GoKawiil
Tech News
A new project called 'How Stale Is Your AI?' compiles release dates and training cutoffs for 20 models from 8 major labs, sorted from stalest to freshest. It finds that only 9 of the 20 models have a training cutoff date that is actually published by their maker, with data also offered as a downloadable models.json file.
Mozilla's latest State of Open Source AI report, using data through September 1, finds top Chinese open-weight models are closing in on closed US frontier systems, lagging by only a few percentage points on key benchmarks while costing far less to run. The report estimates the capability gap at roughly 4.4 months based on METR task-horizon data, close to Epoch AI's independent four-month estimate, and notes one Chinese model scored within a point of a leading Claude model on Terminal-Bench at under a fifth of the price.
Chinese filmmakers are increasingly turning to AI tools like ChatGPT, Kimi and DeepSeek to draft screenplays and generate footage, with one studio producing a 90-minute AI-assisted film for roughly $500,000 versus the millions typical productions cost. Director Lu Beike, known for the series 'Three-Body,' says AI-generated scenes still fall short of live performances, sometimes forcing crews to make real actors look less realistic so they blend with AI-generated environments.
Leading AI developers are now using their most advanced models to help design and accelerate the next generation of AI systems, according to researchers tracking the industry's progress. This shift is prompting fresh warnings that the pace of development could compound on itself, moving faster than safety and alignment work can keep up with.
Major AI labs have moved their attention from building ever-larger models to running inference—the process of using trained models to generate text, code, and images. This shift is driven by growing real-world use of large language models, the rise of reasoning models that repeatedly reprompt themselves, and autonomous AI agents that run continuously rather than just responding to single queries. Amazon Web Services, for instance, has split inference tasks between its Trainium chips and Cerebras's wafer-scale hardware.
A new State of Open Source AI report from Mozilla, shared with Ars ahead of its September 15 release, finds that top open-weights models from Chinese developers now perform nearly on par with closed frontier models from US firms like Anthropic and OpenAI. It cites Moonshot AI's Kimi K3 scoring only three points below Anthropic's Fable 5 on the Artificial Analysis Intelligence Index, despite costing roughly 30 percent as much to run.
Climate scientist Zeke Hausfather analyzed his own usage of Anthropic's Claude Code assistant and calculated that a typical day of work consumed between 1.2 and 5.9 kilowatt-hours of electricity, comparable to running two refrigerators nonstop for a day. He found that 96 percent of the roughly 3.2 billion tokens processed over eight weeks came from the agent repeatedly re-reading its own accumulated memory at each step, rather than generating visible output.
Broadcom CEO Hock Tan told CNBC that his company's AI semiconductor revenue forecasts remain unchanged despite market jitters over Anthropic CEO Dario Amodei's call to slow frontier AI development. Broadcom shares fell 4.8% and chip stocks broadly dropped after Amodei's comments, which drew backing from Sam Altman and Elon Musk. Tan said demand for AI compute infrastructure remains strong and durable, and reaffirmed that Anthropic is set to become Broadcom's largest custom chip customer by 2027.
A leaker known as pdfu shared iOS 27 and macOS Golden Gate framework code showing Siri AI can hand off tasks to Anthropic's Claude or OpenAI's GPT-5.6 models, with Siri acting as an intermediary for system-level access. One demo even shows GPT-5.6 fully replacing Siri as the on-device agent, searching emails and sending iMessages while logging activity in OpenAI's interface.
With today's iOS 27 rollout of Siri AI, Apple has replaced its old 'Improve Siri and Dictation' opt-in with a new setting called 'Improve Siri and Apple Intelligence.' Choosing 'Share Audio and Text' now lets Apple use voice and text interactions from Siri, Dictation and Translate to train its Foundation Models, not just refine Siri as before. Even users who previously consented will be prompted again during setup to confirm their choice under this expanded scope.
Alongside its new macOS 27 Golden Gate release, Apple has pushed out macOS Tahoe 26.7 and macOS Sequoia 15.8 as updates for users who haven't moved to the newest OS. Apple describes both as containing important security fixes recommended for all users, though it has not yet published detailed release notes on its security page.
An Entrepreneur contributor and founder of roof restoration company Roof Maxx argues that rapid AI-driven change has eliminated the possibility of a stable, 'set-and-forget' business strategy. He contends leaders should replace the pursuit of security with cultivated adaptability, using productive paranoia to question their own assumptions before overhauling their business model.