目录
AI基础设施
条目：50
2026年三月
3 篇
| 类型 | 阅读 | 条目 |
|---|---|---|
[自动]
[BLOGS_PODCASTS] | 2min | mic
SemiAnalysis谈2000亿美元AI支出与谷歌2027年盈利前景 03-01
SemiAnalysis
AI资本支出
谷歌 |
[自动]
[BLOGS_PODCASTS] | 2min | mic
SemiAnalysis创始人谈2000亿美元AI资本开支与2027年谷歌盈利前景 03-01
SemiAnalysis
Dylan Patel
资本开支 |
[自动]
[BLOGS_PODCASTS] | 3min | mic
OpenAI与亚马逊达成战略合作，在AWS部署Frontier模型 03-01
OpenAI
AWS
亚马逊 |
2026年二月
42 篇
| 类型 | 阅读 | 条目 |
|---|---|---|
[自动]
[BLOGS_PODCASTS] | 3min | mic
OpenAI与亚马逊达成战略合作：在AWS上引入Frontier平台扩展AI基础设施 02-28
OpenAI
AWS
亚马逊 |
[自动]
[BLOGS_PODCASTS] | 3min | mic
OpenAI与亚马逊战略合作：将Frontier模型引入AWS 02-28
OpenAI
AWS
亚马逊 |
[自动]
[BLOGS_PODCASTS] | 3min | mic
OpenAI与亚马逊达成战略合作：Frontier平台接入AWS 02-27
OpenAI
AWS
亚马逊 |
[自动]
[BLOGS_PODCASTS] | 3min | mic
OpenAI与亚马逊达成战略合作，将Frontier模型引入AWS 02-27
OpenAI
AWS
亚马逊 |
[自动]
[BLOGS_PODCASTS] | 3min | mic
OpenAI与亚马逊达成战略合作，将Frontier模型引入AWS 02-27
OpenAI
AWS
亚马逊 |
[自动]
[BLOGS_PODCASTS] | 5min | mic
Meta 开源 RCCLX：优化 AMD 平台 GPU 通信性能 02-25
Meta
RCCLX
AMD |
[自动]
[BLOGS_PODCASTS] | 3min | mic
亚马逊Bedrock新推亚太六区：Anthropic Claude模型支持全球跨区域推理 02-24
Amazon Bedrock
Anthropic
Claude |
[自动]
[BLOGS_PODCASTS] | 4min | mic
Hexagon 利用 SageMaker HyperPod 加速分割模型预训练 02-24
SageMaker
HyperPod
AWS |
[自动]
[BLOGS_PODCASTS] | 2min | mic
Taalas HC1 芯片实测：Llama 3.1 8B 跑出 16960 tok/s 02-24
Taalas
定制芯片
ASIC |
[自动]
[BLOGS_PODCASTS] | 3min | mic
GGML与llama.cpp加入HF以保障本地AI长期发展 02-23
GGML
llama.cpp
Hugging Face |
[自动]
[BLOGS_PODCASTS] | 3min | mic
GGML与llama.cpp加入HF推动本地AI长期发展 02-23
GGML
llama.cpp
Hugging Face |
[自动]
[BLOGS_PODCASTS] | 4min | mic
GGML与llama.cpp加入HF以保障本地AI长期发展 02-23
GGML
llama.cpp
Hugging Face |
[自动]
[BLOGS_PODCASTS] | 2min | mic
a16z深度对话：Anthropic与OpenAI的博弈、Noam Shazeer及AI创业趋势 02-23
a16z
Anthropic
OpenAI |
[自动]
[BLOGS_PODCASTS] | 3min | mic
GGML与llama.cpp加入HF以保障本地AI长期发展 02-22
GGML
llama.cpp
Hugging Face |
[自动]
[BLOGS_PODCASTS] | 3min | mic
GGML与llama.cpp加入Hugging Face推动本地AI发展 02-21
GGML
llama.cpp
Hugging Face |
[自动]
[BLOGS_PODCASTS] | 3min | mic
a16z深度对话：Anthropic与OpenAI的博弈、Noam Shazeer及AI基础设施投资 02-21
a16z
Anthropic
OpenAI |
[自动]
[HACKER_NEWS] | 5min | newspaper
Ggml.ai加入Hugging Face推动本地AI长期发展 02-21
GGML
Hugging Face
本地AI |
[自动]
[HACKER_NEWS] | 5min | newspaper
Ggml.ai加入Hugging Face以推动本地AI长期发展 02-21
GGML
Hugging Face
本地AI |
[自动]
[HACKER_NEWS] | 5min | newspaper
Ggml.ai 加入 Hugging Face 推动本地 AI 长期发展 02-21
Hugging Face
GGML
本地AI |
[自动]
[HACKER_NEWS] | 5min | newspaper
通往普及AI之路：实现每秒1.7万tokens推理 02-21
推理加速
Token生成
性能优化 |
[自动]
[BLOGS_PODCASTS] | 4min | mic
GGML与llama.cpp加入HF以推动本地AI长期发展 02-21
GGML
