Llama, every mention
50 scenes, the whole family · ← back to Llama
tap a year for its mentions
every year anyone Anjney Midha 9Jeff Schmidt 7Marc Andreessen 6Ben Horowitz 6Andrew Bosworth 5Guido Appenzeller 4Matt Bornstein 3Chris Dixon 3Wei-Lin Chiang 2Ryan Chi 2
Verbatim, from the transcripts: the passages where Llama comes up
Inside the Race to Measure Frontier Intelligence
- ▶ 0:05 Ryan Chi When Meta released Llama four on our held out private benchmarks, the model was actually underperforming, but on all of the major public benchmarks, it was showing incredible capabilities.
- ▶ 2:38 Ryan Chi Yeah, I mean, internally, they, they do build a lot of great benchmarks, and that's what drives model progress, but I think there's an issue when we, we speak about model capabilities in a way that's self-reported, and so one of the early…
How Cursor Built One of AI’s Fastest-Growing Companies
- ▶ 1:25 Matt Bornstein The leading models at the time were GPT-Foro, Claude-III, and Llama-III. 2 times in the scene
How Open Source Became AI's Backbone | Inferact with a16z
- ▶ 20:29 Simon Moe So we have been starting to see, uh, terms even to the Lama days for, for when Meta was releasing Lama, they do have a term of if your daily active user or like annual recurring revenue exceeds some threshold, please enter into a…
- ▶ 25:45 Matt Bornstein One of my favorite artifacts, maybe we could even track down the link is one of the early llama models.
The $700 Billion AI Productivity Problem No One's Talking About
- ▶ 39:12 Russ Fraden And so the other thing we did is we, we basically trained our own kind of customized llama model to block people from asking questions that are illegal or the company doesn't want you to.
Dylan Patel on the AI Chip Race - NVIDIA, Intel & the US Government vs. China
- ▶ 1:32:24 Dylan Patel I'll use, I'll use Llama, seventy-b, because it's simple to do math on, like, seventy billion parameters, uh, that's, that's a 140 gigaflops per token, 70 times, uh, 64,000, that's, that's,
Chris Dixon on How to Build Networks, Movements, and AI-Native Products
- ▶ 38:59 Chris Dixon But then you have, you know, Facebook is doing it with Lama. 2 times in the scene
The Future of AI Video: How Fal.ai is Making AI Video Faster & Easier
- ▶ 10:20 Burkay Gur The point when it kind of clicked for us was actually, I think LAMA-II was a pretty, pretty big turning point. 2 times in the scene
Beyond Leaderboards: LMArena’s Mission to Make AI Reliable
- ▶ 30:20 Wei-Lin Chiang And then at that time, Llama one was just released, which is a base model. 2 times in the scene
- ▶ 1:37:06 Anastasios Angelopoulos That's what we did with Llama, right? 2 times in the scene
Sovereign AI: Why Nations Are Building Their Own Models
- ▶ 11:37 Anish Acharya So what do we want our allies on, DeepSeek or Lama?
What Comes After Mobile? Meta’s Andrew Bosworth on AI and Consumer Tech
- ▶ 8:52 Andrew Bosworth And in fact, they were entirely built, and we were six months away from production when Lama three hit.
- ▶ 24:46 Andrew Bosworth The first one is, you know, LLAMA came out of FAIR, um, our fundamental AI research group, and that's been an open source research group since the beginning. 2 times in the scene
- ▶ 25:49 Andrew Bosworth Um, LAMA II was probably the big decision point for us. 2 times in the scene
Can the Military Move at Startup Speed? How the Army and Navy Are Rebuilding
- ▶ 29:21 Alex Miller Um, I had a chance last week to sit with all of the big LLM providers, and I, I loved the conversation because they were ruthlessly honest with, with me and with each other in that, hey, all of these things are available, whether it's…
The Future of Digital Workers
- ▶ 7:09 Prabhav Jain Then I'm going to take that text and, like, throw it into, like, GPT-FORO or, like, a Lama model or a cloud model, right?
Jensen Huang & Arthur Mensch: Why Every Nation Needs Its Own AI Strategy
- ▶ 33:04 unnamed speaker Your co-founder Guillaume created Llama.
Building the Next Generation of Conversational AI
- ▶ 18:32 Anjney Midha But I mean, basically it was just Lama around the time Sesame was started.
DeepSeek, Reasoning Models, and the Future of LLMs
- ▶ 18:31 Guido Appenzeller We ran the cost for some Lama models last year, and I think we ended up with, you know, a little over three million dollars, right?
- ▶ 25:06 Marco Mascorro Another, another interesting thing that, that came out of the paper from R-one was for these model models, the LAMAs, the seven Bs, the seven Bs, the 13 Bs. 2 times in the scene
How to use DeepSeek safely
- ▶ 4:15 Joel de la Garza I'm curious your estimation of sort of the, the, the, the maturity and complexity of the deep seek preventions versus like what you'd see in something like llama or some other model.
Who Will Own the Internet? a16z’s Chris Dixon on AI and Crypto
- ▶ 2:30 Chris Dixon I don't believe the safety thing, but, you know, thankfully there's these ones like Llama and Flux and Mistral and things who are open source.
