<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	xmlns:sy="http://purl.org/rss/1.0/modules/syndication/"
		>
<channel>
	<title>Comments on: TwitterFone: Voice recognition can be dangerous here</title>
	<atom:link href="http://almaer.com/blog/twitterfone-voice-recognition-can-be-dangerous-here/feed" rel="self" type="application/rss+xml" />
	<link>http://almaer.com/blog/twitterfone-voice-recognition-can-be-dangerous-here</link>
	<description>blogging about life, the universe, and everything tech</description>
	<lastBuildDate>Sat, 08 Sep 2012 07:06:53 -0700</lastBuildDate>
	<generator>http://wordpress.org/?v=2.8.4</generator>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
		<item>
		<title>By: tobias</title>
		<link>http://almaer.com/blog/twitterfone-voice-recognition-can-be-dangerous-here/comment-page-1#comment-38681</link>
		<dc:creator>tobias</dc:creator>
		<pubDate>Fri, 23 May 2008 17:35:14 +0000</pubDate>
		<guid isPermaLink="false">http://almaer.com/blog/twitterfone-voice-recognition-can-be-dangerous-here#comment-38681</guid>
		<description>ha ha, i saw your Man U tweet and wondered... now I get it :)
Having spent 3 years in the speech recognition world, i agree that it always seems to be &quot;10 year out&quot;. Quite frankly, it just doesn&#039;t work very well. As you point out, Tell Me and Goog 411 are the right applications - consumer-input-only and non-critical. Anything that converts to consumer-output (e.g. twitterphone) or is critical (i would agrue, depending on the circumstance, a lot of call center automation - especially travel), is a bloody nightmare! I haven&#039;t tested MSFT&#039;s latest and greatest in-car voice rec tech yet. I&#039;d imagine that&#039;s ok (limited set of grammars to recognize - e.g. &quot;turn radio on&quot; - makes the job a lot easier). But, yes, HAL-style VUI is still a way off :)</description>
		<content:encoded><![CDATA[<p>ha ha, i saw your Man U tweet and wondered&#8230; now I get it :)<br />
Having spent 3 years in the speech recognition world, i agree that it always seems to be &#8220;10 year out&#8221;. Quite frankly, it just doesn&#8217;t work very well. As you point out, Tell Me and Goog 411 are the right applications &#8211; consumer-input-only and non-critical. Anything that converts to consumer-output (e.g. twitterphone) or is critical (i would agrue, depending on the circumstance, a lot of call center automation &#8211; especially travel), is a bloody nightmare! I haven&#8217;t tested MSFT&#8217;s latest and greatest in-car voice rec tech yet. I&#8217;d imagine that&#8217;s ok (limited set of grammars to recognize &#8211; e.g. &#8220;turn radio on&#8221; &#8211; makes the job a lot easier). But, yes, HAL-style VUI is still a way off :)</p>
]]></content:encoded>
	</item>
</channel>
</rss>
