Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

"Civilization is the only example of super intelligence." Our CEO @davidpstout on Startup Project explaining why we're building millions of specialized models across devices — not one giant model in the cloud. A few highlights: – Why prompting loses money – 8B models matching GPT-3.5 – How we beat... show more

webAI

5,195 subscribers

325,138 views • 9 months ago •via X (Twitter)

News & Politics Education Science & Technology

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

Last week we released OpenAI Developers gpt-realtime our latest speech-to-speech model. Prompting S2S models can be extremely powerful. To get the most out of the model we compiled a series of prompting tips of what we saw work with early customers. To present them here's Cedar our new voice in gpt-realtime walk you through it. If you want to read it yourself check the links below 👇

Last week we released OpenAI Developers gpt-realtime our latest speech-to-speech model. Prompting S2S models can be extremely powerful. To get the most out of the model we compiled a series of prompting tips of what we saw work with early customers. To present them here's Cedar our new voice in gpt-realtime walk you through it. If you want to read it yourself check the links below 👇

dominik kundel

66,904 views • 11 months ago

Revolutionizing Move Programming with OpenLedger In this demo, we showcase how Move datasets contributed by data providers to OpenLedger’s datanets are used to fine-tune specialized models with LoRA fine-tuning. As seen in the video, we showcase an example on how builders can deploy a Move-specialized model that powers Co-pilot agents using our no-code model fine-tuning platform. This is the future of AI and Web3 innovation. Watch this space to see more specialised models and data feeds being built for next generation agents on top of OpenLedger #Move

Revolutionizing Move Programming with OpenLedger In this demo, we showcase how Move datasets contributed by data providers to OpenLedger’s datanets are used to fine-tune specialized models with LoRA fine-tuning. As seen in the video, we showcase an example on how builders can deploy a Move-specialized model that powers Co-pilot agents using our no-code model fine-tuning platform. This is the future of AI and Web3 innovation. Watch this space to see more specialised models and data feeds being built for next generation agents on top of OpenLedger #Move

OpenLedger

61,662 views • 1 year ago

“Our compute nodes don’t run a full copy of the models we train on IOTA. They actually run a small sliver of the model. This means you can train really large frontier sized models using very small building blocks”. CTO crux unpacks how IOTA ・ SN9’s model parallel architecture allows us to train at scale by splitting our models across multiple machines, and then sewing them together. This is core to SN9’s architecture. See the full Eye On A.I. podcast below.

“Our compute nodes don’t run a full copy of the models we train on IOTA. They actually run a small sliver of the model. This means you can train really large frontier sized models using very small building blocks”. CTO crux unpacks how IOTA ・ SN9’s model parallel architecture allows us to train at scale by splitting our models across multiple machines, and then sewing them together. This is core to SN9’s architecture. See the full Eye On A.I. podcast below.

IOTA ・ SN9

15,465 views • 2 months ago

The founder of a $4B inference company says that if you're building agents, foundational models could become your IP. According to Lin Qiao, 90% of the world's data is still private and locked inside applications and enterprises. Foundation models are trained on public internet + labeling company data, which is barely 10% of all the data. And this is why your application and foundation models are misaligned by definition. Companies building agents don't treat models as APIs. They opt for a product-model co-design. - Models continuously learn from your private data - Pick up domain-specific intelligence - Run faster - Cost less - Scale to millions of users

The founder of a $4B inference company says that if you're building agents, foundational models could become your IP. According to Lin Qiao, 90% of the world's data is still private and locked inside applications and enterprises. Foundation models are trained on public internet + labeling company data, which is barely 10% of all the data. And this is why your application and foundation models are misaligned by definition. Companies building agents don't treat models as APIs. They opt for a product-model co-design. - Models continuously learn from your private data - Pick up domain-specific intelligence - Run faster - Cost less - Scale to millions of users

