TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Generating Videos with Scene Dynamics

64 pointsby Ivoahover 8 years ago

5 comments

undershirtover 8 years ago
Once, when waking from a nap, I was able to consciously keep my &quot;dream eyes&quot; open with my real eyes closed. I could vividly see images my dream was still generating, and it looked a lot like this.<p>Makes me wonder if I&#x27;m just influenced by the work happening in ML, or if we are really approaching what the brain is already doing.
gallerdudeover 8 years ago
People may call BS, but I do think we&#x27;ll reach a point where we can generate coherent books or movies. It&#x27;ll take many more neurons, but I think the possibility is out there.<p>The real question is where these &quot;fake&quot; pieces of art will be placed in our society.
评论 #13298704 未加载
goblin89over 8 years ago
Future prediction capability here can greatly improve video monitoring of all kinds. Run a constantly trained system like this in real time over incoming streams and let the agent observe generated predictive videos N units of time ahead.
jedimastertover 8 years ago
&gt;These videos are not real; they are hallucinated by a generative video model.<p>I&#x27;m not sure why, but the fact that they used the term &quot;hallucinated&quot; is a little unsettling.
评论 #13298943 未加载
myowncrapulenceover 8 years ago
Interesting this only used 11tb worth of video.<p>Makes me wonder what this could produce with youtube&#x27;s entire library.