LumaFold: Screen Light
Makeup Mirror & Color Effects for iPhone Duo
Surviving "The Graduate"
What the Founders Who Dared to Start Actually Have in Common
Every founder story looks inevitable in hindsight. The garage, the first customer, the pivot that changed everything. But if you strip away the mythology, the founders who make it past the early years tend to share a handful of unglamorous habits. None of them are about being fearless. Most of them are about reducing the cost of being wrong.
This piece looks at what the data says about starting a company, and what the people who dared to start did differently.
The odds are real, but they are not the whole story.
Let's start with the uncomfortable numbers. According to the U.S. Bureau of Labor Statistics, of all private business establishments born in March 2013, 79.6% were still operating one year later, 50.6% after five years, and only 34.7% after ten years.

Read that again: roughly half of new businesses do not reach their fifth birthday. That is not a reason to stay put. It is a reason to design your first years around survival, not around a perfect launch.
The founders who last usually treat those statistics as a design constraint. They ask a simple question early: what has to be true for us to still be here in year three?
Habit 1: They start from a gap, not from an idea.
Ideas are cheap. Gaps are expensive, because someone is already feeling the pain. The founders who build durable companies tend to start from a problem they have seen up close: in a previous job, in their neighborhood, in their own daily frustrations.
A practical test: can you name ten people who have this problem today, and would three of them pay to make it go away? If not, you have a hypothesis, not a business. That is fine, as long as you treat it as one.
Habit 2: They keep the first version small and cheap.
Survivors rarely bet everything on version one. They launch something narrow, charge for it early, and use real customer behavior to decide what to build next.

This is less about frugality and more about learning speed. Every month of runway you save is a month you can spend being wrong in useful ways. A smaller first version also makes it easier to admit when something is not working, because you have not sunk two years into it.
Habit 3: They do not romanticize failure.
There is a popular belief that failure is a rite of passage and that second-time founders are automatically better. The evidence is more sobering. A study by Professor Francis Greene of the University of Edinburgh Business School, which followed nearly 8,400 German start-ups, found that entrepreneurs who had failed before did no better than first-timers, and were often more likely to fail again.
The lesson is not "never fail." It is that failure only teaches you something if you deliberately extract the lesson. Founders who improve between ventures write down what went wrong, test those assumptions in the next project, and change their behavior instead of just their branding.
Habit 4: They borrow experience through mentors and networks.
Nobody builds alone, and the numbers back this up. A SCORE survey on small businesses found that mentored businesses were 12% more likely to remain in business after one year than the national average, and that business growth rose with the number of mentoring sessions.
A good mentor will not hand you a playbook. What they offer is pattern recognition: the ability to say "I have seen this go wrong before" while there is still time to change course. Peer networks do something similar. Talking to founders who are six months ahead of you is often more useful than reading another framework.
Habit 5: They study other people's journeys, critically.
Founder stories are useful when you read them as case studies rather than as motivation. The interesting parts are rarely the big wins. They are the boring decisions: the first hire, the pricing change, the market they chose not to enter.
If you want a starting point, French-speaking readers can browse a collection of entrepreneur journeys and success stories on henrypanky.com, which focuses on the courage to start, spotting market gaps, and learning from setbacks. Whatever you read, ask the same questions each time: what constraint did they face, what did they try first, and what would I have done differently?
Habit 6: They tell a clear story early.
Marketing is not something you bolt on after the product is ready. The founders who gain traction early can explain in one sentence who they help and why it matters. That sentence shapes the product, the pricing, and the first hires.
Clarity also makes you easier to help. Mentors, investors, and early customers can only point you to the right people if they understand exactly what you are doing.
A short checklist before you leap.
- Can you describe the problem without mentioning your solution?
- Have you talked to at least ten people who have that problem?
- What is the smallest version you could charge for within 90 days?
- How long can you survive if revenue takes twice as long as planned?
- Who are two people you can call when you get stuck?
- What would make you stop, and have you written it down?
The Bottom Line
Daring to start is not about ignoring the odds. It is about respecting them enough to build a company that can survive its own mistakes. The founders who make it are rarely the boldest people in the room. They are the ones who learn fastest, ask for help early, and keep their bets small until the market tells them to go big.
So before you write the business plan, write down your first three experiments. That might be the most entrepreneurial thing you do this year.
I Gave My App Away for Free — And It Almost Killed It
Generosity is a marketing strategy. Unit economics is a survival strategy. Here is how my "Founder's Launch Program" saved my indie app.
The best week of my indie life.
It started with a stranger on Reddit.
He'd tried my native notes app, EpicNotes, on his Xiaomi phone and left a comment: "Feels fast and nice. But there's some minor visual weirdness on my phone." I replied offering him a free 3-month Pro code as a thank-you for the feedback. He said yes. I generated the code in the Play Console, felt like a real company handing out a real perk, and went to sleep happy.
Then came more. A fellow tester who cancelled his subscription on principle but still sent me bug reports with screenshots. A student from India who emailed: "I'm a student, so I can't afford it. If you could make a lifetime promo code available for me, I would be able to fully use it and provide suggestions and bug reports. Please 🙏"
I granted him lifetime Pro in about two minutes. It felt amazing. People were using my app. People were talking to me. After six months of building native Kotlin Multiplatform code at night after my day job as a product manager, this was the part that finally felt real.
I thought I was doing indie development right. I was doing indie development feelingly. Not mathematically.
The math I refused to do.
Here's the embarrassing part: I'm a product manager by day. I build business cases for a living. I would never ship a feature at work without a rough cost-benefit estimate.
And yet at 1 a.m., I was handing out free Pro access to an app with AI features without ever opening a spreadsheet.
So let me open it now, in public.
What a free Pro user costs me:
- Every AI summary and smart web clip is a call to a third-party AI API. Pennies per call — but power users summarize a lot.
- Every synced note is compute, storage, and bandwidth on my cloud backend.
- Every 90-day code, and especially every lifetime grant, is a commitment to pay both of those bills for months or years.