Ivan Burazin

54,964 views • 3 months ago

Physical Intelligence (Physical Intelligence) is building a foundation model that can control any robot to do any task — what the team describes as the GPT moment for robotics. The company's cross-embodiment approach trains across many different robot platforms, and recent results show tasks being performed zero-shot that last year required hundreds of hours of data collection. In this episode of the Lightcone Podcast , co-founder Quan Vuong (Quan Vuong) sat down with Garry Tan, Jared Friedman, Diana, and Harj Taggar to talk about why robotics is finally ready for its scaling moment, how PI runs its models in the cloud rather than on-device, and the playbook for what Quan sees as a Cambrian explosion of vertical robotics companies. 00:00 — Robotics just got cheaper 00:41 — The GPT moment for robotics 02:24 — Why robots didn’t work before 05:30 — The breakthrough that changed everything 09:12 — The data problem 13:33 — Robots learning without data 15:05 — Robots folding laundry (for real) 22:18 — From engineering problem → ops problem 29:12 — The startup playbook 38:46 — Thousands of robotics startups are coming

Physical Intelligence (Physical Intelligence) is building a foundation model that can control any robot to do any task — what the team describes as the GPT moment for robotics. The company's cross-embodiment approach trains across many different robot platforms, and recent results show tasks being performed zero-shot that last year required hundreds of hours of data collection. In this episode of the Lightcone Podcast , co-founder Quan Vuong (Quan Vuong) sat down with Garry Tan, Jared Friedman, Diana, and Harj Taggar to talk about why robotics is finally ready for its scaling moment, how PI runs its models in the cloud rather than on-device, and the playbook for what Quan sees as a Cambrian explosion of vertical robotics companies. 00:00 — Robotics just got cheaper 00:41 — The GPT moment for robotics 02:24 — Why robots didn’t work before 05:30 — The breakthrough that changed everything 09:12 — The data problem 13:33 — Robots learning without data 15:05 — Robots folding laundry (for real) 22:18 — From engineering problem → ops problem 29:12 — The startup playbook 38:46 — Thousands of robotics startups are coming

Y Combinator

115,590 views • 3 months ago

Dylan Patel says GPT-5 is not a bigger model and did not gain intelligence through scale but through optimization. He highlights how OpenAI improved efficiency by cutting down wasted tokens compared to earlier o-series models. He adds that GPT-5’s new router system decides whether to use the base model, mini, or thinking, leading to faster and more efficient results. He points out that o3 could think for around 30 seconds, while GPT-5 only thinks for 5–10 seconds, meaning less compute is used per query.

Dylan Patel says GPT-5 is not a bigger model and did not gain intelligence through scale but through optimization. He highlights how OpenAI improved efficiency by cutting down wasted tokens compared to earlier o-series models. He adds that GPT-5’s new router system decides whether to use the base model, mini, or thinking, leading to faster and more efficient results. He points out that o3 could think for around 30 seconds, while GPT-5 only thinks for 5–10 seconds, meaning less compute is used per query.

Wes Roth

87,396 views • 11 months ago

Introducing autoresearch with GPT 5.6 We had GPT 5.6 Sol reproduce the key findings from “Towards Mechanistically Understanding Why Memorized Knowledge Fails to Generalize in LLM Finetuning” Compared to GPT-5.5 and even Fable 5, GPT-5.6 stayed more focused on a few, critical experiments and spent less time on peripheral details. It also asked fewer “clarification” questions and independently resolved ambiguities instead of pushing them back to us OpenAI pushing the boundaries of the automated research loop with models that don’t have handcuffs

Introducing autoresearch with GPT 5.6 We had GPT 5.6 Sol reproduce the key findings from “Towards Mechanistically Understanding Why Memorized Knowledge Fails to Generalize in LLM Finetuning” Compared to GPT-5.5 and even Fable 5, GPT-5.6 stayed more focused on a few, critical experiments and spent less time on peripheral details. It also asked fewer “clarification” questions and independently resolved ambiguities instead of pushing them back to us OpenAI pushing the boundaries of the automated research loop with models that don’t have handcuffs

