Twitter | Search | |
will knight
I write about mechanical brains for MIT Tech Review will@technologyreview.com
3,673
Tweets
3,793
Following
11,119
Followers
Tweets
will knight Jan 15
Definitey going to use this excuse when missing a deadline
Reply Retweet Like
will knight Jan 11
Replying to @0x00B1 @techreview
It's a secret.
Reply Retweet Like
will knight Jan 11
Will do.
Reply Retweet Like
will knight retweeted
wilhelm nylund Jan 11
Reply Retweet Like
will knight Jan 10
I found what must be one of the only LISP machines still in operation at MIT (note: it is preferred to a newish Mac) viva
Reply Retweet Like
will knight Jan 10
We are looking for a new group of "innovators under 35" to highlight in . Submit your favorite 🤖 and 🤖 people here:
Reply Retweet Like
will knight Jan 10
GANs can figure out categories of physical object as well as rules about how they should appear--cool story by mit-ibm project page:
Reply Retweet Like
will knight retweeted
Jeremy Howard Jan 2
It was a surprising claim from that AI ethics does not have "huge life and death issues yet that need to be urgently solved". This is not correct. itself has been involved in many life and death issues, eg:
Reply Retweet Like
will knight Dec 30
Good points, Thomas. I know people are trying to apply RL in lots of domains nowadays. Interesting argument from Russell.
Reply Retweet Like
will knight Dec 30
Good question, Mark. Perhaps we should try to replicate the simple things that humans can do and computers can't, rather than compete with what are essentially unusual feats of human intelligence (eg chess expertise)?
Reply Retweet Like
will knight Dec 30
Replying to @rachelmetz @jsomers
The emoji equivalent of a terminator pic!
Reply Retweet Like
will knight Dec 30
Replying to @halhod @jjaron and 4 others
I wouldn't say that at all. It just seems like an unnecessarily simple way to define intelligence. BTW, I don't know enough about cybernetic, but it seems sort of relevant to Weiner's focus on feedback loops etc.
Reply Retweet Like
will knight Dec 30
Replying to @halhod @jjaron and 4 others
what Jacob said…
Reply Retweet Like
will knight Dec 30
I think there's more to it than that, Mark. Perhaps a bigger question is how much these game playing algorithms constitute real process towards replicating the only example of AGI we know of...?
Reply Retweet Like
will knight Dec 30
An good and often overlooked point regarding AI game playing in ⁦⁩ nyorker peice:
Reply Retweet Like
will knight Dec 16
Replying to @mims
Yeah, I've often wondered about the surveillance potential of self driving cars. It would presumably be pretty trivial to develop algorithms that automatically make a note of bad driver behavior.
Reply Retweet Like
will knight retweeted
MIT Technology Review Dec 15
These microchips China wants to design aren’t your traditional chips. They are key to new weapons systems, better cryptography, and more powerful supercomputers. by
Reply Retweet Like
will knight Dec 14
The quantum trolley problem... (attn: )
Reply Retweet Like
will knight Dec 14
Replying to @rachelmetz
ha, okay. was it about a feline lazy-boy? (a *great* idea).
Reply Retweet Like
will knight Dec 14
Replying to @rachelmetz
You know I have a cat called Sausage.
Reply Retweet Like