Skip to playerSkip to main content
  • 7 hours ago

Category

🤖
Tech
Transcript
00:00Google launches Gemini 3 with major improvements.
00:04Google rolled Gemini 3 into its app and AI search, its most powerful, hope this goes
00:09well foundation model yet.
00:11If this feels fast, that's because it is.
00:14Gemini 2.5 debuted 7 months ago, roughly a decade and a half in AI years.
00:21The launch also lands amid an industry-wide AI sprint that's starting to look like a
00:25nerdy arms race.
00:27OpenAI launched GPT 5.1, Anthropics shipped Sonnet 4.5 earlier, and Google is flashing
00:34benchmark charts like a proud parent.
00:37A turbocharged, research-focused edition Gemini 3 DeepThink is waiting in the wings for Google
00:42AI Ultra subscribers.
00:45Google says the model needs safety testing before release, ensuring it doesn't invent
00:49its own meaning of Ultra.
00:51According to the company, the model's biggest trick is a serious leap in reasoning.
00:56It's responding with a level of depth and nuance that we haven't seen before, it said.
01:01Benchmarks back it up.
01:02Gemini 3 just posted a record 37.4 on humanity's last exam, a test meant to evaluate general reasoning.
01:10For context, the previous record holder, GPT 5 Pro, scored 31.64.
01:16And Google also claims it now tops LM Arena, a human-driven ranking of models.
01:21Google are definitely showing up for it.
01:24The Gemini app has more than 650 million monthly users, and 13 million developers use it, so
01:29a not-insignificant portion of the world's code is now being ghostwritten by an algorithm.
01:35Google launched Antigravity, a Gemini-powered coding tool blending ChatGPT, command line, and
01:40browser, revealing exactly how your app breaks now.
01:44DeepMind CTO, Kore Kavukchoglu says the agent fluidly shifts between editor, terminal, and browser
01:51to help developers build applications optimally.
01:53you
Be the first to comment
Add your comment

Recommended