What a free Pro user pays me: $0.
My actual revenue at the time: three paying subscribers. Single-digit monthly recurring revenue.
The scary part wasn't that the number was big. It was that the number was small enough to ignore — and big enough to bleed. A leak doesn't need to be a flood to sink a boat that isn't moving.
Why free users felt like growth (and weren't).
The indie playbook says: give things away early, collect feedback, build goodwill. And the playbook is right — partially.
That Xiaomi tester? His free code paid for itself ten times over. He sent screenshots of two real bugs (a settings sheet clipped by system insets, a formatting popup buried under the soft keyboard), his exact device model, and unprompted pricing research: "I have a rule against using subscription-based services." That's the kind of intelligence money can't buy.
But here's what nobody tells you: most free Pro users don't do that. They redeem the code, run a few AI summaries, and go quiet. No feedback. No review. No referral. Every summary they run is a cost with no return.
The uncomfortable reframe I had to accept: a free user who costs you money every month is not an asset waiting to convert. It's a liability waiting for your motivation to run out.
The Pivot: The Founder's Launch Program
I realized I was trying to force a subscription model on a market that was screaming for a one-time purchase. Multiple users told me, in different words, the same thing: I will not pay monthly.
So instead of defending my subscription, I stopped fighting subscription fatigue and decided to price for it.
I just launched the EpicNotes Founder's Launch Program. It is a one-time Lifetime Deal for the first 100 users.
Here is why I structured it this way:
- It solves the subscription fatigue: Users pay once and own the Pro features (including the AI engine) forever. No monthly anxiety.
- It solves my unit economics: The upfront cash injection covers my API and server costs for the next year, giving me the runway to keep building without bleeding money every month.
- It builds a core community: The 100 "Founders" get direct access to me. Their feedback shapes the roadmap.
If you are tired of bloated Electron wrappers and want a truly native, privacy-first notes app for Mac and Android, this is your chance to support an indie dev and lock in a lifetime license before the doors close. (Link to the Founder's Launch Program is in my bio/below).
What I changed (the boring, unsexy fixes).
1. I hard-paywalled the AI. The free tier now includes the full notes experience — writing, formatting, sync — but zero AI. AI is the only feature with a real marginal cost, so it's the only feature behind the paywall. A free user now costs me roughly the price of storage: pennies.
2. I put my generosity on a budget. Twenty goodwill codes, total. They now go to people who give something back: detailed feedback, honest reviews, real reach. The tester who sends bug reports with screenshots? Instant yes. A drive-by "send code pls"? A polite no with a link to the free tier.
3. I asked. I reached out to my core users and told them the truth: solo dev, real server costs, if this app helps you, here's how to keep it alive. Honesty converts better than guilt — and both convert better than silence.
The Lesson
Generosity is a marketing strategy. Unit economics is a survival strategy. You need both — but in this order: survive first, then be generous on purpose.
I'm still giving EpicNotes away to students, testers, and kind strangers on Reddit. I just open the spreadsheet before I open the billing dashboard now.
If you want to try a native, local-first notes app without the subscription trap, grab a spot in the Founder's Launch Program. Let's build the best productivity tool on the Mac and Android ecosystems together.
Have you ever been burned by subscription fatigue? Tell me in the comments. And if you want to support an indie dev building native apps, check out EpicNotes below.
https://epicapps.app/epicnotes
Grab a new 3D printer this Halloween weekend, with prices as low as $199
[分享创造] 做了个给 2D 游戏用的 AI 精灵图生成器:一张角色图生成待机、攻击动画,直接导出精灵图和 JSON
大家好,距离上次发帖已经过去挺久了。原因是 7 月底,我又失业了,哈哈哈。
不过这次失业我到没有那么焦虑了,原因是通过这一年的折腾,副业还是有了点成果。
然后就是哪怕失业了,我跟朋友那边也在合作做一些游戏项目,目前有稳定的订单,虽然赚的不多,但是起码饿不死。
然后我最近几个月呢,也在打磨自己的全新的产品SpriteCanvas,一个给 2D 游戏做角色动画的工具。
起因是自己做小游戏:角色立绘用 AI 很快就能出,但一到“让它动起来”就卡住。逐帧手画不现实;用视频模型生成一段动画,又得自己抽帧、抠背景、对齐脚底、拼成图集、写每帧时长,一套下来比做游戏还累。SpriteCanvas 就是把这条链路做完,最后给你一张游戏引擎能直接用的精灵图。
效果
一张参考图,生成待机和攻击动画( 35 帧,透明背景):
![]()
剑士三连击,38 帧:
![]()
技能特效也能单独生成,每个 6 帧、透明背景:
![]()
导出时可以勾“像素风”,每帧重绘到小像素网格上,整张精灵图共用一套 32 色调色板,帧与帧之间颜色不闪:
![]()
怎么用
- 描述角色,或者上传你已有的角色图。一次出几张候选,挑一张。
- 选动作:待机、行走、奔跑、攻击、跳跃是内置的,也可以自己写动作描述。所有动作都从同一个角色生成,造型不会跑偏。
- 逐帧检查,选帧范围和帧数。
- 导出 PNG 精灵图和 JSON 图集。JSON 是 Phaser 、PixiJS 能直接读的 JSON Hash 格式,带每帧时长和脚底锚点。
做完还能在内置试玩场里用键盘跑、跳、打几下,先看看手感:
![]()
首页就有一个能玩的示例场景,不用注册:
![]()
视频演示( YouTube Shorts ): https://youtube.com/shorts/vfHBoUoWIm4
不用登录的免费工具
顺手做了几个小工具,都在浏览器里本地处理,不用注册:
还有一套 CC0 免费素材包(苔藓骑士 5 个动作、剑士,加史莱姆、蘑菇怪、哥布林、石头巨人几个怪物和一组特效),可商用,拿去测试游戏代码也行。
价格
- 邮箱验证后送 40 积分,7 天内有效;之后每天送 3 积分,最多攒到 12 ,免费积分只能用于标准画质。
- 每个任务开始前会显示要花多少积分;任务失败、没有产出时,积分会退回。
- 付费套餐 Starter 起步价 $12/月;也可以不订阅,单独买积分包。
目前的不足
- 它不是逐像素的像素画编辑器。像素风是在导出时重绘的,单个像素还得去 Aseprite 这类编辑器里修。
- 动画是排队生成的,要等一会儿,不是秒出。
- Unity 、Godot 的导入流程还没正式验证。现在确定能直接用的是 PNG 精灵图加 JSON ( Phaser 、PixiJS 已验证)。
想听听大家的意见:你做游戏时最缺哪类动作或素材?想要哪个引擎的导出?用下来哪里卡住了?评论区或者站内反馈都可以,谢谢!