alphaXiv

85,137 views • 18 days ago

✈️ Starting the day in Istanbul. Let's talk AI. The future of AI won't be shaped by size, but by precision. Lately I've been diving into what OpenLedger has been building, and I think we're witnessing one of the most important hard forks in AI: 👉 From giant generalist models 👉 To focused, hightrust AI agents Here's why that shift matters , and why OpenLedger's vision makes perfect sense: ✅ Specialized > Generalized General AI can do many things. But specialized AI? It does one thing extraordinarily well. Tailored models don't waste compute on irrelevant context, every parameter is purposedriven. ✅ Explainability isn't optional anymore In highstakes sectors like finance or healthcare, because the model said so won't cut it. We need transparent reasoning paths. Models must show how they reached conclusions , not just what they concluded. ✅ Trust comes from traceability With OpenLedger's Proof of Attribution, each AI decision is traceable, verifiable, and tamperproof. We're talking onchain records of who contributed what , accountability by design. ✅ Less hallucination, more signal Smaller, specialized models trained on clean, domainspecific data are far less prone to hallucinations. Clear boundaries = higher reliability. ✅ Efficiency is the real scalability Deploying one massive model for everything? Expensive, slow, and unsustainable. Specialized AI is leaner, faster, and far more costeffective. In short, OpenLedger isn't just following a trend. They're laying the rails for the infrastructure layer of verifiable, domainaware AI. And in a space flooded with blackbox models and hype, that clarity hits different.

✈️ Starting the day in Istanbul. Let's talk AI. The future of AI won't be shaped by size, but by precision. Lately I've been diving into what OpenLedger has been building, and I think we're witnessing one of the most important hard forks in AI: 👉 From giant generalist models 👉 To focused, hightrust AI agents Here's why that shift matters , and why OpenLedger's vision makes perfect sense: ✅ Specialized > Generalized General AI can do many things. But specialized AI? It does one thing extraordinarily well. Tailored models don't waste compute on irrelevant context, every parameter is purposedriven. ✅ Explainability isn't optional anymore In highstakes sectors like finance or healthcare, because the model said so won't cut it. We need transparent reasoning paths. Models must show how they reached conclusions , not just what they concluded. ✅ Trust comes from traceability With OpenLedger's Proof of Attribution, each AI decision is traceable, verifiable, and tamperproof. We're talking onchain records of who contributed what , accountability by design. ✅ Less hallucination, more signal Smaller, specialized models trained on clean, domainspecific data are far less prone to hallucinations. Clear boundaries = higher reliability. ✅ Efficiency is the real scalability Deploying one massive model for everything? Expensive, slow, and unsustainable. Specialized AI is leaner, faster, and far more costeffective. In short, OpenLedger isn't just following a trend. They're laying the rails for the infrastructure layer of verifiable, domainaware AI. And in a space flooded with blackbox models and hype, that clarity hits different.

Crypto Sinan

13,636 views • 11 months ago

We asked (Co-Founder & CEO, webAI) about the advantages of running AI on devices. "There's no free lunch, but there are things cloud players can't do. It's RAM intensive, yes, but by distributing across devices, we can unlock more accuracy and context. But if you're hosting for a million users on Nvidia, you can't do that." "With our stack, everything's downstream only. Think about health data, like you want that to be private. The dream is personalized models for millions that never leave their device."

We asked (Co-Founder & CEO, webAI) about the advantages of running AI on devices. "There's no free lunch, but there are things cloud players can't do. It's RAM intensive, yes, but by distributing across devices, we can unlock more accuracy and context. But if you're hosting for a million users on Nvidia, you can't do that." "With our stack, everything's downstream only. Think about health data, like you want that to be private. The dream is personalized models for millions that never leave their device."

TBPN

10,084 views • 11 months ago

Spent the last few days vibe coding on my NVIDIA DGX Spark. Here's what I learned. Qwen 3.5 122B took one minute and nine seconds to respond "Hi how are you doing". Unusable for vibe coding. Gemma 4 was fast but built a dot instead of a first person shooter game. GPT-OSS 120B was the sweet spot. Fast, capable, and actually produced working HTML. Open source models running locally are not replacing Claude Opus 4.6 or Codex with GPT 5.4. Not even close. But they're getting better every month. The new DGX Spark Bench is live on Real-world benchmarks for local models on local hardware. This is just the start. Full video below.

Spent the last few days vibe coding on my NVIDIA DGX Spark. Here's what I learned. Qwen 3.5 122B took one minute and nine seconds to respond "Hi how are you doing". Unusable for vibe coding. Gemma 4 was fast but built a dot instead of a first person shooter game. GPT-OSS 120B was the sweet spot. Fast, capable, and actually produced working HTML. Open source models running locally are not replacing Claude Opus 4.6 or Codex with GPT 5.4. Not even close. But they're getting better every month. The new DGX Spark Bench is live on Real-world benchmarks for local models on local hardware. This is just the start. Full video below.

