The 5 hottest AI features from Amazon’s September hardware reveal (www.techradar.com)

Amazon’s fall hardware reveal centered on Alexa+, a system-level AI layer baked into new Echo speakers and Shows (Echo Dot Max, Echo Studio, Echo Show 8/11) and pushed across Fire TV, Kindle Scribe, Ring/Blink and other devices. The upgraded Echos include more powerful local processing so Alexa+ can track conversations over time (persistent context and memory), carry richer multi-turn dialogue without constant cloud round-trips, and run some inference on device. Fire TV gains scene search that uses semantic video indexing to jump to vague moments (“the big fight”), plus context-aware follow-ups about performers. Kindle Scribe adds handwriting recognition and searchable/summarizable notes without forcing transcription to typed text. Ring doorbells add Alexa+ Greetings, Familiar Faces (on-device facial recognition) and a “Search Party” feature that uses AI object recognition over participating neighborhood cameras to locate lost pets. Technically, Amazon is stitching more on-device ML, cross-device state, and API integrations to make Alexa+ act like an agent: booking reservations, coordinating calendars, and following up across apps using past preferences. That hardware+software push reduces latency and enables proactive automation, but raises usual trade-offs around privacy, data routing and interoperability. For developers and AI practitioners it signals a stronger emphasis on embedded inference, multimodal semantic indexing (audio, video, handwriting) and agent orchestration at consumer scale—while Amazon still competes with Google, Microsoft and OpenAI for assistant dominance.