News
What's happening in robotics and AI — curated by the wujec.ai editors.
AWS took the promise that a Bedrock model lives a year out of the rulebook — and put it in the model card
Amazon has split the rules that decide how long a model on Amazon Bedrock is guaranteed to answer. Models launched before 7 September 2026 keep the old policy; everything launched on or after that date falls under a new one, and the new one no longer contains the two sentences customers used to plan around. The old page — still live, now titled "Model lifecycle (Legacy)" — makes two platform-wide promises. First: "Once a model launches on Amazon Bedrock, it will remain on Amazon Bedrock for at least 12 months before the EOL date." Second: "A model will be in the Legacy state for at least 6 months before the EOL date" — that is, at least half a year of notice before the switch-off. Neither sentence appears on the new page. What appears instead is a per-model arrangement. The new policy says every model card shows two things: an "EOL no sooner than" date, and the Legacy period, meaning the notice given before end of life. And the notice period is no longer one number: "There are two Legacy periods: 6 months and 45 days. Most models have a 6-month Legacy period." A 45-day wind-down is now written into the rules as an option a provider can take. So far nobody has taken it. We read nine model cards across Amazon, OpenAI, Anthropic, xAI, Z.ai, MiniMax, Moonshot and TwelveLabs, including the newest arrival on the platform: GPT-6 Astra, launched 8 September 2026 — one day after the cutover, and therefore among the first models governed by the new policy. Its card reads "EOL no sooner than: September 8, 2027" and "Legacy period: at least 6 months". Twelve months of life, six months of notice: exactly the old terms, now stated for one model rather than promised to everyone. One clause went the customer's way. The old policy describes a "public extended access" phase: after at least three months in Legacy, a model enters a stretch in which it keeps running but, as AWS puts it, "you should expect higher pricing, which will be set by the model provider". The new policy does not mention public extended access at all. The last months of a model's life were the leg where the bill could rise; that leg is not in the new rules. The practical consequence is small today and structural tomorrow. Anyone building on a hosted model has to answer one architectural question — how long will this thing keep answering — and until now Bedrock answered it once, for the whole catalogue. Now the answer sits in each model card, set per model, and a provider who wants to move fast has a 45-day door available. The number to check is no longer in the terms; it is on the card. The old rules, meanwhile, are still doing their work. On 14 September 2026 Bedrock switched off Amazon Nova Sonic, the company's first speech-to-speech model, and Amazon Nova Premier, its first frontier-class model and the designated teacher for the cheaper Nova line. Both were moved to Legacy on 13 March 2026 — six months' notice to the day. Nova Canvas and Nova Reel, Amazon's image and video models, go on 30 September, and Amazon has named no successor for either.
Amazon Nova Premier →Amazon says drones will reach nearly 500 US towns this year — that is a sixfold jump in place names, not in launch sites
On 19 August 2026 Amazon said Prime Air would serve "nearly 500 cities and towns" in the United States by the end of the year. Today the service flies from 11 locations in 10 metro areas across seven states — Arizona, Florida, Kansas, Louisiana, Michigan, Nebraska and Texas. New sites were named in the Chicago, Atlanta, Cleveland, Syracuse and Boise metro areas. The two numbers are not measured in the same unit. A Prime Air site works a radius of about 7.5 miles, which is roughly 175 square miles of ground, and an American metropolitan area of that size routinely contains dozens of separately incorporated cities, towns and villages. Adding a handful of sites therefore adds hundreds of place names. "We've already delivered hundreds of thousands of packages to customers by drone this year," said David Carbon, vice president of Prime Air — a figure that, spread over a year and 11 sites, describes a working service rather than a mass one. The aircraft doing the flying is the MK30, in customer service since November 2024. It lifts off on six motors, transitions to wing-borne cruise at up to 73 mph, flies up to 400 feet above ground and returns to base within a 7.5-mile radius. Its maximum take-off weight is 83.2 lb, of which the customer's parcel may be no more than 5 lb — a ceiling Amazon says still covers about 60% of what its customers buy most often. The constraint that decides whether the target is met is not the drone. Prime Air holds an FAA Part 135 air-carrier certificate and flies beyond visual line of sight, but each new site needs its own regulatory clearance and environmental review before a single package leaves the ground. Amazon has published a target measured in towns; what is worth watching between now and December is the number measured in sites.
