Karl,
I'll be happy to participate in the new thread; I had already bookmarked it. I had started a similar thread last year, but it got little interest compared to the individual fonix and lernout threads. Perhaps this year, with interest much higher, the thread will be a success.
And hello to Ed; I've occasionally checked in on the fonix thread, which is how I learned of Karl's new thread, so I've seen a couple of your posts.
With regard to the companies that look attractive to me, I think it's no secret that I think LHSPF is the leader, although I think it's too expensive right now to rate as a buy. Still has great upside potential, but also substantial downside risk. Financial considerations aside, here's a few random thoughts:
Seems to me ASR will be integral to a lot of microprocessor controlled applications in the not too distant future, including things which aren't yet microprocessor based (such as your home telephone service, home appliances, etc.) Some of these will be very focused applications, which can use ASR with a limited vocabulary, and be tolerant of errors (if it's too noisy, or you say something nonsense, it will ask you to repeat). This type of ASR, as well as text-to speech, will become a commodity, with thin profit margins, lots of players, no major opportunity. It seems likely that the bigs will either develop the capability in-house (as IBM, Philips, MSFT are doing), partner with another company (as MSFT/LHSPF, SEG/Dragon) are doing, or license or buy out a smaller company's product when the market is a little more mature.
As far as I can tell, the real potential lies in those companies who can integrate a complex product for such tasks as multilingual, multi-user dictation, conferencing, and so on. Here, performance will command a premium price. For this, ASR, translation, noise cancellation, and perhaps other technologies will be vital. LHSPF obviously is developing this way, IBM seems to be also, though they can keep their efforts a little quieter because most of it is in-house; fonix seems to have recognized ASR by itself won't be enough to compete, with its acquisition of AcuVoice, though I suspect there were other, perhaps more important factors operative there as well. Clearly, applications that make the internet more accessible (especially for commerce) are where the volume considerations become massive, especially if you have a multilingual product.
Finally, I think partnerships are vital, unless you have such an outstanding and unique product that everyone has to play ball with you (as MSFT has had with their Windows OS). Right this moment, I think LHSPF is in that position with its linguistics capabilities, as I expressed in a post which someone copied on the Yahoo fonix thread - I think MSFT needs LHSPF more than the other way around, for the moment. As long as LHSPF can do multilingual much better than anyone else, they'll be hard to beat, even if someone else's ASR or TTS works a little better. On the internet, when writing software, when building a product that you want to sell and service internationally - English only won't cut it for much longer - even for those if us for whom it's the native language. Perhaps I'm overemphasizing the importance of this factor, because it obviously plays in favor of LHSPF, but I'm open to other opinions!
Bob |
