The 30-Year AI Scare Cycle
Artificial intelligence might seem like Silicon Valley'southward latest obsession, but it'due south nothing new. I went through AI mania the start time around 1981, when robots and AI were all anyone talked about. This continued until the middle of the decade and died for expert when the Japanese economy tanked and took with it the so-chosen Fifth Generation Calculator.
HTML MODULE 4010
I wouldn't be surprised if the same thing happened again. Though Elon Musk is concerned that AI could rising up and kill everyone nether some imaginary circumstance, he's really only been reading too much science fiction.
Currently, AI can barely practice optical character recognition; I have however to encounter text analysis worth a powder. Until AI tin translate French to English language flawlessly, I'm not buying any of information technology. Hitting 99 percent is not good enough, not in OCR or AI. I'k reminded of the airline that fabricated the proud claim that 99 percent of all baggage got to its destination as promised. But that means three people on a plane with 300 passengers routinely lose their luggage. That's t errible.
One of the bulwarks of the electric current AI craze is IBM'due south Watson. I've discussed it in the past, but in the 1980s, the same AI engineering science was all over the place. It was called an "expert arrangement" dorsum then and was supposed to, amidst other things, perform medical diagnosis—only like Watson.
Lately, every new startup pitch mentions AI. For case, if I went out looking for venture money, I'd put together a concern programme for a new electronic mail customer that took advantage of both Popular and IMAP using AI to categorize, prioritize, and sort messages. The AI component would kick spam into its ain bin. Information technology would quarantine all links and attachments into a sandbox where everything would be opened and examined for ransomware and trojans.
Does any of this require actual AI? Or is this just clever coding of countless IF-THEN statements that I just decide to call AI?
A genuine artificial intelligence should spend its spare time tracking down people who send email viruses, independently call the FBI, and report the evil-doer without my cognition. This not happening anytime soon.
The AI mania of the 1980s faded considering it produced nil of any commercial value. Afterward the collapse, everything AI was spun downward, forgotten, and discarded, because zilch had come from information technology. A similar excitement happened thirty years earlier in 1956 with the invention of IPL (Information Processing Language), which chop-chop morphed into LISP. This could easily be a 30-year cycle. The next appearance of this nonsense should be around 2045.
Equally a cycle, it's hard to take seriously until there is a true quantum, which I have yet to see. In the meantime, don't fret over AI "taking over" and beingness much of a threat to humanity. Simply maybe next fourth dimension!
Nearly John C. Dvorak
Source: https://sea.pcmag.com/opinion/17736/the-30-year-ai-scare-cycle
Posted by: lucaswastor.blogspot.com
0 Response to "The 30-Year AI Scare Cycle"
Post a Comment