Gemini 3, every mention

25 scenes, the whole family · ← back to Gemini 3

mentionsepisodesper episode tap a year for its mentions

Verbatim, from the transcripts: the passages where Gemini 3 comes up

loading…

▶ 14:39 John Coogan Now, what's interesting is that, uh, Gemini III felt like a new base pre-trained, felt like it had some of that big model smell, felt like it was, uh, you know, really delightful to talk to, and I think a lot of people were expecting…

▶ 22:14 John Coogan The company's Gemini three model has put it on top of a heap of performance for AI models, while

▶ 13:22 unnamed speaker You're like, oh, yeah, like, OpenAI codex is this quality, Gemini three is good for this, and

▶ 9:44 unnamed speaker Would you rather have Gemini 2.5 and an app or Gemini three and no app?

▶ 17:24 unnamed speaker November 18th, Google releases Gemini three.
▶ 17:52 unnamed speaker And then December 17th, Google releases Gemini three flash.

▶ 20:59 John Coogan The first Google blow was Gemini III, which scored better than OpenAI's state-of-the-art model on a host of benchmarks, even if actual real-world usage was a bit more uneven. 3 times in the scene

▶ 7:14 unnamed speaker And then they went and did a pre-train and, and Gemini three outperformed, uh, Anthropic also pre-trained. 2 times in the scene

▶ 17:46 unnamed speaker Apparently, I asked GBT, 4.5, and I asked Gemini three pro.

▶ 12:29 unnamed speaker It's state of the art for image generation, editing with more advanced world knowledge, text rendering, precision plus controls, built on Gemini three.

▶ 0:01 John Coogan My take is, do you want iMessage in Gemini three? 4 times in the scene
▶ 5:33 John Coogan Gemini Three Pro is the first LLM to beat professional human players at GeoGuessr. 4 times in the scene
▶ 7:01 John Coogan And then three flash, if that comes out.
▶ 7:04 John Coogan Or three lights.
▶ 7:54 John Coogan Google trained Gemini three pro on Google's own TPUs.

▶ 0:00 unnamed speaker Gemini three pro Google's most intelligent model yet with state of the art reasoning, next level vibe coding and deep multimodal understanding. 3 times in the scene
▶ 3:16 unnamed speaker Do you want to read your, uh, the, the results of the, the standup comedy bit from Gemini three? 3 times in the scene
▶ 6:21 unnamed speaker And you can see that with the Gemini three pro model card, uh, which is doing very well. 3 times in the scene
▶ 7:35 unnamed speaker Uh, Gemini three deep think preview gets a 45% on it at 77 dollars a task.
▶ 13:01 unnamed speaker Gemini III is our latest such model advancement, so we went out to build the next step change of an IDE.
▶ 16:57 unnamed speaker It's Gemini three day. 4 times in the scene
▶ 18:27 unnamed speaker So the other, the other OpenAI news that dropped on, you know, around Gemini three day, Gemini three week, is this profile, um, in the, in Wired of Fiji Simo.
▶ 29:12 unnamed speaker He's like, my idea was Gemini three.

▶ 1:47:38 John Coogan Demoing Gemini three should be a good rebuttal from the Geminis.

▶ 6:30 unnamed speaker Demoing Gemini III should be a good rebuttal from the Geminis.

Turn any episode into a week of clips.