22 July 2002

People want to talk to computers - if it's rewarding current [speech] technologies, if deployed in psychologically sound designs, can be sufficiently reinforcing to ensure recurrent, even enthusiastic use. The key is to design interactions that provide a progressively rewarding experience while adequately compensating for their own conversational shortcomings. People will gladly talk to computers, as long as talking to computers remains a rewarding experience. The underlying thought is that same as Walt Tetschner's in It's natural speech, not natural language, stupid!: focus on the interfaces the technology can provide well today, rather than over-reaching. 4:49:29 PM     comments

V-World: it's all in software and hosting Mark Plakias goes to V_World and reports: Speechrec, interface and back-end integration add up to a powerful solution. But costs of speechrec licenses and hardware are holding companies back, so why not outsource to a company that offers managed hosting services, like many telcos and others. 4:48:01 PM     comments

Loquendo perfects speaker verification technology "Perfects": Apparently it's ready and perfected. Rodcorp admit there will be much interest in biometric security, but many deals will fail at the due diligence stage when customers get into the tricky security questions. But at least Loquendo have text-prompted verification... 3:58:19 PM     comments

Listen to email and dictate a reply with Copytalk CopyTalk's service is similar to others in the market but adds in human editors for their dictate-your-reply feature. It's $50/year. We think a reply-by-text-template service might fit the user/use-case model better than dictate-to-humans but good luck to them. 12:43:18 PM     comments