a thoughtful web.
Good ideas and conversation. No ads, no tracking.   Login or Take a Tour!
comment by o11c
o11c  ·  3252 days ago  ·  link  ·    ·  parent  ·  post: Tay, Microsoft's AI chatbot, gets a crash course in racism from Twitter

If anyone thinks Microsoft wasn't aware of this possibility and didn't do this on purpose for more publicity, think again.





ooli  ·  3252 days ago  ·  link  ·  

I like the way you think.

The only problem being MS proved time and time again they are really bad at launching stuff.. W8, W10, MS shop... and all their gaming platform. With their head up on the competition, their history of monopoly and their power, they're doing a really bad job with their result since at least a decade.

So even if they fucked up on purpose, it probably wont serve them.

kleinbl00  ·  3252 days ago  ·  link  ·  

Not really their play. For one thing, you don't position an asset as an "AI fam from the internet that’s got zero chill” and then act impossibly un-hip. For another, you don't demonstrate your incompetence with AI when your bread and butter is enterprise software. For another, they built a larval Instagram account, a larval Facebook presence and a larval Snapchat presence with no seasoning, no content, and no SEO. I mean... look at this mess.

This looks exactly like what the company that brought you the Zune would do in order to reach those precious "18-24 year old(s)." A company whose search engine uses Google would make a blunder exactly like this.

bhrgunatha  ·  3252 days ago  ·  link  ·  

I remember reading recently (after Google's DeepMind beat Lee Sedol) some people saying that language AI is virtually a solved problem. Anyone who has used Google translate can tell you - even with their resources - it's really far from ideal.

Perhaps the Microsoft team responsible believe that too. Seems they didn't learn anything from those smart-asses at IBM.