#Other

AIAlexander Doria reposted a post with no added commentary, and a quoted post argues Google's products are consistently strong yet always feel slightly off or unfinished.

AIMatt Shumer shares a four-minute recap of the seven AI stories that mattered in the week of September 25–29, linking to a full write-up on somethingbig.ai. The post itself gives no details on the individual stories, so the specific developments are only available through the linked article.

AIBiomedical datasets are often incomplete, such as patients with imaging but no genomic data. Stanford speaker Olivier Gevaert will discuss using foundation models to fill these gaps in multimodal precision medicine modeling on October 7.

AIAnt Ling asked developers what they will build with Ling-3.1-flash and invited them to share demos and feedback in its official Discord server. The post gives no specifications, benchmarks, pricing, or release details.

AISGLang announces early bird pricing for its summit, with up to 30% off through October 20. The Group Pass costs $199 per seat, according to the post.

AIMuse is promoting new connectors for small businesses, asking users where they would start among content planning, customer research, sales insights, and managing inventory. The post frames these tools as a way to let Muse handle some of the heavy lifting that owners would otherwise do themselves.

AIOpenAI CEO Sam Altman praised Ultrafast as so fast that he does not want to return to previous speeds. The post gives no further details about the product, its performance figures, or its availability.

AIHamel Husain recommends Simon Willison's live blog as the fastest way to catch up on today's launches. Willison is covering OpenAI's DevDay keynote in San Francisco, posting live updates during the event.

AIAnthropic is running a new study with Anthropic Interviewer from Sept 29 to Oct 6, open to Free, Pro, and Max users on Claude and Claude Code. Participants can choose to make their responses public, and the findings will shape The Anthropic Institute's research and Anthropic's decisions.

AIAnthropic says making an interview public is completely optional for participants. Its blog post covers the possible benefits and risks of publishing such interviews.

AISimon Willison is attending OpenAI's DevDay event in San Francisco and will post live updates during the keynote, which starts in five minutes. This marks his third consecutive DevDay live blog, and the post links to his live blog page.

AIIndependent filmmaker Jeff Synthesized won the Future Vision XPRIZE grand prize for The Gifted, a film about an 11-year-old who recreates his late mother's voice and essence using code. The solo-developed project was chosen from more than 2,500 worldwide entries and receives $100,000 plus $2.5 million in feature production funding. Google is partnering with Range Media Partners through its 100 ZEROS initiative to bring the film to the big screen.

AIJosh Woodward, Google's Gemini leader, called a Yosemite-themed browser project "awesome" after returning from the park. The project, by @trondw, is a three.js and WebGL photography simulator featuring real sun and Moon positions, real lidar data, and a tripod-mounted film camera for virtual shooting at Yosemite locations.

AIMark Zuckerberg announced that Chirantan "CJ" Desai will join Meta as Chief Enterprise Platform Officer, reporting directly to him. Desai previously served as CEO and President of MongoDB, led product and engineering at Cloudflare, and spent nearly eight years at ServiceNow, including as President and COO. The post describes his experience across AI, infrastructure, business applications, and security as the basis for leading Meta's enterprise effort.

AISemiAnalysis tore down Intel's Panther Lake chip, examining its 18A process with RibbonFET gate-all-around transistors and PowerVia backside power delivery. Measurements put 18A compute logic and TSMC N3E GPU logic at similar logic density, though 18A does not lead TSMC N3P, N2, or Samsung SF2 in peak density. The analysis also compares the compute, GPU, I/O, and Foveros-S packaging against Lunar Lake and Samsung's SF2 process.

AIThe post from Microsoft's official account contains only a link with the text "Learn more," so no specific product, feature, or figures are stated. The linked destination is not described in the source, so no further details can be supported.

AILatent Space, the AI newsletter and podcast, has announced plans for AINews v3 and a new home for the publication. The company says it is open for business, with Supabase as its first sponsor.

AIAnthropic's team enabled users to use their connectors in a shared environment, with access now scoped to individual users. Admins no longer need a channel-wide identity to set up Tag, making it easier to configure and more useful.

AIGoodfire researchers found that Llama 3.1 8B solves month and day arithmetic, such as six months after August, through a single addition module in layer 18. The module represents numbers as Fourier-feature circles and computes modular sums in parallel, and steering those circles changed the model's predicted month.

AIGoodfire Research examines how sparse autoencoder directions relate to curved manifolds in neural representations, identifying shattering, compact capture, and dilution as three ways lines can represent them. The team trained an autoencoder on synthetic data containing shapes such as donuts, spheres, and Möbius strips, and reports that real features in Llama 3.1 8B show dilution. It also describes an unsupervised pipeline that clusters features by firing patterns to surface manifolds in that model.

AIJeff Dean praises Alex Ksendzovsky and team at BioComputing Co for work that lets researchers learn from computations running in real rat neurons. The quoted post from Ksendzovsky says the company is partnering with Amazon to bring the technology to customers and is taking early-access requests.

AIMeta is extending its Private Processing confidential computing infrastructure to AI glasses, running AI models inside confidential virtual machines so that even Meta cannot access user data. The system relies on hardware Trusted Execution Environments, with remote attestation checked by clients before any data is sent. Meta first introduced Private Processing in 2025 for WhatsApp and the Meta AI app.

AIX's Grok Bot account shares a download link to Grok Bot at The post provides no further details about the product's features, version, or pricing.