped after 8 minutes and 12 seconds without finding a planted bug. The result raises questions about what it means to “own” code.

Meta has announced “Muse Spark Open,” an open-weight model with 816 billion parameters and a context window of 1 million tokens. It is designed to run on Macs, though it reportedly requires at least 1TB of memory. We look at where open-weight AI could be useful, given the gap between the model’s performance and the hardware needed to run it.

Amazon-owned Bee has announced two AI devices, Bee Frontier and Bee Eclipse, that summarize conversations and turn them into calendar events and tasks. Priced at $74.99 and $99.99, respectively, they point to a vision of personal AI that handles everything from recording to execution—and raise privacy questions about entrusting it with conversations.