Slashdot Mirror
Will Apple Lose Siri's Core Tech To Samsung?
An anonymous reader writes Apple bought Siri in 2010, but its core technology is owned by Nuance, maker of Dragon NaturallySpeaking. Now Samsung is looking to buy Nuance. From the article: "This past June, Nuance and Samsung began merger talks, but nothing came of it. At the time, the two companies said talks had 'slowed' due to 'complexities.' But they didn't say it was dead. Guess what? The talks are back on. The first hint came in June, after the company missed the quarterly projections. The Wall Street Journal then brought up the talks with Samsung and also noted the company had taken financial steps that could indicate a buyout was imminent. The company’s earnings report for June stated that Nuance was redeeming $250 million in 2027 convertible notes. By calling back the debt, that would save the future acquirer around $50 million from a debt-to-share conversion."
Can't innovate man. The tech is covered by patents.
You're not allowed to build your own Siri from scratch.
Give me a break, "they do what they do in India Russia and China because it's the right thing to do."
They follow the law over there while trying to make a buck, same as we do here.
Don't blame our people for obeying the law.
There was never a "community feel." You were too young to understand how businesses work.
SIRI does not work well, on the other Google Now works every time even in a noisy coffee shop.
"If any question why we died, Tell them because our fathers lied."
Exactly. This is nothing more than a case of nostalgia for 'the good old days' that never actually existed.
Watt and his competitors were having patent fights hundreds of years ago. Bell and his competitors. Edison. The list goes on.
Well, this article is incredibly misleading.
Nuance is NOT the core-technology behind SIRI.
SIRI's search AI was the core technology.
Nuance was just the convenient front-end (speech recognition) - and that technology is actually very "old" (in internet terms): Mid-1990's.
What Nuance has that its competitors don't have, is a method of using pronunciation guides, which were captured using very painstaking intensive studies of speakers. Variations in individual speech patterns (mannerisms, dialects, accents) have been boiled down to statistical models. These models inform the recognizer, so that "training" isn't necessary, and speech recognition can be more of a plug-n-play thing. It's a huge boost in usability over other speech recognition software. And it required a huge up-front investment. There are few cases where intellectual property is a legitimate idea to protect in software - and this is one of them. Otherwise, there would have been no incentive for the geniuses at Nuance to have come up with this scheme, and invest in the research required to create the statistical models.
If other companies want to build a competing solution for speech recognition, I'm not aware of any encumbrance on the method (software patents are a thing, but this method is not patented afaik) - but they still have to come up with their own statistical models, because Nuance's are protected.
These are my friends, See how they glisten. See this one shine, how he smiles in the light.