BridgeMind

57,304 views • 3 months ago

Excited to share a glimpse of what’s possible with specialized models on OpenLedger🐙 In this video, you’ll see a chat agent in action—think of it as the Perplexity of crypto. Powered by tool calling, this specialized model showcases how OpenLedger enables domain-specific intelligence for AI-driven apps. Vision of these specialised models is to use english to interact, query, and transact on-chain.🐙

Excited to share a glimpse of what’s possible with specialized models on OpenLedger🐙 In this video, you’ll see a chat agent in action—think of it as the Perplexity of crypto. Powered by tool calling, this specialized model showcases how OpenLedger enables domain-specific intelligence for AI-driven apps. Vision of these specialised models is to use english to interact, query, and transact on-chain.🐙

OpenLedger

87,899 views • 1 year ago

UPDATE: Four new open models on the Text to Speech Arena! 🔥 *sound on🔉* As the Text-to-Speech ecosystem is heating up, we decided to add more competition. > Parler TTS > VoiceCraft > Vokan > GPT-SOVITS Why is this important? The TTS ecosystem is riddled with opaque metrics and meaningless MOS scores. By crowdsourcing the evals, we test these models in real-life conditions and much more methodically. ⚡ Rank one, rank'em all! 🚀

UPDATE: Four new open models on the Text to Speech Arena! 🔥 sound on🔉 As the Text-to-Speech ecosystem is heating up, we decided to add more competition. > Parler TTS > VoiceCraft > Vokan > GPT-SOVITS Why is this important? The TTS ecosystem is riddled with opaque metrics and meaningless MOS scores. By crowdsourcing the evals, we test these models in real-life conditions and much more methodically. ⚡ Rank one, rank'em all! 🚀

Vaibhav (VB) Srivastav

40,944 views • 2 years ago

offers full-stack privacy across queries, data, payments and devices / hardware 1. The best open-source models (eg. Kimi K2.5) served in secure enclaves (TEEs) 2. The best closed-source models (eg. GPT-5.2 Pro) with queries pooled into proxy servers and PII anonymized with a specialized Silo model 3. The only available implementation of private Deep Research for secure reasoning 4. Fully private crypto payments via Zcash or discounted payments with FAI 5. A hardware offering to run everything on-prem from your own home with 288 GB GPU RAM (Silo Box) All with encrypted sync between devices so users can uniformly + privately access their AI sessions

offers full-stack privacy across queries, data, payments and devices / hardware 1. The best open-source models (eg. Kimi K2.5) served in secure enclaves (TEEs) 2. The best closed-source models (eg. GPT-5.2 Pro) with queries pooled into proxy servers and PII anonymized with a specialized Silo model 3. The only available implementation of private Deep Research for secure reasoning 4. Fully private crypto payments via Zcash or discounted payments with FAI 5. A hardware offering to run everything on-prem from your own home with 288 GB GPU RAM (Silo Box) All with encrypted sync between devices so users can uniformly + privately access their AI sessions

Freysa

16,647 views • 4 months ago

Sam Altman recently said: “GPT-OSS has strong real-world performance comparable to o4-mini—and you can run it locally on your phone.” Many believed running a 20B-parameter model on mobile devices was still years away. At Nexa AI, we’ve built our foundation on deep on-device AI technology—turning that vision into reality. Today, GPT-OSS is running fully local on mobile devices through our app, Nexa Studio. Real performance on Snapdragon Gen 5: - 17 tokens/sec decoding speed - < 3 seconds Time-to-First-Token Developers can now use NexaSDK to build their own local AI apps powered by GPT-OSS. What this unlocks: - Real reasoning models running locally - The next wave of AI Agent use cases—without cloud limits - True data privacy: your data never leaves your device Appreciation to OpenAI Sam Altman Greg Brockman and the entire leadership team for pushing the boundaries of the open-source AI community. Thanks to Qualcomm for the partnership, and to Manoj Khilnani and Chun-Po (Jerry) Chang for the incredible collaboration.

