{"id":470,"date":"2024-02-15T16:34:45","date_gmt":"2024-02-15T21:34:45","guid":{"rendered":"https:\/\/ozer.gt\/log\/?p=470"},"modified":"2024-02-25T13:19:07","modified_gmt":"2024-02-25T18:19:07","slug":"prompt-to-video-but-not-cause-to-effect","status":"publish","type":"post","link":"https:\/\/ozer.gt\/log\/2024\/02\/15\/prompt-to-video-but-not-cause-to-effect\/","title":{"rendered":"Prompt to video, but not cause to effect"},"content":{"rendered":"

The output of Sora, OpenAI’s latest tool, looks really impressive for an off-the-shelf tool. What I found even more interesting is that OpenAI explicitly defines the weakness of the model as not understanding “cause and effect.”<\/p>\n

Their example is a person biting into a cookie in a video, but potentially not leaving a bite mark on the cookie. There is also a reverse treadmill scene.<\/p>\n

Yet OpenAI downplays the absolute lack of cause-and-effect reasoning:
\n\ud835\ude10\ud835\ude35 \ud835\ude62\ud835\ude56\ud835\ude6e \ud835\ude68\ud835\ude69\ud835\ude67\ud835\ude6a\ud835\ude5c\ud835\ude5c\ud835\ude61\ud835\ude5a \ud835\ude38\ud835\ude2a\ud835\ude35\ud835\ude29 \ud835\ude22\ud835\ude24\ud835\ude24\ud835\ude36\ud835\ude33\ud835\ude22\ud835\ude35\ud835\ude26\ud835\ude2d\ud835\ude3a \ud835\ude34\ud835\ude2a\ud835\ude2e\ud835\ude36\ud835\ude2d\ud835\ude22\ud835\ude35\ud835\ude2a\ud835\ude2f\ud835\ude28 \ud835\ude35\ud835\ude29\ud835\ude26 \ud835\ude31\ud835\ude29\ud835\ude3a\ud835\ude34\ud835\ude2a\ud835\ude24\ud835\ude34 \ud835\ude30\ud835\ude27 \ud835\ude22 \ud835\ude24\ud835\ude30\ud835\ude2e\ud835\ude31\ud835\ude2d\ud835\ude26\ud835\ude39 \ud835\ude34\ud835\ude24\ud835\ude26\ud835\ude2f\ud835\ude26, \ud835\ude22\ud835\ude2f\ud835\ude25 \ud835\ude62\ud835\ude56\ud835\ude6e \ud835\ude63\ud835\ude64\ud835\ude69 \ud835\ude6a\ud835\ude63\ud835\ude59\ud835\ude5a\ud835\ude67\ud835\ude68\ud835\ude69\ud835\ude56\ud835\ude63\ud835\ude59 \ud835\ude34\ud835\ude31\ud835\ude26\ud835\ude24\ud835\ude2a\ud835\ude27\ud835\ude2a\ud835\ude24 \ud835\ude2a\ud835\ude2f\ud835\ude34\ud835\ude35\ud835\ude22\ud835\ude2f\ud835\ude24\ud835\ude26\ud835\ude34 \ud835\ude30\ud835\ude27 \ud835\ude24\ud835\ude22\ud835\ude36\ud835\ude34\ud835\ude26 \ud835\ude22\ud835\ude2f\ud835\ude25 \ud835\ude26\ud835\ude27\ud835\ude27\ud835\ude26\ud835\ude24\ud835\ude35.<\/p>\n

