| DAILY NEWS FOR YOU AI / APPLE / SIRI |
| |
They Invented Siri Before Apple Bought It. Sixteen Years Later, They're Still Waiting for the Original Vision. |
| Siri's creators never wanted a voice-controlled alarm clock. They wanted a “do engine” that understood what a person wanted and then handled the task. Apple's new Siri AI is much closer to that idea — but the people who built the original still say the full vision has not arrived. |
| SIRI / ORIGINAL MISSION | | “DO ENGINE” Not just find information.
Understand the request — then complete the task. | |
| Before Siri became the familiar voice built into Apple's devices, it was a small startup created by Dag Kittlaus, Adam Cheyer and Tom Gruber out of years of artificial-intelligence research at SRI International. |
| Their ambition was larger than today's familiar commands for timers, weather or music. Cheyer told CNN that the team thought of Siri as the counterpart to a search engine: Google could help someone find a web page; Siri was supposed to help someone actually get something done. |
| THE ORIGINAL SIRI TIMELINE | | |
| 2007 | SIRI, INC. SPINS OUT SRI turns years of assistant research into a consumer-focused startup. | | FEB 2010 | SIRI APP LAUNCHES The standalone iPhone app could book restaurants, find movie tickets and connect requests to outside services. | | APR 2010 | APPLE ACQUIRES SIRI The purchase comes only weeks after the consumer app reaches the market. | | OCT 2011 | SIRI REACHES THE IPHONE 4S Apple introduces Siri as an integrated iPhone feature. | |
| Siri itself grew out of a much larger research effort. SRI led the Cognitive Assistant that Learns and Organizes project under DARPA's Personalized Assistant that Learns program — a five-year effort involving more than 300 researchers across 22 institutions. |
| CALO / THE RESEARCH GOAL | LEARN | REASON | ACT | EXPLAIN | ADAPT | | |
| The research goal was strikingly modern: build an assistant that could learn from experience, reason about tasks, execute commands, explain what it had done and respond when circumstances changed. |
| ORIGINAL VISION Understand intent
Coordinate services
Handle complex tasks
Act for the user | SIRI FOR YEARS Timers and alarms
Music playback
Basic questions
Device controls | |
| What reached hundreds of millions of Apple users was much narrower. Siri became enormously familiar, but many of its most common uses remained simple commands rather than the broad task-completion model its creators had imagined. |
| WHY THE OLD VISION WAS SO HARD | | |
| 01 LANGUAGE Understand intent | 02 APPS Take action | 03 CONTEXT Know the user | 04 PRIVACY Protect data | |
| An assistant that reliably completes real-world tasks needs more than speech recognition. It has to understand flexible language, retain relevant context, interact with outside software and do all of that without creating unacceptable privacy or security risks. |
| Large language models have changed that equation. Modern AI systems are much better at interpreting broad instructions, maintaining conversational context and breaking complex requests into multiple steps. |
| |
| iOS 27 / NEW CAPABILITY STACK | PERSONAL
CONTEXT | SCREEN
AWARENESS | APP
ACTIONS | WEB
KNOWLEDGE | | |
| Apple's iOS 27 now introduces Siri AI, a rebuilt version of the assistant that can use information across messages, emails, photos and other personal data, understand what is currently displayed on the screen and perform a wider range of actions across apps. |
| It can also search the web for current information and maintain conversational history through a dedicated Siri app that syncs across supported Apple devices. |
| CNN / BETA TEST | | FIND A
DENTIST APPOINTMENT FROM MESSAGES | IDENTIFY A
BAND FROM A TEXT | SAVE A
RECIPE ACROSS APPS | | |
| Those examples are much closer to the original Siri philosophy than asking a voice assistant for the weather. They require the system to understand personal context, locate relevant information and move between services without forcing the user to manage every individual step. |
| CLOSER ≠ COMPLETE THE FOUNDERS STILL DO NOT THINK ANYONE HAS FULLY BUILT THE “DO ENGINE.” | |
| | DAG KIT TLAUS / SIRI CO-FOUNDER “That vision hasn't changed one iota. But nobody's done it yet.” | |
| THE NEXT PROBLEM MAY NOT BE AI | | |
| USER ASKS
SIRI | → | SIRI CHOOSES
THE FLOW | → | SERVICE
COMPLETES TASK | | |
| Kittlaus and Cheyer told CNN that the remaining obstacle may increasingly be commercial rather than technical. If a person asks Siri to book dinner, shop, arrange transportation or plan travel, the assistant could become the interface between the user and the companies providing those services. |
| That creates a new question: how willing are major apps and online services to let an outside AI assistant control the customer experience that their own businesses were built around? |
| THE INTERFACE SHIFT | OPEN APP Find → tap → navigate | → | ASK AI Intent → execution | | |
| 2010 VISION / 2026 INFRASTRUCTURE | | |
| THEN Narrow language systems
Limited app connections
Rigid commands
Smaller compute base | NOW Large language models
Cross-app actions
Personal context
Cloud + device AI | |
| The unusual part of Siri's history is therefore not that its founders imagined something impossible. Much of the industry is now trying to build essentially the same thing. |
| The technology has spent sixteen years moving toward an idea that was already clear before Apple bought the company: computers should move beyond answering questions and begin handling the work behind them. |
| | THE ORIGINAL BET Search engines helped people find information. Siri's founders wanted the next interface to remove the steps between wanting something done and actually getting it done. | |
| 03 | CORE NOTES What sixteen years of Siri actually show | |
| 01 | Siri, Inc. spun out of SRI in 2007, launched its standalone iPhone app in February 2010 and was acquired by Apple roughly two months later. | |
| 02 | The original product concept was a “do engine”: an assistant intended to understand a goal and coordinate the services required to complete it. | |
| 03 | Siri AI in iOS 27 now adds personal context, onscreen awareness, broader knowledge and deeper app actions, bringing Apple closer to that early vision without establishing that the fully autonomous assistant has arrived. | |
| |
| SOURCES They Created Siri Before Apple Bought It. They're Still Waiting for Their Original Vision to Come True — CNN
Siri — SRI International
Artificial Intelligence: CALO — SRI International
Major Updates for Apple's Software Platforms Are Now Available — Apple
Apple Introduces Siri AI, a Profoundly More Capable and Personal Assistant — Apple |