llama.cpp
Hugging Face |
[自动]
[HACKER_NEWS] | 5min | newspaper
迈向通用AI：17k tokens/sec的推理性能路径 02-20
推理性能
Token生成
通用AI |
[自动]
[BLOGS_PODCASTS] | 5min | mic
GGML与llama.cpp加入HF以保障本地AI长期发展 02-20
GGML
llama.cpp
Hugging Face |
[自动]
[HACKER_NEWS] | 6min | newspaper
Ggml.ai加入Hugging Face以推动本地AI长期发展 02-20
GGML
Hugging Face
本地AI |
[自动]
[HACKER_NEWS] | 4min | newspaper
Ggml.ai加入Hugging Face以推动本地AI长期发展 02-20
GGML
Hugging Face
本地AI |
[自动]
[BLOGS_PODCASTS] | 4min | mic
a16z对话：Anthropic与OpenAI的博弈及AI基础设施投资逻辑 02-20
a16z
投资逻辑
算力 |
[自动]
[HACKER_NEWS] | 5min | newspaper
通往泛在AI之路：实现每秒1.7万tokens推理 02-20
推理加速
Token生成
泛在AI |
[自动]
[HACKER_NEWS] | 4min | newspaper
通往无处不在的AI：实现每秒1.7万tokens推理 02-20
LLM
推理加速
Token |
[自动]
[BLOGS_PODCASTS] | 3min | mic
a16z合伙人解析AI资本飞wheel：算力、人才与资本争夺战 02-19
a16z
资本飞轮
AI基础设施 |
[自动]
[BLOGS_PODCASTS] | 2min | mic
Jeff Dean：重写搜索栈、复兴稀疏模型与TPU协同设计 02-18
Jeff Dean
TPU
稀疏模型 |
[自动]
[BLOGS_PODCASTS] | 4min | mic
Jeff Dean：重塑搜索、TPU与稀疏模型的AI技术栈 02-17
Jeff Dean
Google
TPU |
[自动]
[BLOGS_PODCASTS] | 3min | mic
Jeff Dean：重塑搜索架构、复兴稀疏模型与设计TPU 02-17
Jeff Dean
Google
TPU |
[自动]
[BLOGS_PODCASTS] | 2min | mic
Codex与Claude支持定制内核以适配各类应用 02-14
Claude
Codex
定制内核 |
[自动]
[BLOGS_PODCASTS] | 2min | mic
Jeff Dean：重塑Google搜索栈与TPU联合设计之路 02-14
Jeff Dean
TPU
稀疏模型 |
[自动]
[BLOGS_PODCASTS] | 3min | mic
Jeff Dean：重塑Google搜索架构与TPU及稀疏模型的技术历程 02-13
Jeff Dean
Google
TPU |
[自动]
[BLOGS_PODCASTS] | 3min | mic
Jeff Dean：重塑谷歌搜索架构与TPU及稀疏模型的技术演进 02-13
Jeff Dean
Google
TPU |
[自动]
[HACKER_NEWS] | 7min | newspaper
仅更换框架，一下午提升15个大模型代码能力 02-13
LLM
代码生成
模型评估 |
[自动]
[BLOGS_PODCASTS] | 3min | mic
Jeff Dean：重写谷歌搜索栈与TPU共设计之路 02-12
Jeff Dean
Google
TPU |
[自动]
[HACKER_NEWS] | 4min | newspaper
仅替换调度框架，一下午提升15个大模型编程能力 02-12
LLM
代码生成
调度框架 |
[自动]
[BLOGS_PODCASTS] | 3min | mic
Goodfire AI：打造首个机制可解释性实验室与企业级API 02-10
Goodfire AI
机制可解释性
Mechanistic Interpretability |
[自动]
[BLOGS_PODCASTS] | 3min | mic
AI vs SaaS：从OpenClaw到Cursor的AI中心化演进 02-09
AI中心化
OpenClaw
Cursor |
[自动]
[BLOGS_PODCASTS] | 3min | mic
ElevenLabs获5亿美元融资，Cerebras估值达230亿美元 02-05
ElevenLabs
Cerebras
融资 |
2026年一月
5 篇
| 类型 | 阅读 | 条目 |
|---|---|---|
[自动]
[BLOGS_PODCASTS] | 3min | mic
Moltbook：首个面向 AI 智能体的社交网络 01-31
AI Agents
社交网络
Moltbook |
[自动]
[HACKER_NEWS] | 4min | newspaper
OpenAI 与英伟达价值千亿美元芯片交易暂停 01-31
OpenAI
英伟达
Nvidia |
[自动]
[HACKER_NEWS] | 4min | newspaper
OpenAI 与英伟达百亿美元芯片采购谈判暂停 01-31
OpenAI
英伟达
芯片 |
[自动]
[HACKER_NEWS] | 5min | newspaper
OpenAI 与英伟达价值千亿美元芯片交易搁浅 01-31
OpenAI
英伟达
Nvidia |
[自动]
[HACKER_NEWS] | 7min | newspaper
ShapedQL：支持多阶段排序与RAG的SQL引擎 01-29
ShapedQL
SQL引擎
RAG |
无匹配条目