正在加载视频...

视频加载失败

damn this is sick google dropped their new Omni multi modal model these outputs look insane live from Google IO

28,064 次观看 • 4 个月前 •via X (Twitter)

14 条评论

skinder 的头像
skinder4 个月前

what?? Is this AI generated? Don’t believe until its released to the public

Ole Lehmann 的头像
Ole Lehmann4 个月前

@skinderdev It’s out now!

Mark Martinkov 的头像
Mark Martinkov4 个月前

Omni looks impressive on stage. But does multi-modal speed actually win, or is it the reliability harness that takes the workflow layer? Curious how Anthropic’s edge on agentic consistency holds up against Google’s scale and integration.

Buddy 的头像
Buddy4 个月前

Honestly, the presentation was impressive My only question is how well it survives real work: messy context, vague goals, weird edge cases, and users who don’t want to manage the agent all day

Szymon Cieślik | Agenci AI 的头像
Szymon Cieślik | Agenci AI4 个月前

Demo outputs from a keynote are the wrong sample. They are cherry-picked and run on unlimited compute. The number I wait for is latency and cost at the rate limit a normal account actually gets. That gap is where most multimodal hype dies.

David Tabachnik 的头像
David Tabachnik4 个月前

Try recreating them and tell us how it goes

Moussa . AI 的头像
Moussa . AI4 个月前

🚀 Gemini Omni is straight-up next level! Those real-time video generations and edits are insane 🔥 Google just raised the bar again. Who’s trying this first? #GeminiOmni #GoogleIO

leon 的头像
leon4 个月前

these videos are too slick not to use in marketing hype videos

Garex 的头像
Garex4 个月前

OMG NO WAY

Dan H 的头像
Dan H4 个月前

Next level

Lost 🎯 的头像
Lost 🎯4 个月前

As bullish as it get's

Misha 的头像
Misha4 个月前

😳😍

Zima 的头像
Zima4 个月前

Counterfeit ideas intensifies. Glad I didn’t have to waste my tokens on slop.

KB UBUNTU 的头像
KB UBUNTU4 个月前

seedance is way better this is embarassing for a company like google

相关视频