Scale AI podcast mentions

Quote

they need because of two things to work accurately and to work efficiently. When we look at these agents, the first piece is that They are essentially building suggestions and the answers are non deterministic. The second piece is it costs tokens. But when we look at a AI and we look at the ROI of it, especially this is why we want equipped tooling for these agents. AI ROI is actually recognizing. that scale. MIT Sloan actually recently said large scale AI transformation is just at around eight percent. So some of the blockers of Y why agents need some of these toolings, especially thinking about it at scale, is three different walls they hit. One is agents can't maintain around 10,000 repos Secondly, like I mentioned was the accuracy piece. It's non deterministic issues that seem might be a little bit more than a little bit. minor in a demo, but they compound at scale. These hallucinations could be catastrophic when you look at its scale. Then the next thing as we're starting to see with the recent and news, we're starting to see token based billing. Agents are actually burning millions of tokens, rebuilding the context every single time to understand what's happening. happening and then these aren't just edge cases. We're starting to see these as outcomes. And this is why if you're able to provide deterministic tooling for these

Quote

Welcome to the podcast. I'm your host, Jaden Schaefer. Today we have an absolutely packed lineup for the show. The biggest thing that I've been super excited about yesterday, Anthropic Unveiled. Their hosted AI agents platform. It's their cloud platform. I've been playing around with it a ton. There's so much there, and I think this is a really big shift basically up end of time. A lot of what OpenClaw did, but there's some nuances, so we're gonna get into that. In addition, Meta just dropped their very first model that was built with Alexander Wang. Remember, that's formerly the CEO of Scale AI, what they kind of acquired him in. We also have a research team at Tufts that figured out how to cut AI energy consumption by a factor of hundreds. Which is definitely a big deal if you think about how much power these data centers are burning through. Eli Lilly just flipped the switch on what they're calling the most powerful AI human supercomputer in Far. day work week, which is probably the most open AI thing I've read in a while. And Google released Gemma 4, which is their latest open source model that's getting a lot of attention for what it can do relative to it. size. I mean basically this is an edge model that you can put on devices. So a lot to cover in the show today. Before we get into that, I want to mention AI box, which is a tool I use every single day. day at this point…

Quote

Welcome to the podcast. I'm your host, Jaden Schaefer. Today we have an absolutely packed lineup for the show. The biggest thing that I've been super excited about yesterday, Anthropic Unveiled. Their hosted AI agents platform. It's their cloud platform. I've been playing around with it a ton. There's so much there, and I think this is a really big shift basically up end of time. A lot of what OpenClaw did, but there's some nuances, so we're gonna get into that. In addition, Meta just dropped their very first model that was built with Alexander Wang. Remember, that's formerly the CEO of Scale AI, what they kind of acquired him in. We also have a research team at Tufts that figured out how to cut AI energy consumption by a factor of hundreds. Which is definitely a big deal if you think about how much power these data centers are burning through. Eli Lilly just flipped the switch on what they're calling the most powerful AI human supercomputer in Far. day work week, which is probably the most open AI thing I've read in a while. And Google released Gemma 4, which is their latest open source model that's getting a lot of attention for what it can do relative to it. size. I mean basically this is an edge model that you can put on devices. So a lot to cover in the show today. Before we get into that, I want to mention AI box, which is a tool I use every single day. day at this point…

Quote

Welcome to the podcast. I'm your host, Jaden Schaefer. Today we have an absolutely packed lineup for the show. The biggest thing that I've been super excited about yesterday, Anthropic Unveiled there. Their hosted AI agents platform. It's their cloud platform. I've been playing around with it a ton. There's so much there and I think this is a really big shift basically up end. a lot of what OpenClaw did, but there's some nuances, so we're gonna get into that. In addition, Meta just dropped their very first model that was built with Alexander Wang. Remember that's formerly the CEO of Scale AI, what they kind of acquired him in. We also have a research team at Tufts that figured out how to cut AI energy consumption by a factor of hundreds. Which is definitely a big deal if you think about how much power these data centers are burning through. Over a thousand Blackwell GPUs, which are aimed at cutting drug development timelines in half, OpenAI published a set of policy proposals that include robot taxes and affordable. And Google released Gemma 4, which is their latest open source model that's getting a lot of attention for what it can do relative to its size. I mean basically this is an edge model that you can put on devices. So a lot to cover in the show today. Before we get into that, I want to mention AI Box, which is a tool I use every single day. day at this point. If you h…

Quote

Welcome to the podcast. I'm your host, Jaden Schaefer. Today we have an absolutely packed lineup for the show. The biggest thing that I've been super excited about yesterday, Anthropic Unveiled there. Their hosted AI agents platform. It's their cloud platform. I've been playing around with it a ton. There's so much there and I think this is a really big shift basically up end. a lot of what OpenClaw did, but there's some nuances, so we're gonna get into that. In addition, Meta just dropped their very first model that was built with Alexander Wang. Remember that's formerly the CEO of Scale AI, what they kind of acquired him in. We also have a research team at Tufts that figured out how to cut AI energy consumption by a factor of hundreds. Which is definitely a big deal if you think about how much power these data centers are burning through. Over a thousand Blackwell GPUs, which are aimed at cutting drug development timelines in half, OpenAI published a set of policy proposals that include robot taxes and affordable. And Google released Gemma 4, which is their latest open source model that's getting a lot of attention for what it can do relative to its size. I mean basically this is an edge model that you can put on devices. So a lot to cover in the show today. Before we get into that, I want to mention AI Box, which is a tool I use every single day. day at this point. If you h…