Sam Altman recently said: “GPT-OSS has strong real-world performance comparable to o4-mini—and you can run it locally on your phone.” Many believed running a 20B-parameter model on mobile devices was still years away. At Nexa AI, we’ve built our foundation on deep on-device AI technology—turning that vision into reality. Today, GPT-OSS is running fully local on mobile devices through our app, Nexa Studio. Real performance on Snapdragon Gen 5: - 17 tokens/sec decoding speed - < 3 seconds Time-to-First-Token Developers can now use NexaSDK to build their own local AI apps powered by GPT-OSS. What this unlocks: - Real reasoning models running locally - The next wave of AI Agent use cases—without cloud limits - True data privacy: your data never leaves your device Appreciation to OpenAI Sam Altman Greg Brockman and the entire leadership team for pushing the boundaries of the open-source AI community. Thanks to Qualcomm for the partnership, and to Manoj Khilnani and Chun-Po (Jerry) Chang for the incredible collaboration.

NEXA AI

197,453 views • 9 months ago

.The Information Reporter Stephanie Palazzolo on OpenAI's GPT-5—a major step up for coding and what it means for competitors. "GPT-5 is a step up on a lot of different domains...one area that really stood out to [sources] was with coding." "Not only is GPT-5 better on more academic...tasks, but it's also better on the more practical programming tasks...working with very large and complex code bases." "If GPT-5 is going to be significantly better at these more practical everyday programming tasks, that could prove to be bad news for Anthropic." Watch the full episode on

.The Information Reporter Stephanie Palazzolo on OpenAI's GPT-5—a major step up for coding and what it means for competitors. "GPT-5 is a step up on a lot of different domains...one area that really stood out to [sources] was with coding." "Not only is GPT-5 better on more academic...tasks, but it's also better on the more practical programming tasks...working with very large and complex code bases." "If GPT-5 is going to be significantly better at these more practical everyday programming tasks, that could prove to be bad news for Anthropic." Watch the full episode on

The Information

52,067 views • 1 year ago

GPT-5.6 Sol just changed how I build websites with AI. In this full 10-minute video, I test Sol and Terra inside Codex to build a cinematic landing page from a video reference. I cover: → Which GPT-5.6 model to use → How I analyze a full website experience → How I recreate the layout, motion and interactions → Why better prompting matters more than one-shot results → How to avoid generic AI slop The model matters—but the way you guide it matters even more. Watch the full process below 👇 Want the free template and the exact prompt I used? They’re available exclusively in the description of the YouTube version:

GPT-5.6 Sol just changed how I build websites with AI. In this full 10-minute video, I test Sol and Terra inside Codex to build a cinematic landing page from a video reference. I cover: → Which GPT-5.6 model to use → How I analyze a full website experience → How I recreate the layout, motion and interactions → Why better prompting matters more than one-shot results → How to avoid generic AI slop The model matters—but the way you guide it matters even more. Watch the full process below 👇 Want the free template and the exact prompt I used? They’re available exclusively in the description of the YouTube version:

Sourasith

261,047 views • 19 days ago

Never bet against Cursor. Subagents solve so many issues between different models it's kind of insane, especially with Composer 1. Composer 1 can quickly gather context for a model like GPT-5.2 Codex, which otherwise could take on the order of 10 mins just to search files. Definitely one of the most underrated models right now. Now you get the benefits of it's speed, but can handoff to a smarter model at implementation. The video I attached is just a simple example, but you can see how long GPT 5.2 takes to gather context on a very simple problem. For harder problems, the search time grows significantly.

Never bet against Cursor. Subagents solve so many issues between different models it's kind of insane, especially with Composer 1. Composer 1 can quickly gather context for a model like GPT-5.2 Codex, which otherwise could take on the order of 10 mins just to search files. Definitely one of the most underrated models right now. Now you get the benefits of it's speed, but can handoff to a smarter model at implementation. The video I attached is just a simple example, but you can see how long GPT 5.2 takes to gather context on a very simple problem. For harder problems, the search time grows significantly.

ostyn

19,631 views • 6 months ago

$AI agents are about to redefine the internet. The mistake we made with Large Language Models? We let a handful of corporations capture all the value. Action Model is building a different path. By training through our extension, users gain fractional ownership in the Large Action Model, giving them a real stake in the future of AI. When LLMs emerged, the upside flowed to Big Tech. This time, it doesn’t have to. They’re building AI on our data, and keeping the upside for themselves. Community-owned Large Action Model is how we take it back.$

AI agents are about to redefine the internet. The mistake we made with Large Language Models? We let a handful of corporations capture all the value. Action Model is building a different path. By training through our extension, users gain fractional ownership in the Large Action Model, giving them a real stake in the future of AI. When LLMs emerged, the upside flowed to Big Tech. This time, it doesn’t have to. They’re building AI on our data, and keeping the upside for themselves. Community-owned Large Action Model is how we take it back.

