Llama, every mention

18 scenes, the whole family · ← back to Llama

mentionsepisodesper episode tap a year for its mentions

Verbatim, from the transcripts: the passages where Llama comes up

loading…

▶ 3:13:55 Ben Gilbert That was people searching Instagram who accidentally hit a llama model that made some things happen, and they were like, oh, go away.

▶ 5:20:10 David Rosenthal I mean, this is part of what FAIR is, this is what LAMA is, et cetera, et cetera.
▶ 5:22:16 Ben Gilbert It's the llama models underneath it.
▶ 5:26:42 Ben Gilbert It's worth disambiguating Meta AI from Llama.
▶ 5:26:45 Ben Gilbert Llama is the name of their family of open source models.
▶ 5:42:14 Ben Gilbert One is llama, and this is the family of foundational models that Meta has developed, and these are competitive with OpenAI and Anthropix Claude and Google Gemini, etc.
▶ 5:42:33 Ben Gilbert That is the way that you interact with Meta's self-hosted version of Llama, or maybe give it a little bit more credence than that.
▶ 5:42:41 Ben Gilbert It is an application that Meta has that uses Llama in the background, but provides meta-specific AI experiences, some of them bundled into apps, like in WhatsApp chat or in Instagram.
▶ 5:43:00 Ben Gilbert But Llama is the models themselves.
▶ 5:43:05 Ben Gilbert And Llama, they have spent billions and billions of dollars, huge amounts of R&D, huge GPU clusters to train, big data centers.
▶ 5:43:14 Ben Gilbert Interestingly, it is all open source.
▶ 5:43:34 Ben Gilbert Meta is putting huge amounts of CapEx and OpEx, like huge amount of dollars into willing Llama into existence.

▶ 2:18:38 Mark Zuckerberg and, you know, how open these platforms, whether it's things like llama and AI or, um, the glasses or different things should be for developers, like an individual, someone getting started in their dorm room like me to not have to ask for…

▶ 34:49 Mark Zuckerberg I mean, around Lama.
▶ 1:10:51 Mark Zuckerberg and, you know, how open these platforms, whether it's things like llama and AI or, um, the glasses or different things should be for developers, like an individual, someone getting started in their dorm room like me to not have to ask for…

▶ 56:56 David Rosenthal A cloud that allows customers to train and run their own instances of open source models like Llama two and stable diffusion.

▶ 30:37 David Rosenthal GPT-III, GPT-IV, Lama, Bard, these others, they can guess who the criminal is.
▶ 2:35:03 Ben Gilbert And over time, like, you look at, uh, Google's Chinchilla or Llama II, they actually use less parameters than GPT-IV and have equivalent quality, or, you know, many other people can be the judge of that, but we're high-quality models with…

Turn any episode into a week of clips.