[推广] [自荐] AIRAP Video:把人物照片变成单人或双人 AI 说唱视频
大家好,自荐一下最近在做的 AIRAP Video,一个用人物照片生成 AI 说唱和音乐表演短片的网站。
地址:AIRAP Video
可以用它给朋友做生日 Rap 、生成双人表演,或者让自己出现在录音棚和现场演奏的场景里。
目前支持:
- 单人和双人主题:包括橙色摄影棚、绿色录音棚、单人现场演奏等方向,页面可以先看示例。
- 上传人物照片:按所选主题添加照片,生成围绕这些人物的新场景和表演。
- 可选生成音频:可以描述主题、情绪或祝福内容,给音频提供方向。
- 生成前查看报价:按主题支持的选项选择时长、比例和清晰度,确认积分费用后再生成。
目前还是早期版本。浏览主题和示例免费,视频生成需要消耗积分,自助购买积分暂未开放。 示例展示的是创作方向,实际生成结果会随照片、主题和设置变化。
现阶段想先收集一些反馈:大家更想拿它做生日祝福、朋友之间的整活,还是自己的表演短片?也欢迎反馈想增加的主题,以及页面上不容易理解的地方。
【数智周报】张一鸣以6700亿元登顶中国首富,刷新中国首富财富纪录;阿里巴巴与Apple达成合作,将千问集成至Apple智能;Anthropic提交IPO注册文件,拟以“ANT”为代码在纳斯达克上市
(本文作者为 ITValue,钛媒体经授权发布)

观点
城堡投资Salame警示AI融资风险:浪潮之下“一些船会沉没”
城堡投资联席首席投资官Pablo Salame表示,支撑AI革命的巨大资本需求,正与态度谨慎、不愿承担如此大风险的贷款机构发生冲突。他表示,这将迫使追求数倍投资回报的选股投资者与需要确保资金能够收回的债权人之间重新寻求平衡。Salame说,“AI浪潮虽会水涨船高,但也会让一些船沉没,而且这股浪潮还将充满起伏。”
马斯克与黄仁勋观点打架,摩根大通:超大规模AI数据中心运营商状况良好是市场共识
SpaceX首席执行官马斯克最新警告,200多家新云公司将大规模倒闭或合并,有些公司能挺过去,但不会很多。
而马斯克的这番警告与英伟达首席执行官黄仁勋此前的观点几乎截然相反。黄仁勋在GTC大会上指出,市场对新云的需求远超预期。他虽然也提到了目前新云公司亟需资金支持,而资金正面临短缺的问题,但他认为这一问题一定会得到解决。
杰富瑞分析师指出,马斯克的言论引发了新一轮的恐慌交易。摩根大通则认为,人工智能债务市场对人工智能基础设施变得更加挑剔,但对于更广泛的融资危机不以为然。摩根大通市场情报团队强调,信贷短缺令人担忧,但目前的共识是超大规模数据中心运营商状况良好,大型运营商将继续占据主导地位,包括新型云服务商在内的小型企业的投资回报率也仍然为正。
亚马逊CEO:预计AWS未来两年收入同比增速都将高于40%,5-10年有望成长为万亿美元营收规模业务
亚马逊首席执行官安迪·贾西在亚马逊财报电话会上,对其AWS云业务前景绘制了极为宏大的前景。他表示:“AWS的年化收入已达1840亿美元,但发展机遇依然巨大……我们预计AWS 2027、2028每年同比增速都将高于40%,未来5-10年有望成长为万亿美元营收规模业务。”
国内资讯
中兴通讯:第三季度净利润13.07亿元,同比增长394.46%
中兴通讯公告称,前三季度实现营业收入1168.38亿元,同比增长16.23%;归属于上市公司股东的净利润40.60亿元,同比下降23.71%。第三季度营业收入388.13亿元,同比增长33.99%;归属于上市公司股东的净利润13.07亿元,同比增长394.46%。
张一鸣以6700亿元登顶中国首富,刷新中国首富财富纪录
胡润研究院发布《2026胡润百富榜》。字节跳动创始人张一鸣以6700亿元第二次成为中国首富,刷新中国首富的财富纪录;DeepSeek的梁文锋首次上榜便以2800亿元财富,进入榜单前十。
腾讯汤道生:腾讯云正加快全球数据中心建设
在腾讯全球数字生态大会上,腾讯集团高级执行副总裁、云与智慧产业事业群CEO汤道生现场表示,腾讯云正加快全球数据中心建设。目前,腾讯云已在全球部署超200万台服务器,覆盖25个地理区域和70个可用区。过去一年,腾讯云在德国法兰克福、马来西亚、墨西哥、日本大阪、沙特利雅得及泰国新增或扩容了数据中心。未来一年,腾讯云将进一步扩展亚洲、中东、欧洲、美洲的整体布局,持续为客户打造高性价比的基础设施,以满足AI时代激增的通算与存储需求。
蚂蚁集团发布灵光2.0
蚂蚁集团发布灵光2.0,可支持在开放环境中调用MCP、CLI等工具。据了解,灵光App是蚂蚁集团推出的全模态通用AI助手,于2025年11月18日正式上线,目前已内置支付宝账单、会员、出行,以及钉钉、飞书、邮箱、米家等服务能力。
宇树科技:第三季度净利润1.13亿元,同比下降17.37%
宇树科技公告称,第三季度营业收入为6.65亿元,同比增长69.93%;归属于上市公司股东的净利润为1.13亿元,同比下降17.37%。前三季度营业收入为18.17亿元,同比增长55.71%;归属于上市公司股东的净利润为3.87亿元,同比增长270.54%。
豆包工作升级电脑操作能力:支持技能录制与锁屏执行
豆包工作升级电脑操作能力,屏幕点击精度与速度提升。用户授权后,其可自动查看屏幕并模拟键鼠操作。通过“更多技能”入口可录制电脑操作,演示一遍流程,豆包即记录鼠标点击、文字输入及窗口内容,整理为个人技能,实现同类任务快速复用。
此外,用户可在设置中开启“允许豆包在锁屏时工作”。开启后,豆包可在Mac锁屏后继续执行任务,支持手机下达指令及查看结果。
豆包工作向企业客户开放配置三方模型
豆包工作向客户开放三方大模型服务,首批将接入 Kimi 3.1 Pro、GLM 5.3、DeepSeek V4.1 Flash 等系列模型。本次更新面向团队订阅版和企业订阅版客户开放。功能上线后默认直接开启,企业无需另行向模型厂商采购,也无需额外申请,后续按照用量扣除企业已有的AI点数,真正实现开箱即用。
腾讯WorkBuddy上线“远程控制”