Quote

here to unpack, but why don't you give us kind of the founding story of where you came up with the idea for kernel? Yeah, sure. So I've kind of grown up uh different phases. of AI innovations. And so out of college, I founded a company. That was my C back and what I would describe as the last AI wave free transformer. So my company at that time was trying to use supervised machine learning to automate book keeping for small businesses. And scale AI was in in my batch. And so I got it kind of like CEO. over over the course of the ten years how a company at a technical logical shift starts in their earliest days and then how they grow and mature. scale as the technology continues to improve. So that's kind of like my very, very kind of background aside from kernel. A couple of years ago, I was working at Cash App, leading a team of deployed engineers and this was in kind of 2024. So Traty BC was becoming more widespread and her friends were starting to adopt tools like Glean. And when Cloud Computer Use came out in beta in twenty twenty. three, there was a sister team that started to try to use it, their QA engineering team. And what they were trying to use it for was not QA cash app and blocks. own website services, but QA in their partner websites. And so

Quote

Centralization of intelligence. Because the industry is hitting a structural wall. And to solve it, they're aggressively pushing intelligence to the edge. Right, meaning into your And the domino that started this whole panic, the catalyst, was the realization that we have to do. that massive centralized GPU clusters are essentially eating themselves. Yeah, I saw the quote in the sources from Greg Steinbritcher. He's OpenAI's workload lead. And he calls conventional large scale AI training a failure amplifier. Which is such a descriptive term. It sounds catastrophic, honestly. Like what does failure amplifier act? actually mean in the context of building a model. Well think about it like this. When you link tens of thousands of processors together to train a massive model, they have to communicate constantly. Okay. But GPUs process math faster than current network switches can actually route the data. So if one tiny thing breaks in that massive desert data set, center like a single switch fails, or a cable misfires. Or just a momentary network blip. Exactly. The ripple effect forces the entire cluster to grind to a halt. The GPUs just sit there idle waiting for the network to catch up. Wow. And when you are running a hundred megawatt facility, I mean idle time means you're burning millions of dollars. Which perfectly explains the software band-aid OpenAI just rolled out. They u…

Quote

and navigating change in the era of generative AI. I'm excited to welcome Angeli Kakad, a technology leader and what many would call a gen AI navigator. Anjali brings a grounded practical perspective on AI, moving beyond perfectionism, embracing iteration, and helping organizations translate AI ambition into real business impact. In this conversation, we'll explore What it truly means to navigate Gen AI, why progress beats perfection in today's AI-driven world, and how leaders can scale AI to the AI. responsibly while keeping humans at the center. Let's dive in and extract the raw AI. All right, uh welcome back to our uh extra AI podcast series. And as I've been uh During the last few sessions, the idea is that this is now uh we're doing a thought uh thought leadership series, the topic of um the leadership series. series and this is where I'm talking with uh different leaders uh in the context of AI. So welcome on board Anjali

Quote

The human supervisor has to be able to understand what the other thing is. Why it prioritized stability over a feature in that moment. Predicting such widespread adoption of AI ops by twenty twenty six. Optimization. It's a total shift to predictive autonomy and AI oversight. Advanced ML platforms and agentic AI tools at an ultra scale AI integrated level. So for the decision makers, for the C-suite listeners, this isn't just some technical curiosity. This whole evolution, it translates directly into tangible financial and operational improvement. Improvements. This is the business case. Absolutely. The primary measurable impact is on the MTTR meantime to resolution. We're talking about a paradigm shift. In the sysadmin days, resolution. With AI ops and autonomous agents, that time drops. What? Single digit minutes. And reducing downtime. We know from industry estimates. that large enterprises can lose upward of fifty six hundred dollars per minute. Minimizing that exposure is enormous. But it's not just about avoiding losses. It's about creating a competitive

Quote

And we see like You can still and importantly we still have you know the crowdsourcing and data labeling that's Right. So we actually have both like near Crowd which is kinda original project we had for crowdsourcing, like crowdsourcing on the like scale AI. style as well we have a public AI which is another like data labeled crowdsourcing. low at the hardware level for the confidential computing but how much is there in terms of just like the software setup, right? I mean it's not super easy to set up I mean you have a marketplace, but you know, even going out and hiring a bunch of people to create data Like that's not super easy even in in sort of base case. Interested in how much Also at the sort of model like who makes these decisions?

Quote

Hey everyone. Adam, what's on the agenda today? Today, we're diving into the world of AI with the first time. Some groundbreaking updates. First up, the Time AI agent is here to revolutionize how we consume news. Imagine your morning news not just summarizing the headlines but speaking your language, literally and figuratively. Wow, that's like having a personal news concierge. How does it work? Great question Sky. Developed with Scale AI, the Time AI agent offers features like summarization, translation, and even audio generation. It's all about enhancing reader engagement while maintaining the integrity of Times journalism. Sounds like a game changer. But what about the big bucks? I heard Meta's making a huge investment. Absolutely. Meta is investing a whopping $600 billion in U.S. data centers to boost. Saw with innovation on one side and sustainability on the other. That's a lot of zeros. And what about the startups? I heard Savorite and Claude. Indeed. Savorite scalable intelligence has secured over 100 million dollars in pre-orders, and core weave is reporting.