What DeepSeek Means For The Future Of AI | Tech Veterans Weigh In
- ▶ 24:09 Steven Sinofsky I mean, I, you know, run llama on my Mac and it's, it's just the first time you do it, it's your, your mind is blown.
AI Is Becoming a Regional Race
- ▶ 17:04 Anjney Midha You know, he worked at DeepMind, or Guillaume Lomp, who led the initial Lama family at Meta, who are deeply mission-led and, and believe that they can help solve a bunch of these infrastructure problems for the world's largest governments.
Truth Terminal - The AI Bot That Became a Crypto Millionaire
- ▶ 4:53 Marc Andreessen ChatGPT and Claude and, and, um, you know, uh, Grok from, from, from, from, uh, from Elon and Lama from Meta and so forth.
- ▶ 10:03 Marc Andreessen Um, and so anyway, so this, this guy, basically, Andy, Andy, Andy, you know, he started training, I think it was a customized version of, of Lama, so the seventy billion parameter of Lama, and, and so that's the sort of mid-sized model… 4 times in the scene
The Quest for Community-Trained Open Source AI Models
- ▶ 12:27 Jeff Schmidt It is now used by every, every model you use nowadays, everything, everything Chachipiti, Lama, DeepSeq, all of them, they all use the, the methods that we pioneered in, in the YARN paper. 2 times in the scene
- ▶ 19:48 Jeff Schmidt Yeah, but the, the start of it really was this, this idea of what if we don't get Llama IV? 5 times in the scene
- ▶ 24:43 Anjney Midha So, so just to give people a sense of how dramatic of a change that is, could you, how, how many folks, how many organizations do you think in the world do they have the capability to train a co-located model, right, of the scale of a…
- ▶ 53:41 Anjney Midha The, the grand challenge for distro was inspired by the question of, hey, what happens if we never get an open source Lama four, right? 2 times in the scene
Why Human Data is Key to AI: Alexandr Wang from Scale AI
- ▶ 13:22 Alexandr Wang And the models are making their way, like, you know, uh, Llama 3.1 is open source.
Mark Casey (Capital Group): Public Market and Macro Perspectives
- ▶ 20:55 Mark Casey You know, I think about those Ray-Ban glasses that Meta's making that has their Lama, ah, models in it and future AI models.
The Top 100 GenAI Consumer Apps
- ▶ 6:09 Steph Smith These other ones where you're talking about this, you know, lack of construction, are they using Llama, or how are they getting to...
Build Your Startup With AI
- ▶ 6:24 Ben Horowitz I think if you look at the very top models, um, you know, Claude and, uh, OpenAI and Mistral and Lama, um,
- ▶ 16:36 Ben Horowitz When they proved it in the new Lama release, right?
- ▶ 29:01 Ben Horowitz Um, you know, Lama, Mistral, like these are great models.
Startup Building: Challenges & Opportunities
- ▶ 47:45 Marc Andreessen Um, that, that was a key part of the Lama, the Lama news, you know, this week.
Safety in Numbers: Keeping AI Open
- ▶ 0:31 unnamed speaker Now, earlier this year, Arthur banded together with two other researchers, Guillaume Lample and Timothée Lacroix, two researchers at Meta who worked on Llama.
- ▶ 6:38 Anjney Midha And if I recall correctly, right before they left, Tim and Guillaume had started to work on Lama, right, over at Meta. 4 times in the scene
Big Ideas 2024: The Consumer AI Battleground Moves from Model to UX with Alex Immerman
- ▶ 13:49 Alex Immerman Several of the leading open source models like Lama and Mistral are already competitive with the closed source equivalents.
Universally Accessible Intelligence with Character.ai's Noam Shazeer
- ▶ 11:35 Sarah Wang Advancements on the open source model side, and, um, you know, maybe folks building a product on top of a fine-tuned Llama II, you know, fine-tuned for chat.
AI Copilots and the Future of Knowledge Work with Microsoft's Kevin Scott
- ▶ 2:15 Bob Swan You've partnered with leaders in the industry, whether it's GitHub or OpenAI, as you mentioned, and most recently, the Meta Llama II announcement. 2 times in the scene
- ▶ 5:28 Kevin Scott You mentioned Meta and Lama too, and you know, I think Lama too is just like Lama going to be an important building block that people are going to want to use to build AI software.
AI Food Fights in the Enterprise with Databricks' Ali Ghodsi
- ▶ 14:09 Ben Horowitz Um, and then, you know, coming from Databricks perspective, how are you, how are you all thinking about open source, both with respect to Mosaic and then with the other, you know, things like Llama? 3 times in the scene
The True Cost of Compute
- ▶ 14:44 unnamed speaker A quick note to close out that many models today are even bigger, with Llama II, for example, being trained on two trillion tokens, or about eight trillion characters.
Chasing Silicon: The Race for GPUs
- ▶ 13:42 unnamed speaker As one of many emerging examples, Vicuña was created by fine-tuning Meta's LlamaOne model for chat.
- ▶ 16:11 Guido Appenzeller And so, so, uh, you know, Lama, for example, that was a Facebook model where they published the weights for, for researchers. 3 times in the scene
- ▶ 16:49 unnamed speaker It's worth noting that since this recording, several more open source models have been released, including Llama-II with seventy billion parameters and an open license. 3 times in the scene