腾讯WorkBuddy宣布正式上线“远程控制”,在手机上打开App,一键连上电脑,就能看到电脑桌面的实时画面,并直接动手接管。
华为将韬定律思想引入基站设计,发布新一代移动通信网产品及解决方案
在Decoding China 2026活动上,华为宣布首次将“韬(τ)定律”的时间缩微理念引入基站产品设计,并用“算不完、传不全”概括传统基带在单个TTI内面临的挑战。同时,华为发布OSCAR智能信道处理方案与新一代UBBPi基带系列。围绕单个TTI内的计算与传输约束,方向给出对整网协同的表述。
阿里巴巴与Apple达成合作,将千问集成至Apple智能
阿里巴巴与Apple达成合作,将千问集成至Apple智能,中国用户可以在Siri、写作工具及视觉智能中,直接调用千问的文本与图像理解、内容生成等能力。这些功能将于11月4日起随最新的iOS开发者Beta版开放测试,并将于12月iOS 27.3正式发布时面向中国用户推出。
海外消息
Anthropic提交IPO注册文件,拟以“ANT”为代码在纳斯达克上市
人工智能初创公司Anthropic向美国证券交易委员会提交S-1注册文件,计划首次公开发行A类普通股,并申请在纳斯达克全球精选市场以“ANT”为代码上市。文件暂未披露发行股数、发行价格区间及具体发行日期。
Meta预计2026年资本开支最高1450亿美元,计划通过股权融资等方式支持AI投资
Meta表示,预计2026年资本开支为1350亿至1450亿美元。公司称,AI投资已通过提升用户参与度、改善广告表现及提高变现效率带来实际回报,未来还将投资个人超级智能、消费级及企业级AI智能体、AI眼镜等业务。Meta计划通过经营现金流、投资级债务、战略合作及股权资本等多种方式为相关投资提供资金。2026年前三季度,公司经营现金流为970亿美元;截至三季度末,现金、现金等价物及有价证券约820亿美元。
甲骨文联合OpenAI推出企业AI产品套件,覆盖办公、开发与网络安全
甲骨文宣布与OpenAI扩大合作,将其模型、ChatGPT和Codex整合至甲骨文的应用、数据及云基础设施。双方推出多项企业AI产品,包括可通过ChatGPT调用企业系统的Oracle AI Workspace、利用AI模型检测和修复漏洞的Oracle AI Security Mesh,以及支持员工借助Codex开发企业应用的Oracle App Composer。
谷歌推出最新世界模型Genie 4,模拟时长翻倍至2分钟、首次实现音视频同步生成
谷歌DeepMind推出最新世界模型Genie 4,并宣称是其迄今最逼真、交互能力最强的世界模型。与主要根据提示词生成视频片段的传统视频生成模型不同,Genie 4能够根据文本指令构建可探索的虚拟环境,并在用户持续输入指令时实时调整场景。谷歌表示,其最长可以维持两分钟连贯不变的场景,时长约为上一代Genie 3的两倍。该模型还能生成和虚拟世界同步的实时音频,涵盖音乐、环境音以及可交互对话。
融资并购
Meta拟融资350亿美元并计划推出600亿美元股票发行计划
Meta宣布启动总规模约350亿美元的同步公开发行,包括225亿美元A类普通股和125亿美元强制可转换优先股存托股份,并计划于2027年第一季度推出最高600亿美元的场内股票发行计划。英伟达、AMD和博通已表示有意合计认购最高50亿美元的A类普通股,但尚未作出有约束力的承诺。
OpenAI据称将新一轮融资目标下调100亿美元,拟以1.25万亿美元估值融资200亿美元
OpenAI已将最新一轮融资目标下调100亿美元,计划融资200亿美元,对应估值1.25万亿美元。此前,该公司原计划以1.4万亿美元估值融资300亿美元。
中科慧灵完成数亿元A++轮融资
北京中科慧灵机器人技术有限公司,完成数亿元人民币A++轮融资。本轮融资由知来投资、涌铧投资、茂天投资、元禾辰坤、国科科智、鲁宁投资等机构联合投资。投资方涵盖市场化股权投资机构、专业母基金及产业投资平台,投资布局涉及人工智能、硬科技、先进制造等领域。

超球面科技完成数亿元天使轮融资
深圳超球面科技有限公司完成数亿元天使轮融资,由君联资本、某超一线美元基金、中科创星、商汤国香资本、华山资本WestSummit Capital、庚辛资本中国 等机构共同投资,庚辛资本中国担任独家兼长期战略财务顾问。公司由香港中文大学教授刘威杨创立,聚焦下一代基础模型研发。
趋境科技完成B轮超13亿元融资
AI Token生产服务商趋境科技(Approaching.AI)完成B轮融资,本轮融资金额超13亿元,今年内已累计募得24亿元。本轮融资由浙财资本、萧山资本、工银投资、中关村科学城、星连资本及钱江世纪城共同领投,云创资本、湖南高新纵横、建投华科、申万创新投、中信聚信、珠海科技产业集团、商汤科技、网宿科技、甘肃树帜、常金控、元智未来、清证星竹、小咖资本等跟投,老股东尚势资本继续追投。
文远知行子公司景烁科技首轮融资数千万美元,布局物理AI
文远知行旗下子公司景烁科技获得数千万美元Pre-A轮融资,投资方包括嘉御资本、HKSTP香港科技园、天狼星资本等。景烁科技主要布局物理AI数据与模型基础设施,目前已推出数据-模型一体化平台WorldEngine及世界模型GENESIS-Robotics。据公司披露,目前其在手订单金额达数十亿元,年收入增长超100%。
趋势
宇树科技:人形机器人行业尚处商业化早期,物理AI技术从实验室演示走向真实场景规模化落地仍面临多重挑战
宇树科技公告称,人形机器人行业尚处于商业化早期,技术成熟度、下游市场需求培育、产业链配套能力均存在不确定性。同时,物理AI技术从实验室演示走向真实场景的规模化落地仍面临感知精度、运动控制稳定性、数据积累及软硬件协同等多重挑战,相关技术路径与商业模式的成熟尚需时间验证。若行业大规模商业化落地进度不及预期,公司未来营收扩张及业绩释放可能受到不利影响,存在估值回归的投资风险。
中信证券:GPT-6 Astra进入机器人控制,物理AI迎来重大拐点