Amazon Prime Air MK30 →Amazon's newer voice model listens 12% cheaper — and writes 11 times dearer
Amazon never published a price comparison between its two speech-to-speech models, but the Bedrock price list does it for anyone who reads all four rates instead of the headline one. In the us-east-1 list dated 13 August 2026, Nova 2 Sonic charges $3.00 per million speech input tokens and $12.00 per million speech output, against $3.40 and $13.60 for the original Nova Sonic. That is a cut of 11.8% on both sides of the audio. The text rates moved the other way, and much harder. Text input went from $0.06 to $0.33 per million tokens — five and a half times more — and text output from $0.24 to $2.75, a rise of 11.5 times. Both models bill audio and text separately, so a session's bill depends on how the two mix. That mix has an arithmetic break-even. Moving to Nova 2 Sonic saves $1.60 per million speech output tokens and adds $2.51 per million text output tokens. The newer model is therefore cheaper only while text output stays below roughly 64% of speech output volume. A model that mostly talks stays cheap; one that also returns transcripts, tool calls and structured replies crosses the line. What this does not prove: we have not measured how either model actually splits its tokens in a real session, and Amazon publishes no such breakdown, so the 64% figure is a property of the price sheet, not a verdict on any application. Regional lists other than us-east-1 may differ, and both sets of rates can change without notice.
Amazon Nova 2 Sonic →Amazon puts end dates on four of its own Nova models — and bets the house on one frontier system
Amazon is winding down most of the Nova family it launched barely eighteen months ago. Its own lifecycle table for Amazon Bedrock now carries end-of-life dates for four Nova models: Nova Premier, the family's largest, and Nova Sonic, its speech model, both stop working on 14 September 2026, while the image model Nova Canvas and the video model Nova Reel follow on 30 September 2026. All four entered the Legacy state in March 2026, which under Amazon's own rules already blocks new customers and new fine-tuning jobs. The strategy behind the dates has been reported rather than announced. Amazon is said to be folding its scattered specialist models — separate systems for text, image, audio and video — into a single frontier-scale model intended to compete directly with OpenAI, Google and Anthropic, with a debut targeted at re:Invent in late 2026 and the Nova name likely retained. The work sits with the Frontier Model Research team that Pieter Abbeel was brought in to lead in December 2025. Amazon has confirmed that Nova 2 Lite stays. What makes this unusual is not the retirement itself but who is doing it. Amazon is the cloud provider hosting everyone else's models, and the same lifecycle page that ends Nova Premier also schedules Anthropic's Claude Sonnet 4 for 14 October 2026, Cohere's Command R and Command R+ for 19 August 2026, and AI21's Jamba 1.5 pair for 26 November 2026. A company that sells shelf space has decided it would rather stock one strong own-brand product than five weak ones. For anyone building on Bedrock the practical point is the calendar. Legacy models do not fail politely: after the EOL date requests simply stop working, and Amazon states plainly that migration does not happen automatically.
US appeals court: an AI agent is a tool, not a trespasser
On 4 August 2026 the US Court of Appeals for the Ninth Circuit vacated the injunction that had blocked Perplexity's Comet assistant from logging into Amazon accounts on a user's behalf, in Amazon.com Services LLC v. Perplexity AI (No. 26-1444). The lower court had granted that injunction on 9 March 2026. The reasoning matters more than the outcome. Amazon had argued that Perplexity gained unauthorised access to its servers under the Computer Fraud and Abuse Act — the 1986 anti-hacking statute. The appeals court disagreed on who was doing the accessing: the assistant is a tool, not a person for statutory purposes, and it is the user, logging in with their own credentials, who reaches Amazon's servers. The panel acknowledged there is "little to no existing caselaw directly dealing with how to ascribe responsibility for AI agents", let alone under the CFAA, and applied the rule of lenity — where a criminal statute is ambiguous, it is read narrowly. This is the first appellate answer to a question every agentic product now runs into: when software acts on your behalf inside your own account, is that you or the vendor knocking on the door. The answer here is "you". Amazon's trademark and state-law claims survive, so the dispute is not over, and a decision from one circuit does not settle the country — but automation vendors now have something to point at. wujec.ai does not yet have a profile for Perplexity or its Comet browser assistant; that gap is on our list.