LogoThread Easy
  • Explorar
  • Criar thread
LogoThread Easy

Seu parceiro completo para threads do Twitter

© 2025 Thread Easy All Rights Reserved.

Explorar

Newest first — browse tweet threads

Keep on to blur preview images; turn off to show them clearly

even the competitors are moved to tears by their graceful gift to Sama on the anniversary of ChatGPT. Speciale, truly

even the competitors are moved to tears by their graceful gift to Sama on the anniversary of ChatGPT. Speciale, truly

We're in a race. It's not USA vs China but humans and AGIs vs ape power centralization. @deepseek_ai stan #1, 2023–Deep Time «C’est la guerre.» ®1

avatar for Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)
Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)
Mon Dec 01 12:55:18
RT @JZ281C: @teortaxesTex Ground based solar installation is one of the easiest to automate among all energy infrastructure. It's easier th…

RT @JZ281C: @teortaxesTex Ground based solar installation is one of the easiest to automate among all energy infrastructure. It's easier th…

We're in a race. It's not USA vs China but humans and AGIs vs ape power centralization. @deepseek_ai stan #1, 2023–Deep Time «C’est la guerre.» ®1

avatar for Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)
Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)
Mon Dec 01 12:52:42
> dsv3.2 speciale version thinks much more than any other model
On 4/8 evals V3.2-Speciale thinks *less* than Kimi-K2-Thinking. On 3, more; for codeforces no Kimi datapoint
A 7 months older, 33% smaller, AdamW checkpoint with gimped attention, explicitly let loose, crushes K2

> dsv3.2 speciale version thinks much more than any other model On 4/8 evals V3.2-Speciale thinks *less* than Kimi-K2-Thinking. On 3, more; for codeforces no Kimi datapoint A 7 months older, 33% smaller, AdamW checkpoint with gimped attention, explicitly let loose, crushes K2

We're in a race. It's not USA vs China but humans and AGIs vs ape power centralization. @deepseek_ai stan #1, 2023–Deep Time «C’est la guerre.» ®1

avatar for Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)
Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)
Mon Dec 01 12:51:32
刚刚,DeepSeek 发布了最新 V3.2 模型,主打高效推理与强大的 Agent 能力。

引入稀疏注意力(DSA) 机制,在保持模型性能的同时大幅降低计算复杂度,专门对长上下文场景做了优化。

在大规模 Agent 任务合成中实现了突破,将推理能力整合进工具调用场景,显著提升模型在复杂交互环境下的表现。

模型地址:https://t.co/VwGFgEylOz

同时,还推出了一个增强版本 DeepSeek-V3.2-Speciale 表现更是亮眼。

在强化学习框架加持下,推理能力号称超越 GPT-5 与 Gemini-3.0-pro  相当,并在 IMO 2025 和 IOI 等国际竞赛中斩获金牌。

模型地址:https://t.co/yWt5pX9UH2

除此之外,还更新了 Chat Template 以支持 thinking with tools 能力,新增 developer 角色专用于搜索代理。

总的来看,DeepSeek 这一次升级相当硬核,值得大家关注一下。

刚刚,DeepSeek 发布了最新 V3.2 模型,主打高效推理与强大的 Agent 能力。 引入稀疏注意力(DSA) 机制,在保持模型性能的同时大幅降低计算复杂度,专门对长上下文场景做了优化。 在大规模 Agent 任务合成中实现了突破,将推理能力整合进工具调用场景,显著提升模型在复杂交互环境下的表现。 模型地址:https://t.co/VwGFgEylOz 同时,还推出了一个增强版本 DeepSeek-V3.2-Speciale 表现更是亮眼。 在强化学习框架加持下,推理能力号称超越 GPT-5 与 Gemini-3.0-pro 相当,并在 IMO 2025 和 IOI 等国际竞赛中斩获金牌。 模型地址:https://t.co/yWt5pX9UH2 除此之外,还更新了 Chat Template 以支持 thinking with tools 能力,新增 developer 角色专用于搜索代理。 总的来看,DeepSeek 这一次升级相当硬核,值得大家关注一下。

💡 挖掘开源的价值 🧑🏻‍💻 坚持分享 GitHub 上高质量、有趣、实用的教程、AI工具、前沿 AI 技术 🧐 A list cool, interesting projects of GitHub. ✏️ 公众号:GitHubDaily

avatar for GitHubDaily
GitHubDaily
Mon Dec 01 12:44:57
I built a marketplace to buy/sell startups:

📋 Founders list startups with open metrics
👀 Buyers make private offers
🤝 TrustMRR connects both

I sold 3 startups in 2023, and it changed my life because it gave me enough bandwidth to focus on the next thing for a few months.

So I built this marketplace to help founders do the same.

Link below ↓

I built a marketplace to buy/sell startups: 📋 Founders list startups with open metrics 👀 Buyers make private offers 🤝 TrustMRR connects both I sold 3 startups in 2023, and it changed my life because it gave me enough bandwidth to focus on the next thing for a few months. So I built this marketplace to help founders do the same. Link below ↓

→ https://t.co/RqaLO6oMGz If you want to sell your startup: 1. List your startup on TrustMRR 2. On the startup page, click "edit" and tick "list startup for sale" 3. I'll email you the acquisition offers Listings are sorted by most recent. If you buy/sell startups, please let me know your feedback!

avatar for Marc Lou
Marc Lou
Mon Dec 01 12:44:06
DeepSeek wishes ChatGPT a happy birthday by releasing model DeepSeek-v3.2-Speciale that exceeds both GPT-5-High and Gemini-3-Pro.

@teortaxesTex - Liang Wenfeng doesn't disappoint!

DeepSeek wishes ChatGPT a happy birthday by releasing model DeepSeek-v3.2-Speciale that exceeds both GPT-5-High and Gemini-3-Pro. @teortaxesTex - Liang Wenfeng doesn't disappoint!

Artificial Intelligence @amazon. RL, General Purpose Agents, OSS AI All views personal!

avatar for GDP at NeurIPS 2025
GDP at NeurIPS 2025
Mon Dec 01 12:42:39
  • Previous
  • 1
  • More pages
  • 1900
  • 1901
  • 1902
  • More pages
  • 5634
  • Next