Curating Content To Support Learning About Humanity's Transition

Posts tagged with:  Jim Rutt - Twitter

By: The Posts Author | Posted on: 31 Mar 23

@th0masschindler actually I did! I feeling kind of GPTish this afternoon.

@th0masschindler actually I did! I feeling kind of GPTish this afternoon.


By: The Posts Author | Posted on: 31 Mar 23

@speakerjohnash I dunno, it might be interesting to take everything I interact each day, index it into a latent sematic space, cluster it into topics, and have tools for browsing. Also some quality filters and such. And LLM summaries. etc.

@speakerjohnash I dunno, it might be interesting to take everything I interact each day, index it into a latent sematic space, cluster it into topics, and have tools for browsing. Also some quality filters and such. And LLM summaries. etc.


By: The Posts Author | Posted on: 31 Mar 23

@speakerjohnash We get very little traffic there except occasionally when an episode goes a little bit viral.

@speakerjohnash We get very little traffic there except occasionally when an episode goes a little bit viral.


By: The Posts Author | Posted on: 31 Mar 23

@wolfejosh Whatever became of Bill Joy?

@wolfejosh Whatever became of Bill Joy?


By: The Posts Author | Posted on: 31 Mar 23

@speakerjohnash no video, though.

@speakerjohnash no video, though.


By: The Posts Author | Posted on: 31 Mar 23

@JeffLadish Kind of bogus. There is no chance of a Yudkowskian AGI risk from a feedforward neural net like the LLMs. Bad faith for him to lead people to believe there is when knows there isn’t.

@JeffLadish Kind of bogus. There is no chance of a Yudkowskian AGI risk from a feedforward neural net like the LLMs. Bad faith for him to lead people to believe there is when knows there isn’t.


By: The Posts Author | Posted on: 31 Mar 23

@IntuitMachine I think “magic words” are a subset of “skyhooks” which us where we get GPT to as much as possible of the work building the context. Many of the most useful require moving output to an external system which then processes it (maybe using GPT) before returning it tomGPT as part…

@IntuitMachine I think “magic words” are a subset of “skyhooks” which us where we get GPT to as much as possible of the work building the context. Many of the most useful require moving output to an external system which then processes it (maybe using GPT) before returning it tomGPT as part…


By: The Posts Author | Posted on: 31 Mar 23

@p_millerd Would need considerably more data. DM me if you wanna chat. A fair bit of experience as a deal guy.

@p_millerd Would need considerably more data. DM me if you wanna chat. A fair bit of experience as a deal guy.


By: The Posts Author | Posted on: 31 Mar 23

@IntuitMachine Workman-like not mind blowing. Fine for writing a 10 line function, setting up a framework, etc. world class replacement for Stack Overflow type question.

@IntuitMachine Workman-like not mind blowing. Fine for writing a 10 line function, setting up a framework, etc. world class replacement for Stack Overflow type question.


By: The Posts Author | Posted on: 31 Mar 23

@johnrobb Generative Worlds will massively reduce the cost and scale up VR, too. It’s building VR material that has been the bottleneck. Current tools are too hard for normies. Text->VR stuff may actually ignite VR.

@johnrobb Generative Worlds will massively reduce the cost and scale up VR, too. It’s building VR material that has been the bottleneck. Current tools are too hard for normies. Text->VR stuff may actually ignite VR.


Scroll to Top