LogoThread Easy
  • 探索
  • 線程創作
LogoThread Easy

Twitter 線程的一站式夥伴

© 2025 Thread Easy All Rights Reserved.

探索

Newest first — browse tweet threads

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

这几天跟大家一起研究各种X 账号的风险控制， 又想起来几年前读过的一篇科幻小说《镜子》 ， 这篇小说被收录在刘慈欣的科幻小说集《时间移民》 中， 非常有趣， 大刘在小说里探讨了人类社会能真正运维是究竟是靠的什么。 不是世界太黑暗， 而是我们无法承受绝对的光明。

这几天跟大家一起研究各种X 账号的风险控制， 又想起来几年前读过的一篇科幻小说《镜子》 ， 这篇小说被收录在刘慈欣的科幻小说集《时间移民》 中， 非常有趣， 大刘在小说里探讨了人类社会能真正运维是究竟是靠的什么。 不是世界太黑暗， 而是我们无法承受绝对的光明。

Solo Developer ｜ Growth Coach｜Helping creators build their personal brand on X 公众号：PandaTalk8

avatar for Mr Panda
Mr Panda
Fri Nov 28 03:46:25
🚀 上新新图片模型 Flux 2 Flex 我记得是 Flux 开源之后，图片模型才迎来起飞的 没想到短短 1 年，各家都已经赶上了啊～ 而 Flux 2 效果看起来很好，但不够惊艳 更何况 Nano Banana 的图片生成能力， 是包含了 Gemini 3 的世界知识和思考模式。 https://t.co/tA79RkoK94

🚀 上新新图片模型 Flux 2 Flex 我记得是 Flux 开源之后，图片模型才迎来起飞的 没想到短短 1 年，各家都已经赶上了啊～ 而 Flux 2 效果看起来很好，但不够惊艳 更何况 Nano Banana 的图片生成能力， 是包含了 Gemini 3 的世界知识和思考模式。 https://t.co/tA79RkoK94

🚧 building https://t.co/AJfZ3LMlgq https://t.co/606cFUoda3 https://t.co/s0m0tpQMDH https://t.co/UQ5vrrYdAG 🐣learning/earning while helping others ❤️making software, storytelling videos 🔙alibaba @thoughtworks

avatar for 吕立青_JimmyLv (闭关ing) 2𐃏25
吕立青_JimmyLv (闭关ing) 2𐃏25
Fri Nov 28 03:45:05
中国人出海可真不用容易： 1. 要解决上网的问题 2. 要解决手机号的问题 3. 要解决身份认证的问题 1. 上网问题， 对于大多数非技术背景的人来讲，已经很难了。 2. 要养个美国的手机号 。 3. 护照、港澳通行证、各类签证。 护照还是比较有用的， 成为一个国际人，护照才是全世界最通用的证件。

中国人出海可真不用容易： 1. 要解决上网的问题 2. 要解决手机号的问题 3. 要解决身份认证的问题 1. 上网问题， 对于大多数非技术背景的人来讲，已经很难了。 2. 要养个美国的手机号 。 3. 护照、港澳通行证、各类签证。 护照还是比较有用的， 成为一个国际人，护照才是全世界最通用的证件。

Solo Developer ｜ Growth Coach｜Helping creators build their personal brand on X 公众号：PandaTalk8

avatar for Mr Panda
Mr Panda
Fri Nov 28 03:35:01
open questions on chatgpt: • what is the best way to morph from a utility app to a habit app? what's the ceiling on engagement for a text-first utility product at scale? • are there ad models OpenAI can operate which Meta cannot copy within a few quarters? which subset of OpenAI's data may be more valuable? • how much revenue can come from the current set of affiliates? how do you avoid incentive conflicts over time (e.g. open ChatGPT to order DoorDash vs. open DoorDash to order DoorDash)? are many partnerships on borrowed time? • how will the 'long tail' of commerce (Shopify and Etsy, which OpenAI already has) compare to the Amazons/Walmarts over the next decade? could 'specialized AI merchandise' (e.g. send me a physical photo frame with my family in it, but anime-style) be a large enough category to notably shift current trajectories? • how can pushing multimodal content be elegantly integrated into the main app (as opposed to pulling/prompting, i.e. slowly changing ChatGPT into a personalized image/video/e2e voice feed (what is everyone doing here! come on!))? especially given the expansive room competitors already have coming from the opposite wedge (e.g. Instagram can add an AI assistant much more easily than ChatGPT can add an Instagram). is this truly the only path forward, or can Pulse-like proactive utility features matter for the average consumer? • when will Meta do the thing? is the potential better for them the longer they wait, or worse? • group chats (and similar 'true social' features) remain a path to explore to avoid some challenging trade-offs above. how sticky can OpenAI make its distribution here? is the winning move to merge both social and entertainment? if so, what stops Meta from copying and out-distributing? • when will OpenAI do the thing? can hardware help? • Google already has most pieces for the above - is Google's innovator's dilemma surmountable, particularly with Search, YouTube Shorts, and Gemini? given their antitrust wins, will Gemini's distribution become even more cutthroat? will consumers ever know what Nano Banana Pro is to the extent they do ChatGPT? Is Gemini O.K.? do Google's AI Overviews truly not cannibalize ad clicks? • will parts of Meta or Google ever suffer engagement losses due to the onslaught of AI slop? if so, will more users flock to a more personal, trusted, and walled garden-esque ChatGPT? • could inference improvements and model efficiency play a larger role than free-tier monetization? most consumers need little intelligence for their problems - when will most inference end up on-device, shifting compute costs to consumer devices? how much of this is solved at the software level via model routing? • should I include a question about Apple? why or why not? • Is the above discounting OpenAI's moonshot efforts and ongoing enterprise deals too much? Should we treat OpenAI's 'Meta-ification' as a make or break inflection point given the cultural changes, or instead view it as an intelligent hedge which helps bridge the next few years? • Fidji saved Instacart, guiding it from an unprofitable pandemic darling into an enduring IPO. Will Fidji save OpenAI?