中信证券研报称,通用大模型正在从机器人的高层任务规划,进一步向动作生成和执行控制延伸。近期,NVIDIA、Stanford等研究机构陆续探索将GPT-6 Astra应用于机器人控制,尝试利用通用模型已有的视觉理解、空间推理和任务规划能力,减少对专用策略训练的依赖。Astra在具身任务的数据集上展现出了较好的泛化性,中信证券认为,展望2027年,通用模型进入具身任务领域将成为AI行业的下一个投入方向,而通用模型与专用模型短期更可能形成分工协同的局面。
人形机器人应用持续拓展,核心零部件投资机遇受关注
商务部发布的数据显示,今年前三季度,重点平台具身智能机器人销售额同比增长22.6%。从消费端需求增长,到具身智能整机及配套项目推进,机器人产业的应用与商业化进展受到关注。分析人士认为,人形机器人产业正处于量产验证与场景落地的重要阶段,投资机会有望向核心零部件、控制平台和数据基础设施等环节延伸。随着产业发展更加重视实际应用,应关注企业技术优势、成本控制和批量交付能力,以及订单落地与模型迭代进展。
SK海力士预计存储市场供不应求局面将持续至2030年
SK海力士表示,受先进制程技术难度提升及新建生产设施耗时较长等因素影响,存储市场供不应求格局预计将持续至2030年。公司计划加快扩产,预计第四季度HBM4在出货量和营收方面均将超过HBM3E;同时计划将321层NAND工艺占韩国本土产能的比重提升至50%以上,以增加高性能、大容量企业级固态硬盘供应。
亚洲芯片股下跌,围绕大型科技公司的AI融资担忧加剧
决定AI交易前景的关键财报季表现平平,导致全球芯片制造商股价下跌,投资者越来越担心成本飙升和需求的可持续性。最新触发因素包括亚马逊的大规模资本支出数字、英特尔因个人电脑处理器需求疲软发布低于预期的销售预测。这进一步加剧了投资者对AI热潮融资成本高昂以及这些巨额投资能否获得回报的担忧。亚洲芯片制造商普跌,其中,韩国SK海力士和日本爱德万测试跌幅最大,均超过6%。韩国Kospi指数一度下跌4.3%。
“本财报季对人工智能企业来说很艰难,大多数企业都受到了人工智能泡沫和通胀担忧的双重打击,”Floodlines Research创始人、投资策略师Anna Wu表示。“更高的融资成本和信用风险最终可能会迫使超大规模数据中心运营商放缓扩张步伐。”
更多精彩内容,关注钛媒体微信号(ID:taimeiti),或者下载钛媒体App
Tackling Multimodal Learning Challenges with Mixture-of-Expert: A Survey
What's Your Dream Job? Mine Is Founder & CEO of My Own Startup 🚀 (Yours Can Be Anything)
Description: "I'm telling you my dream job, and I want to hear yours. Developer, astronaut, CEO, game designer, chef, anything goes."
🎤 The question: If money, fear, and "be realistic" didn't exist, what job would you wake up excited to do every single day?
I'll go first. Then it's your turn. 👇
🎯 My Dream Job
const me = {
dreamJob: "Founder & CEO",
company: "My own startup",
status: "Building toward it. Every. Single. Day.",
fuel: ["dedication fire 🔥", "curiosity 🔍", "stubbornness 💪"],
};
console.log(`${me.dreamJob} of ${me.company}`);
// → "Founder & CEO of My own startup"
Yes, I want to be the person who starts with a blank editor and a wild idea, and ends up with a company that real people use, love, and work at. 🌱➡️🌳
🤔 Why Founder & CEO?
1. I want to build things that matter.
As developers we already turn ideas into reality with a keyboard. A founder takes it one step further: you pick which problem deserves to be solved.
2. The learning curve is a vertical wall, and I love it. 🧗
In one day a founder might be:
| 🕘 Time | 🎭 Role |
|---|---|
| 9:00 AM | Product Manager |
| 11:00 AM | Developer (fixing a production bug) |
| 1:00 PM | Salesperson |
| 3:00 PM | Designer (yes, you picked that button color) |
| 5:00 PM | Recruiter |
| 8:00 PM | Therapist (for yourself) |
| 11:00 PM | Developer again 😅 |
Where else do you get that many careers in one job?
