3 ms·

Speech to text doesn't require AI. I've had it on my Google keyboard for a very long time. Its always been accurate too.

TeMPOraL 2y ago
In my case, it always been garbage. Speech to text doesn't require AI when you do it like Microsoft did it in the aughts; ever since "new, better", cloud-side techniques came along, the technology got worse, and the only meaningful qualitative improvement I've seen in two decades is in the last two years, with new-generation models which may or may not be backed by LLMs now.
- happymellon 2y ago
  Dragon dictate was crappy. Google keyboard was pretty precise. Over time its become worse, but all of Googles products have become worse. That doesn't mean it we couldnt do it, we just can't rely of Google to provide it.
ben_w 2y ago
> Speech to text doesn't require AI. *grumpy old man shakes fist at ever-changing meaning of words* https://en.wikipedia.org/wiki/AI_effect https://en.wikipedia.org/wiki/AI_effect https://www.kurzweiltech.com/kai.html https://www.kurzweiltech.com/kai.html
- happymellon 2y ago
  They didn't call it AI previously. Are we redefining STT as AI?
  - ben_w 2y ago
    > They didn't call it AI previously. "In this excerpt from The Age of Spiritual Machines (Viking, 1999), Ray Kurzweil describes his work in speech recognition." - previous link "The ‘80s saw speech recognition vocabulary go from a few hundred words to several thousand words. One of the breakthroughs came from a statistical method known as the “Hidden Markov Model (HMM)”. Instead of just using words and looking for sound patterns, the HMM estimated the probability of the unknown sounds actually being words." - https://sonix.ai/history-of-speech-recognition https://sonix.ai/history-of-speech-recognition "Voice Recognition", title of page 82, "Who's who in Artificial Intelligence: The AI Guide to People, Products, Companies, Resources, Schools and Jobs" - By Alan Kernoff, 1986 > Are we redefining STT as AI? The opposite, you're redefining it as not AI.

TeMPOraL 2y ago

In my case, it always been garbage. Speech to text doesn't require AI when you do it like Microsoft did it in the aughts; ever since "new, better", cloud-side techniques came along, the technology got worse, and the only meaningful qualitative improvement I've seen in two decades is in the last two years, with new-generation models which may or may not be backed by LLMs now.

happymellon 2y ago
Dragon dictate was crappy. Google keyboard was pretty precise. Over time its become worse, but all of Googles products have become worse. That doesn't mean it we couldnt do it, we just can't rely of Google to provide it.

ben_w 2y ago

> Speech to text doesn't require AI. *grumpy old man shakes fist at ever-changing meaning of words* https://en.wikipedia.org/wiki/AI_effect https://en.wikipedia.org/wiki/AI_effect https://www.kurzweiltech.com/kai.html https://www.kurzweiltech.com/kai.html

happymellon 2y ago
They didn't call it AI previously. Are we redefining STT as AI?
- ben_w 2y ago
  > They didn't call it AI previously. "In this excerpt from The Age of Spiritual Machines (Viking, 1999), Ray Kurzweil describes his work in speech recognition." - previous link "The ‘80s saw speech recognition vocabulary go from a few hundred words to several thousand words. One of the breakthroughs came from a statistical method known as the “Hidden Markov Model (HMM)”. Instead of just using words and looking for sound patterns, the HMM estimated the probability of the unknown sounds actually being words." - https://sonix.ai/history-of-speech-recognition https://sonix.ai/history-of-speech-recognition "Voice Recognition", title of page 82, "Who's who in Artificial Intelligence: The AI Guide to People, Products, Companies, Resources, Schools and Jobs" - By Alan Kernoff, 1986 > Are we redefining STT as AI? The opposite, you're redefining it as not AI.