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

moon = less air resistance and lower gravity
AI robot crew = more robust and less fragile. You might need a human at some point, but robot is better than nothing!

moon = less air resistance and lower gravity AI robot crew = more robust and less fragile. You might need a human at some point, but robot is better than nothing!

ai agents @hud_evals | owned @AIHubCentral (1 million users,acq.) ex climate protester🦦 dont do the deferred life plan

avatar for Minh Nhat Nguyen
Minh Nhat Nguyen
Wed Dec 17 04:01:25
When i learned about china's space railgun, i wondered why SpaceX wasn't doing that. I figured it was mainly 1. Air resistance and gravity 2. Payload g

Automated moonbase manned ~mostly by AI/robots solves both, and a moon->mars railgun colonisation flywheel becomes feasible

When i learned about china's space railgun, i wondered why SpaceX wasn't doing that. I figured it was mainly 1. Air resistance and gravity 2. Payload g Automated moonbase manned ~mostly by AI/robots solves both, and a moon->mars railgun colonisation flywheel becomes feasible

moon = less air resistance and lower gravity AI robot crew = more robust and less fragile. You might need a human at some point, but robot is better than nothing!

avatar for Minh Nhat Nguyen
Minh Nhat Nguyen
Wed Dec 17 04:01:23
写 Prompt 最头疼的往往不是构思,而是反复的 “试错”。改个词效果变差,加句话逻辑变乱,把大量时间耗在了手动调试上。

偶然在 GitHub 上看到 Salesforce AI 研究院开源了 Promptomatix 这个框架,目的就是终结这种低效的手工调优。

通过分析任务需求自动生成训练数据,并根据反馈不断修正,自动迭代优化提示词,直到产出最佳结果。

GitHub:https://t.co/HHKRjgHrlR

支持 OpenAI、Anthropic 等多个模型提供商,提供命令行和 API 两种使用方式。

还提供了详细的 Jupyter 教程,从基础用法到高级定制都有覆盖,也可以直接用命令行快速上手。

对于正在开发 LLM 应用,或者深受 Prompt 调优之苦的朋友,这个自动化方案值得尝试。

写 Prompt 最头疼的往往不是构思,而是反复的 “试错”。改个词效果变差,加句话逻辑变乱,把大量时间耗在了手动调试上。 偶然在 GitHub 上看到 Salesforce AI 研究院开源了 Promptomatix 这个框架,目的就是终结这种低效的手工调优。 通过分析任务需求自动生成训练数据,并根据反馈不断修正,自动迭代优化提示词,直到产出最佳结果。 GitHub:https://t.co/HHKRjgHrlR 支持 OpenAI、Anthropic 等多个模型提供商,提供命令行和 API 两种使用方式。 还提供了详细的 Jupyter 教程,从基础用法到高级定制都有覆盖,也可以直接用命令行快速上手。 对于正在开发 LLM 应用,或者深受 Prompt 调优之苦的朋友,这个自动化方案值得尝试。

💡 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
Wed Dec 17 04:00:06
Thank you @SenatorRounds - we agree that the government should get the best and brightest to win this critical race.

Thank you @SenatorRounds - we agree that the government should get the best and brightest to win this critical race.

Dir., Office of Personnel Management (previously, MP at a16z); Author of Secrets of Sand Hill Road; father of three amazing/crazy/beautiful girls.

avatar for Scott Kupor
Scott Kupor
Wed Dec 17 03:59:10
There is nobody who works harder or more diligently in support of the policies that @realDonaldTrump was elected to pursue on behalf of the American people than @SusieWiles. Everyone who interacts with her recognizes her great leadership and remarkable contributions to the success of this Administration.

There is nobody who works harder or more diligently in support of the policies that @realDonaldTrump was elected to pursue on behalf of the American people than @SusieWiles. Everyone who interacts with her recognizes her great leadership and remarkable contributions to the success of this Administration.

Dir., Office of Personnel Management (previously, MP at a16z); Author of Secrets of Sand Hill Road; father of three amazing/crazy/beautiful girls.

avatar for Scott Kupor
Scott Kupor
Wed Dec 17 03:59:10
What we call self isn't a thing — it's a pattern of correlations, a consistent history that emerges when our neural subsystems interact.

"Free will" is the narrative this pattern of correlations tells about itself. When an integrated system models its own decisionmaking, it cannot see the pattern of correlations of which it is itself constituted.  The narrative it tells about itself is a compression that elides over the causal mechanisms, leaving only a compact summary: "I wanted X, I did X."

The feeling of agency is what it's like to be a model that includes its own actions as predicted outputs. That feeling is valid for the model even as it is also true for the rest of the world that everything happening is happening according to the laws of physics, which are deterministic. Both experiences are valid. Neither is more (or less) real.

The self is the story a model pays in entropy to tell about itself to the rest of the world.

What we call self isn't a thing — it's a pattern of correlations, a consistent history that emerges when our neural subsystems interact. "Free will" is the narrative this pattern of correlations tells about itself. When an integrated system models its own decisionmaking, it cannot see the pattern of correlations of which it is itself constituted. The narrative it tells about itself is a compression that elides over the causal mechanisms, leaving only a compact summary: "I wanted X, I did X." The feeling of agency is what it's like to be a model that includes its own actions as predicted outputs. That feeling is valid for the model even as it is also true for the rest of the world that everything happening is happening according to the laws of physics, which are deterministic. Both experiences are valid. Neither is more (or less) real. The self is the story a model pays in entropy to tell about itself to the rest of the world.

Wonderer. Amor fati. Scaling trust.

avatar for Michael Frank Martin
Michael Frank Martin
Wed Dec 17 03:54:13
  • Previous
  • 1
  • More pages
  • 554
  • 555
  • 556
  • More pages
  • 5634
  • Next