while doubling down on its promise of AGI:
\n\ud835\ude1a\ud835\ude30\ud835\ude33\ud835\ude22 \ud835\ude34\ud835\ude26\ud835\ude33\ud835\ude37\ud835\ude26\ud835\ude34 \ud835\ude22\ud835\ude34 \ud835\ude22 \ud835\ude27\ud835\ude30\ud835\ude36\ud835\ude2f\ud835\ude25\ud835\ude22\ud835\ude35\ud835\ude2a\ud835\ude30\ud835\ude2f \ud835\ude27\ud835\ude30\ud835\ude33 \ud835\ude2e\ud835\ude30\ud835\ude25\ud835\ude26\ud835\ude2d\ud835\ude34 \ud835\ude35\ud835\ude29\ud835\ude22\ud835\ude35 \ud835\ude24\ud835\ude22\ud835\ude2f \ud835\ude36\ud835\ude2f\ud835\ude25\ud835\ude26\ud835\ude33\ud835\ude34\ud835\ude35\ud835\ude22\ud835\ude2f\ud835\ude25 \ud835\ude22\ud835\ude2f\ud835\ude25 \ud835\ude34\ud835\ude2a\ud835\ude2e\ud835\ude36\ud835\ude2d\ud835\ude22\ud835\ude35\ud835\ude26 \ud835\ude35\ud835\ude29\ud835\ude26 \ud835\ude33\ud835\ude26\ud835\ude22\ud835\ude2d \ud835\ude38\ud835\ude30\ud835\ude33\ud835\ude2d\ud835\ude25, \ud835\ude22 \ud835\ude24\ud835\ude22\ud835\ude31\ud835\ude22\ud835\ude23\ud835\ude2a\ud835\ude2d\ud835\ude2a\ud835\ude35\ud835\ude3a \ud835\ude38\ud835\ude26 \ud835\ude23\ud835\ude26\ud835\ude2d\ud835\ude2a\ud835\ude26\ud835\ude37\ud835\ude26 \ud835\ude38\ud835\ude2a\ud835\ude2d\ud835\ude2d \ud835\ude23\ud835\ude26 \ud835\ude56\ud835\ude63 \ud835\ude5e\ud835\ude62\ud835\ude65\ud835\ude64\ud835\ude67\ud835\ude69\ud835\ude56\ud835\ude63\ud835\ude69 \ud835\ude62\ud835\ude5e\ud835\ude61\ud835\ude5a\ud835\ude68\ud835\ude69\ud835\ude64\ud835\ude63\ud835\ude5a \ud835\ude5b\ud835\ude64\ud835\ude67 \ud835\ude56\ud835\ude58\ud835\ude5d\ud835\ude5e\ud835\ude5a\ud835\ude6b\ud835\ude5e\ud835\ude63\ud835\ude5c \ud835\ude3c\ud835\ude42\ud835\ude44.<\/p>\n

Still, the model is clearly useful for a number of business applications, most obviously marketing and promotional videos. It could also be a potential game changer for the creative industries when the 60-second limit is lifted, such as museums, performing and visual arts, galleries, and fashion design.<\/p>\n

Source<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"

The output of Sora, OpenAI’s latest tool, looks really impressive for an off-the-shelf tool. What I found even more interesting is that OpenAI explicitly defines the weakness of the model as not understanding “cause and effect.” Their example is a person biting into a cookie in a video, but potentially not leaving a bite mark […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"cybocfi_hide_featured_image":"","footnotes":""},"categories":[1],"tags":[],"class_list":["post-470","post","type-post","status-publish","format-standard","hentry","category-uncategorized"],"_links":{"self":[{"href":"https:\/\/ozer.gt\/log\/wp-json\/wp\/v2\/posts\/470","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/ozer.gt\/log\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/ozer.gt\/log\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/ozer.gt\/log\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/ozer.gt\/log\/wp-json\/wp\/v2\/comments?post=470"}],"version-history":[{"count":1,"href":"https:\/\/ozer.gt\/log\/wp-json\/wp\/v2\/posts\/470\/revisions"}],"predecessor-version":[{"id":471,"href":"https:\/\/ozer.gt\/log\/wp-json\/wp\/v2\/posts\/470\/revisions\/471"}],"wp:attachment":[{"href":"https:\/\/ozer.gt\/log\/wp-json\/wp\/v2\/media?parent=470"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/ozer.gt\/log\/wp-json\/wp\/v2\/categories?post=470"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/ozer.gt\/log\/wp-json\/wp\/v2\/tags?post=470"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}