花叔专栏 / AI工具实测
这249.99美金/月的Google AI Ultra会员,我先替你们花了
先来看一段关于昨晚Google I/O大会的播客对谈画面,一定要点开,一定要开声音哦🎵
Google在I/O大会上照例又发布了一大箩筐的模型和产品,而且Google还是有些大公司的通病,产品线太多太混乱,我觉得你真正值得关注的是下面三个点:1、Gemini 2.5 Pro新增了Deep Think版本,在各种评测和跑分榜上更近了一步,这个模型在文本对话的能力和代码生成能力在对着OpenAI和Anthropic两头打;2、新推出Gemini 2.5 Flash版本,速度更快,价格更低,但是性能表现还狠客观,在LMarena这个用户做大模型对比评测的平台上也基本处在了前五的位置;3、新推出Veo3视频生成模型,可以生成8秒1080P60帧,带声音和多镜头的视频。A well-lit podcast studio with two tech-savvy hosts casually chatting across a wooden table, each with a microphone and headphones. Behind them is a large screen or LED wall displaying vibrant, animated Google I/O 2025 branding — glowing "Google I/O" logo, abstract shapes, and light blue/purple gradients, matching the official visual identity (VI) of the event.
Host 1 leans in with excitement, gesturing animatedly:
- “Veo 3 can generate video with sound and dialogue — it’s insane!”
Host 2 chuckles, slightly skeptical, raising a coffee mug:
- “Sure, but $249 a month? That’s almost cinematic-level pricing!”
Both laugh. The mood is light, engaging, and natural. Background includes subtle jazz or chill beats, soft ambient mic hiss, and occasional typing sounds.
Camera: Medium two-shot, slight handheld motion for realism. Ends on a quick zoom-in to the background screen showing “Google I/O 2025 – Veo 3 Demo” and a looping video thumbnail.
这篇读完了,下次实践再见。
花叔 Alchain
AI Native Coder · 独立开发者 · 作者
《一本书玩转DeepSeek》作者,小猫补光灯、女娲Skill生态创作者。不会写代码,用AI做产品,分享独立开发、AI工具与内容创作的实践。
本文收录自微信公众号。在原平台阅读 ↗