3. Ownership. 🔑
Every decision, win, and failure is yours. It's terrifying, and it's also the most motivating thing I can imagine.
4. Creating jobs and opportunities. 🌍
Building a team that believes in the same mission and grows together is the part that excites me most.
⚖️ The Honest Part (Because Dreams Deserve Reality Checks)
| ✨ The Instagram version | 🔥 The real version |
|---|---|
| "Being my own boss!" | Your boss is now every customer, investor, and deadline |
| "Big vision, big impact" | Big spreadsheets, small runway |
| "Pitching on stage" | Pitching to 40 people who say "interesting, let's keep in touch" |
| "Launch day 🎉" | Launch day → 3 signups (2 are your mom) |
| "Freedom" | Freedom to work 80 hours on whatever you want |
I know all this and I still want it. Maybe that's how you know it's a dream and not just a job. 😄
🎲 Plot Twist: It Doesn't Have to Be My Startup
The CEO dream has a second version: becoming CEO of someone else's startup. Joining as early talent, growing with the company, earning the trust of the team, and eventually leading it.
Both paths are valid. Both are hard. Both are awesome.
🗳️ Quick Poll (Comment Edition)
Which dream-job category is closest to yours? Reply with the number:
1️⃣ 🚀 Founder / CEO
2️⃣ 💻 Dream dev role (open source, big tech, indie hacker)
3️⃣ 🎨 Creative (game dev, design, music, content)
4️⃣ 🌌 Totally outside tech (chef, pilot, astronaut, etc.)
I'll tally the results in the comments later. 📊
💬 Your Turn! Drop Your Dream Job Below
Any job is welcome. Developer, CEO, CTO, game designer, teacher, footballer, astronaut, farmer, YouTuber, anything. There are no wrong answers here, only inspiring ones. ✨
Copy this template and fill it in:
🎯 Dream job:
🧠 Why this one:
🛠️ What I'm doing right now to get there:
😅 The scariest part:
💡 One piece of advice I'd give someone with the same dream:
Not sure what to answer? Here are some conversation starters:
- 🧒 What did 5-year-old you want to be, and how far off is it from today?
- 🔄 Have you changed your dream job over the years? What changed it?
- 🏆 If you could be great at anything instantly, what would you pick?
- 🧭 Is your dream job something you want or something you were told to want?
🙌 Let's Make This Thread Awesome
I'll reply to every comment, because someone's dream deserves a response. 💙
Maybe someone here will find a co-founder, a mentor, or just the push they needed to start.
So tell me: what's your dream job, and what's one step you can take this week toward it? 👇
(And if your answer is "Founder & CEO" too, let's connect. The world needs more builders. 🚀)
P.S. If this post made you smile, think, or finally admit your dream out loud, drop a ❤️ or 🦄 and share it with someone who needs to hear that dreams are worth chasing.
Here's what I've built so far under the title of Founder & CEO @(Akhouri Systems)
👉https://akhouri-anmol-kumar.github.io/Akhouri-systems/
Can AI Make a Whole Movie From One Prompt? Here's Where It Breaks
The question I get most often about AI film tools isn't about price or models. It's this one: can I type my idea and get the movie?
No, not in 2026. Every AI video model still generates one short clip per request, and no request remembers the one before it, so a film made of hundreds of shots has to be planned shot by shot before anything is generated. One-click tools can hide that planning from you. They can't remove it.
I make Lost Garden, a dark fantasy animated series, mostly alone. Episode 2 runs 21 minutes. It took me 75 hours and about $1,850 in generations. None of those hours were spent waiting for one magic prompt to finish.
Here's where the one-prompt idea breaks, and what to do instead.
What does a "one-click AI movie generator" actually promise?
The pitch is simple and very attractive. Invideo's movie maker page says the tool "turns your movie script into a full-length movie with AI" and makes movie-making effortless "with just a single prompt." Plenty of other sites promise the same thing in different words.

Invideo's AI movie maker page, captured on October 19, 2026 (invideo.io).
What you usually get is a sequence of short clips stitched together, with a voiceover, music and subtitles on top. For a trailer, an ad or a social video, that can be fine.
For a story, the cracks show fast:
- The main character's face drifts from one scene to the next.
- The location changes between two shots that are supposed to happen in the same room.
- The camera does whatever it wants, so the edit has no rhythm.
- You can't fix one shot without regenerating a whole block.
The tool made every decision for you, and you can't take any of them back one shot at a time.
Why can't AI video models generate a full movie in one go?
Because a single generation is short. Google's own Gemini API documentation for Veo 3.1 lists clip durations of 4, 6 or 8 seconds. You can extend a Veo video by 7 seconds at a time, up to 20 times, for a maximum of 148 seconds. That's two and a half minutes, as one continuous take.

Veo 3.1 parameters in Google's Gemini API documentation, captured on October 19, 2026 (ai.google.dev).
Other models go longer per clip. Wan 3.0, which we added to ScreenWeaver on October 18, 2026, generates between 2 and 30 seconds. Still nowhere near a film.

Single clips and chained extensions next to the runtime of one 21-minute episode.
Now do the math on a short episode. Lost Garden episode 2 is 21 minutes, so 1,260 seconds of screen time. At 8 seconds per clip, that's at least 158 generations just to cover the runtime once, before a single retake. And nobody keeps every take.

The minimum number of generations for one episode, before any retake.
A movie is hundreds of short decisions that have to agree with each other.
So the useful question is who makes those 158 decisions, and whether they agree.
Why do AI shots forget each other?
Each generation request starts from zero. The model doesn't know what it generated for you five minutes ago. It knows what's in the request: the prompt, and any images or video you attach.
That's why consistency in AI video always comes from references you pass in, never from the model's memory. If shot 12 and shot 13 show the same knight in the same corridor, it's because you gave both requests the same character reference, the same location reference and a prompt that describes the same light.