open questions on chatgpt: • what is the best way to morph from a utility app to a habit app? what's the ceiling on engagement for a text-first utility product at scale? • are there ad models OpenAI can operate which Meta cannot copy within a few quarters? which subset of OpenAI's data may be more valuable? • how much revenue can come from the current set of affiliates? how do you avoid incentive conflicts over time (e.g. open ChatGPT to order DoorDash vs. open DoorDash to order DoorDash)? are many partnerships on borrowed time? • how will the 'long tail' of commerce (Shopify and Etsy, which OpenAI already has) compare to the Amazons/Walmarts over the next decade? could 'specialized AI merchandise' (e.g. send me a physical photo frame with my family in it, but anime-style) be a large enough category to notably shift current trajectories? • how can pushing multimodal content be elegantly integrated into the main app (as opposed to pulling/prompting, i.e. slowly changing ChatGPT into a personalized image/video/e2e voice feed (what is everyone doing here! come on!))? especially given the expansive room competitors already have coming from the opposite wedge (e.g. Instagram can add an AI assistant much more easily than ChatGPT can add an Instagram). is this truly the only path forward, or can Pulse-like proactive utility features matter for the average consumer? • when will Meta do the thing? is the potential better for them the longer they wait, or worse? • group chats (and similar 'true social' features) remain a path to explore to avoid some challenging trade-offs above. how sticky can OpenAI make its distribution here? is the winning move to merge both social and entertainment? if so, what stops Meta from copying and out-distributing? • when will OpenAI do the thing? can hardware help? • Google already has most pieces for the above - is Google's innovator's dilemma surmountable, particularly with Search, YouTube Shorts, and Gemini? given their antitrust wins, will Gemini's distribution become even more cutthroat? will consumers ever know what Nano Banana Pro is to the extent they do ChatGPT? Is Gemini O.K.? do Google's AI Overviews truly not cannibalize ad clicks? • will parts of Meta or Google ever suffer engagement losses due to the onslaught of AI slop? if so, will more users flock to a more personal, trusted, and walled garden-esque ChatGPT? • could inference improvements and model efficiency play a larger role than free-tier monetization? most consumers need little intelligence for their problems - when will most inference end up on-device, shifting compute costs to consumer devices? how much of this is solved at the software level via model routing? • should I include a question about Apple? why or why not? • Is the above discounting OpenAI's moonshot efforts and ongoing enterprise deals too much? Should we treat OpenAI's 'Meta-ification' as a make or break inflection point given the cultural changes, or instead view it as an intelligent hedge which helps bridge the next few years? • Fidji saved Instacart, guiding it from an unprofitable pandemic darling into an enduring IPO. Will Fidji save OpenAI?

thank you for subscribing to my newsletter please voice your opinion via slack reactions

avatar for near
near
Fri Nov 28 03:31:44
@sarahdingwang @martin_casado @eriktorenberg @appenz @Mascobot @BornsteinMatt @JasonSCui @MichelleLazzar @dttlngr @MaikaThoughts For the people who I couldn't tag, thanks for making us a better team and me a better person. Alicia, Nate, Andrea, @chsrbrts @KatieBaynes @RaghuRaghuram @richardzphotoz @humford @liangsays @vanlegray @shangdaxu @steph_zhang @jameskissell @rajko_rad Michele, Sharon, Lauren 🫶

@sarahdingwang @martin_casado @eriktorenberg @appenz @Mascobot @BornsteinMatt @JasonSCui @MichelleLazzar @dttlngr @MaikaThoughts For the people who I couldn't tag, thanks for making us a better team and me a better person. Alicia, Nate, Andrea, @chsrbrts @KatieBaynes @RaghuRaghuram @richardzphotoz @humford @liangsays @vanlegray @shangdaxu @steph_zhang @jameskissell @rajko_rad Michele, Sharon, Lauren 🫶

GP @a16z AI x Infra. 💙 Data, AI and dev tools. Portcos: @elevenlabsio @FAL @Ideogram_ai @mintlify @motherduck @usepylon @resend @reductoai @StainlessAPI

avatar for Jennifer Li
Jennifer Li
Fri Nov 28 03:30:57
RT @acquiredotcom: Getting acquired isn’t guesswork. It’s process and preparation. How exit-ready is your business? Take the test: https:…

RT @acquiredotcom: Getting acquired isn’t guesswork. It’s process and preparation. How exit-ready is your business? Take the test: https:…

Founder and CEO of @acquiredotcom. https://t.co/wRMIssDmhl has helped 100s of startups get acquired and facilitated $500m+ in closed deals.

avatar for Andrew Gazdecki
Andrew Gazdecki
Fri Nov 28 03:27:19
  • Previous
  • 1
  • More pages
  • 2150
  • 2151
  • 2152
  • More pages
  • 5634
  • Next