Is This the End of LLMs? This Stanford Paper Could Make Nvidia's Memory Chips Unnecessary
Publicado em 8 de nov. de 2026 · 11:22
Curtidas: 23 milComentários: 1 mil+24h: +14 milMédia de visualizações/h: 6 mil
O desempenho em números
Números de 16 de nov. de 2026, 23:00 UTC.
Is This the End of LLMs? This Stanford Paper Could Make Nvidia's Memory Chips Unnecessary, de Alex - a1crew, tem 1 mi visualizações, 8 dias após a publicação em 8 de novembro de 2026.
Está recebendo 6 mil visualizações por hora agora.
Está nos vídeos em alta do YouTube em 1 país, no máximo #44 em Suécia.
Seu melhor dia até agora foi 10 de nov., com 919 mil visualizações; é mais assistido por volta das 19 UTC.
| Visualizações | 1 mi | 1.128.567 |
|---|---|---|
| Visualizações nas últimas 24 horas | +14 mil | +14.382 |
| Visualizações por hora agora | 6 mil | 5.537 |
| Vezes o normal do canal | 56× | 55,7× |
| Melhor dia | 10 de nov. · 919 mil | 919.027 |
| Horário mais forte | 19 UTC | |
| Países em alta | 1 | |
| Melhor posição | #44 · Suécia | |
| Curtidas | 23 mil | 23.238 |
| Comentários | 1 mil | 1.277 |
| Publicado | 8 de novembro de 2026 | |
| Duração | 11:22 | |
| Categoria | Science & Technology | |
| Canal | Alex - a1crew | |
| Inscritos do canal | 16 mil |
Visualizações
1 mi
+14 mil · 24 h
Visualizações/h agora
6 mil
× o habitual
56×
Ainda não há leituras suficientes.
Visualizações por dia
919 mil
melhor dia, 10 de nov.
Em alta em
1
melhor #44 em Suécia
Quando é assistido
19
seu horário mais forte, seu fuso
Onde aparece1
- #44Suécia9 de nov.
Do registro do YouTube
Campos do vídeo como a API do YouTube os retorna.
- Publicado
- 8 de nov. de 2026, 11:15
- Duração
- 11:22
- Categoria
- Science & Technology
- Idioma do áudio
- inglês (en)
- Idioma do título
- inglês (en)
- Legendas
- Sim
- Definição
- HD
- Formato
- 2D
- Licença
- Standard YouTube
- Conteúdo licenciado
- Sim
- Pode ser incorporado
- Sim
- Feito para crianças
- Não
- Restrito por idade
- Não
- Promoção paga
- Não
- Estatísticas públicas
- Sim
Temas
- Knowledge
- Technology
Título e descrição em 1 idiomas
- inglês
Gravado em
Canada · 56.130, -106.347
Tags (27)
- Next Concept Modeling
- NCM
- Stanford AI research
- Nvidia
- HBM
- high bandwidth memory
- memory bandwidth
- KV cache
- GPU memory
- VRAM
- AI hardware
- LLM inference
- how LLMs work
- token by token generation
- parallel text generation
- nested dropout
- conditioning corruption
- test-time scaling
- Minimum Bayes Risk
- MBR
- small language models
- AI reasoning
- AI Sudoku
- consumer GPU
- frontier AI
- local AI
- AI research explained
Descrição
What if a single Stanford research paper could make Nvidia's most expensive memory chips unnecessary? The AI industry has hit a wall, and it isn't the one everyone talks about. Today's language models write one token at a time, so a 1,000-word answer means 1,000 passes through the entire model, and every pass hammers memory bandwidth. That's why the KV cache eats your GPU memory and why High Bandwidth Memory became the most expensive bottleneck in AI hardware. In this video we look at Next Concept Modeling (NCM), a new architecture that generates the whole answer in parallel and needs 32 passes instead of 1,000. We unpack the three technical secrets that make it work, from nested dropout to conditioning corruption to test-time scaling with Minimum Bayes Risk, and a Sudoku experiment where two models of exactly the same size got very different results. Then the big question: if AI stops needing massive memory bandwidth and only needs raw compute, what happens to the hardware monopoly, and could frontier-level AI run on consumer-grade silicon? Do you think consumer graphics cards will finally run frontier models? Original Stanford paper: https://arxiv.org/html/2610.14162v1 00:00 Can one paper beat Nvidia's memory chips? 00:29 How language models write today 00:58 1,000 words, 1,000 passes 01:31 The processor that just waits 01:43 The KV cache problem 02:02 Expensive memory chips 02:17 Enter Next Concept Modeling 02:35 Borrowed from image generation 02:52 Text in layers 03:31 One pass per layer 03:56 1,000 passes vs 32 04:32 Secret #1: nested dropout 05:02 Subscribe 05:29 Learning true concepts 06:04 The vulnerability: compounding errors 06:31 Secret #2: conditioning corruption 06:56 An AI that fixes its own mistakes 07:15 Can a tiny model compete? 07:30 Secret #3: test-time scaling 07:59 Minimum Bayes Risk 08:10 20 juniors vs 1 senior 08:39 The Sudoku test 09:09 The billion-dollar hardware flip 09:24 No free lunch 09:58 Frontier AI on consumer hardware 10:09 Is token-by-token a relic? 10:24 Will it scale? 10:41 Recap 11:05 Your turn #AI #LLM #Nvidia #MachineLearning #AIHardwareMostrar tudo
What if a single Stanford research paper could make Nvidia's most expensive memory chips unnecessary? The AI industry has hit a wall, and it isn't the one everyone talks about. Today's language models write one token at a time, so a 1,000-word answer means 1,000 passes through the entire model, and every pass hammers memory bandwidth. That's why the KV cache eats your GPU memory and why High Bandwidth Memory became the most expensive bottleneck in AI hardware. In this video we look at Next Concept Modeling (NCM), a new architecture that generates the whole answer in parallel and needs 32 passes instead of 1,000. We unpack the three technical secrets that make it work, from nested dropout to conditioning corruption to test-time scaling with Minimum Bayes Risk, and a Sudoku experiment where two models of exactly the same size got very different results. Then the big question: if AI stops needing massive memory bandwidth and only needs raw compute, what happens to the hardware monopoly, and could frontier-level AI run on consumer-grade silicon? Do you think consumer graphics cards will finally run frontier models? Original Stanford paper: https://arxiv.org/html/2610.14162v1 00:00 Can one paper beat Nvidia's memory chips? 00:29 How language models write today 00:58 1,000 words, 1,000 passes 01:31 The processor that just waits 01:43 The KV cache problem 02:02 Expensive memory chips 02:17 Enter Next Concept Modeling 02:35 Borrowed from image generation 02:52 Text in layers 03:31 One pass per layer 03:56 1,000 passes vs 32 04:32 Secret #1: nested dropout 05:02 Subscribe 05:29 Learning true concepts 06:04 The vulnerability: compounding errors 06:31 Secret #2: conditioning corruption 06:56 An AI that fixes its own mistakes 07:15 Can a tiny model compete? 07:30 Secret #3: test-time scaling 07:59 Minimum Bayes Risk 08:10 20 juniors vs 1 senior 08:39 The Sudoku test 09:09 The billion-dollar hardware flip 09:24 No free lunch 09:58 Frontier AI on consumer hardware 10:09 Is token-by-token a relic? 10:24 Will it scale? 10:41 Recap 11:05 Your turn #AI #LLM #Nvidia #MachineLearning #AIHardware
◆ Calculado pelo viralshooter · todos os outros números vêm da API do YouTube