Action Model

77,049 views • 5 months ago

We made a thing! Very happy to announce sqlcoder-pro and the Defog Alignment Platform. Available to use immediately without a wait-list, weights will be open-sourced very soon. The video does a quick show and tell comparison against ChatGPT (with gpt-4o). Read on for more details! TLDR 💪 equal (or better) performance on text-to-SQL as the most capable Claude-3.5 or GPT-4 models 🤝 You can use it today on a free plan/free trial, without a waitlist 🪽 self-hostable on a single RTX4090, with 2 second median generation times for SQL queries 🔁 exactly the same output every time, give the same prompt 👨🏻‍🏫 teachable and steerable: show the model what you want it to do 🛞 debuggable – you can understand WTF is going on inside the model, instead of treating it like a black box Let's dig into each of these one-by-one! Performance SQLCoder-8b-pro significantly exceeds the performance of our previous sqlcoder-8b model on Postgres text-to-SQL (from 88.2% to 90.2% accuracy - gpt-4o is at 87.6%, for reference). It is also better at following instructions. This was done via self-merges, hand crafted fine-tuning data, and adapting the training data to fit our tokenizer. Cost You can host this on the model on a single $3,500 RTX4090, and support ~5 requests/second via VLLM. If you're looking to host on the cloud instead, you can run it on a single L4 GPU that costs $300/mo on GCP Repeatability We have a dense 8b model with no MoE shenanigans. For the same prompt with temperature=0, you'll always get the same answer – which is critical in BI. Teachable In our alignment and feedback modes, you can give the model feedback on how it answered certain questions, and it will automatically adapt to the feedback. Debuggable You can use logprobs and attention scores to determine where, exactly is the model paying attention to inside a prompt + what it's getting confused by when generating outputs. Available today You can use Defog on the cloud today by going to docs[dot]defog[dot]ai, and getting an API key. Excited to hear what you think!

We made a thing! Very happy to announce sqlcoder-pro and the Defog Alignment Platform. Available to use immediately without a wait-list, weights will be open-sourced very soon. The video does a quick show and tell comparison against ChatGPT (with gpt-4o). Read on for more details! TLDR 💪 equal (or better) performance on text-to-SQL as the most capable Claude-3.5 or GPT-4 models 🤝 You can use it today on a free plan/free trial, without a waitlist 🪽 self-hostable on a single RTX4090, with 2 second median generation times for SQL queries 🔁 exactly the same output every time, give the same prompt 👨🏻‍🏫 teachable and steerable: show the model what you want it to do 🛞 debuggable – you can understand WTF is going on inside the model, instead of treating it like a black box Let's dig into each of these one-by-one! Performance SQLCoder-8b-pro significantly exceeds the performance of our previous sqlcoder-8b model on Postgres text-to-SQL (from 88.2% to 90.2% accuracy - gpt-4o is at 87.6%, for reference). It is also better at following instructions. This was done via self-merges, hand crafted fine-tuning data, and adapting the training data to fit our tokenizer. Cost You can host this on the model on a single $3,500 RTX4090, and support ~5 requests/second via VLLM. If you're looking to host on the cloud instead, you can run it on a single L4 GPU that costs $300/mo on GCP Repeatability We have a dense 8b model with no MoE shenanigans. For the same prompt with temperature=0, you'll always get the same answer – which is critical in BI. Teachable In our alignment and feedback modes, you can give the model feedback on how it answered certain questions, and it will automatically adapt to the feedback. Debuggable You can use logprobs and attention scores to determine where, exactly is the model paying attention to inside a prompt + what it's getting confused by when generating outputs. Available today You can use Defog on the cloud today by going to docs[dot]defog[dot]ai, and getting an API key. Excited to hear what you think!

Rishabh Srivastava

13,460 views • 1 year ago