A one-click tool has to invent those references for you, and it invents them from a two-line idea. It doesn't know that your knight has a lantern for a head, that the corridor is lit from the left, or that this scene is supposed to feel cold. It guesses. Then it guesses again on the next shot, slightly differently.
Continuity is information, and a single prompt doesn't carry enough of it. A screenplay does. A shot list does. A storyboard does.
What does a working AI film workflow look like in 2026?
The workflow that holds up is the old one, with generation moved to the end. Here's the order I use on Lost Garden:
- Write the script. Scenes, characters, places, what happens and why. This is the part I never hand to a model.
- Break it into shots. For each shot: framing, angle, camera movement, duration. Short clips stop being a limit when every shot was meant to be short.
- Storyboard it. One image per shot. An image costs far less than a video clip, so this is where you find the bad ideas cheaply.
- Lock your references. Character sheets, locations, props, attached to every shot that needs them.
- Generate the footage last, shot by shot, from the storyboard frame and the references.
- Edit. Cut the takes that don't work, regenerate only those shots.

Steps 01 to 04 are decisions. Generation comes last.
Steps 1 to 4 are where the film gets made. Step 5 is execution.
This is also why I'm skeptical of any tool that skips straight to step 5. You save the first hour and lose the next fifty, regenerating shots that were never planned.
Where does ScreenWeaver fit in that workflow?
I'm a co-founder of ScreenWeaver, so take this paragraph with that in mind. We built it because I was doing those six steps across too many apps, copying prompts from a script file into a generator and losing track of which take belonged to which scene.
ScreenWeaver is an AI production pipeline for screenwriters, filmmakers and studios: you write the script, break it into shots, build the storyboard, then generate the footage, all connected to the screenplay. A project has three tabs, Script, Storyboard and Production. Each shot carries what it needs from one tab to the next:
- its scene in the script, so you always know which line it serves;
- its framing, angle, camera movement and duration;
- its character, set and reference images, the same ones from storyboard to video.
It doesn't make a film in one click. You write, you cut, you validate your shots, then you generate.
The questions people ask us every week
Since the video features went live, the same four questions come back in comments and emails. Here are straight answers.
Is ScreenWeaver still in beta?
No. The beta ended on October 8, 2026, the day video generation opened to subscribers. The full pipeline (script, storyboard, production) is live for paying plans.
Is it free?
The writing side is. The Screenwriter pla
Proof of Usefulness Hackathon Final Winners Announced 🎉
Once you discover truly useful software, it’s hard to ignore. It finds its way into your life and workflow, helping you answer a nagging question, finish tasks more efficiently, make faster decisions, or solve a problem that existing tools made more complicated than it needed to be. Before long, you might even start to wonder how you ever got by without it.
When we launched the Proof of Usefulness Hackathon in January, we wanted to give the people building these kinds of solutions a platform, along with a share of $150,000 in cash prizes and software credits to help them keep building. Alongside Bright Data, Neo4j, Algolia, and Storyblok, we invited builders to show us what their projects could do and provide the evidence to back it up.
And boy, did they show up.
Over the course of the hackathon, we received hundreds of submissions and published just shy of 300 stories on the #proof-of-usefulness-hackathon page. Behind those numbers are builders who took the time to share their work, explain the problems they were tackling, and put their progress out there for others to examine.
To everyone who submitted a project, published a report, or spent time discovering what others had built: thank you for making this hackathon's inaugural edition a success.

In May, we announced our first set of winners, recognizing projects across all five award categories. Today, we’re closing out the hackathon with the final round of winners.
This round covers the remaining $118,000 in prizes, comprising $12,000 in cash and $106,000 in software credits, distributed across the following categories:
|
AWARD CATEGORY |
FINAL-ROUND PRIZE POOL |
|---|---|
|
Bright Data |
$60,000 in software credits + $6,000 in cash |
|
Algolia |
$40,000 in software credits. |
|
Neo4j |
$6,000 in software credits. |
|
Storyblok |
$3,000 in cash. |
|
HackerNoon |
$3,000 in cash. |
Now, without further ado, let’s meet the projects taking home the final Proof of Usefulness Hackathon awards:
Bright Data Awards
$60,000 in Software Credits ($20,000 each for 3 projects)
Read Visby’s full Proof of Usefulness Report here.
Read FoxPlug’s full Proof of Usefulness report here.
Read Siiqo’s full Proof of Usefulness report here.
$6,000 in Cash Prizes ($1,000 each for 6 projects)
Read Autonomous Deep Research Agent’s full Proof of Usefulness report here.
Read Cofoundr’s full Proof of Usefulness report here.
Read the full Proof of Usefulness report for Ella’s AI Agent here.
Read TransactionLink’s full Proof of Usefulness report here.
Google is giving Gemini’s overlay a literal sparkle on your Android phone
- Google is rolling out a new Gemini overlay animation to all Android phones.
- The animation uses Google’s Neural Expressive design language and gives Gemini an ethereal look.
- The redesign is rolling out as part of a Google app update.
Google loves redesigning things, and Gemini is one of the things it loves redesigning the most. We’ve already seen the company experiment with new animations for Circle to Search and Gemini Live. We also spotted a new Gemini animation that later arrived on the Pixel 11 series. Well, if you were hoping to get that new Gemini design without buying a new Pixel, you’re in luck.
A new Google app update is now finally bringing the updated Gemini design to all Android devices (spotted via 9to5Google). The redesign brings Google’s Neural Expressive design language to the Gemini overlay, complete with the sparkling effects and tiny dots surrounding the “Ask Gemini” text box.
Big blow for Qualcomm? This upcoming Ultra flagship could skip Snapdragon flagship SoCs
- Chinese smartphone brand OPPO is reportedly going with the MediaTek Dimensity 9600 Pro chip for the Find X10 Ultra flagship.
- The Find X9 Ultra launched earlier this year with the Snapdragon 8 Elite Gen 5 SoC, making this a big shift for OPPO.
- OPPO’s Find X10 Ultra is expected to include several key camera upgrades, including a 200MP main camera with LOFIC tech.
Qualcomm once dominated the premium smartphone segment with its Snapdragon chipsets, but that grip is gradually slipping. Multiple companies are pivoting to other sources for their high-end devices, with fellow chip producer MediaTek being a key beneficiary. Now, another prominent smartphone maker, OPPO, is rumored to be making a similar switch for its upcoming Find X10 Ultra flagship.
As per leaker Smart Pikachu on Weibo, OPPO could equip the Find X10 Ultra with the MediaTek Dimensity 9600 Pro SoC. Another prominent leaker, Abhishek Yadav, shared similar info on X and also shed light on some other details about the X10 Ultra, supposedly codenamed “xinyun.”
Grab a Nintendo Switch 2 with a free game for as low as $459 in this limited-time deal
Cisco gives 27-year-old engineer unlimited AI tokens to rewrite 98% of the company's legacy code in 24 months
Rocket Report: Winter is coming for Themis; Cowboy Space adds iron to the fire
Welcome to Edition 9.17 of the Rocket Report! It's been a while since an American company announced plans for a medium-lift new launch vehicle, but that's what happened this week when a startup named Cowboy Space revealed specs for its Big Iron rocket, named for the classic country-western song by Marty Robbins. I'm a huge fan of the name. Maybe Cowboy Space is striking while the iron is hot as the space industry frets over demand for launch outpacing supply. However, Cowboy Space lags far behind so many other launch providers nearing the launch pad with their medium-class rockets, and the company didn't have much hardware to show at this week's big reveal. In other words, don't mistake style for substance. But we're rocket fans at Ars, and we wish them luck.
As always, we welcome reader submissions, and if you don't want to miss an issue, please subscribe using the box below (the form will not appear on AMP-enabled versions of the site). Each report will include information on small-, medium-, and heavy-lift rockets as well as a quick look ahead at the next three launches on the calendar.
Virgin's new rocket plane reaches home port. The first of Virgin Galactic's new Delta-class rocket planes left its Arizona factory Wednesday for a trip to Spaceport America in neighboring New Mexico, where it will undergo preparations for a test campaign culminating in suborbital flights to the edge of space, Ars reports. Virgin Galactic expects the craft's first commercial spaceflight to take place "early" in the second quarter of 2027, but like every other projected date in aerospace, this should be taken with a grain of salt. Virgin Galactic halted commercial flights of its previous vehicle, VSS Unity, in 2024.
[推广] 分享一个图片和视频生成工具 Flux 3,做产品图、社交图和视频的朋友们可以试试
给有产品配图、社交媒体素材或短视频制作需求的 V 友们分享一下 Flux 3 AI ,可以在同一个网站里生成图片、修改照片,也可以把静态图片做成视频。
网站地址: https://flux3-ai.com/
目前新用户注册送积分,可以免费生成 1 个视频或 6 张图片,大家可以先体验一下效果。
主要支持这几种用法:
- 文字生成图片:描述主体、场景和风格,生成配图。
- 图片编辑:上传已有图片,用文字描述想修改的地方。
- 图片生成视频:上传静态图片,描述希望出现的动作。
- 文字生成视频:直接用文字描述想要的画面。
也可以先从一张社交媒体配图开始,输入你想要的画面,比如: 一只橘猫趴在咖啡馆窗边,午后的阳光照在木桌上,旁边放着一杯拿铁,整体温暖自然,方形构图。
注册后就能用赠送的积分体验,做图或做视频都可以~
‘Star Trek’ and ‘X-Men’ icon Patrick Stewart dies at 86: ‘We lost one of the greats of stage and screen today’
Patrick Stewart, the classically trained actor who added Shakespearean gravitas to the “Star Trek” and “X-Men” franchises, has died. He was 86.
Stewart died Wednesday in Los Angeles, following “a brief and poignant journey” with Alzheimer’s disease and Lewy Body Dementia, spokesperson Kelly Bush Novak said. Stewart’s diagnoses had not been previously disclosed.
“We lost one of the greats of stage and screen today,” Hugh Jackman, a frequent Stewart co-star in the “X-Men” universe, said on social media. “And, in my book, a tremendous friend and man. I learned so much from him. I will miss you, Patrick Stewart, all of my days.”
While best known as Professor Charles Xavier and Captain Jean-Luc Picard from the “X-Men” and “Star Trek: The Next Generation” series and films, respectively, Stewart spent more of his career onstage than in the X-Mansion or on the USS Enterprise.
Besides playing everyone from Claudius to Macbeth, he often performed a one-man rendition of “A Christmas Carol” and did West End and Broadway runs of Harold Pinter’s “No Man’s Land” and Samuel Beckett’s “Waiting for Godot” in repertory with buddy Ian McKellen.
He won three Olivier Awards, including one for playing Claudius in “Hamlet” opposite David Tennant. And he had a Grammy — in 1996 for best spoken word album for his recording of “Peter and the Wolf.”
He achieved some of the best reviews of his career when he finally tackled Macbeth for the first time. The production, under the direction of Rupert Goold, debuted in May 2007 at England’s Chichester Festival Theatre, switched to London’s West End and then moved to the Brooklyn Academy of Music before transferring to Broadway in 2008.
Stewart’s performance earned him his sole Tony Award nomination. The Times of London said his Macbeth was “an enthralling creation of frailty, appetite, egotism and grim humor.” The Daily Telegraph said he had turned in “a truly great performance” and The New York Times said it was “a must-see.”
“Nearly 50 years I’ve waited,” Stewart told The Associated Press in 2008. “I learned it when I was 14. I memorized all the soliloquies just because I thought they were great. And they stuck, so I didn’t have to learn those bits of the play, which is a blessing, because learning lines has now become the one curse of this job.”
Starship commander grew up without a toilet
Raised in Mirfield in Yorkshire, a working-class corner of northern England, the future 24th-century commander of a starship grew up without a bathroom in his house.
Stewart joined an amateur drama group at 12 and memorized Shakespeare for fun as a teen. He won a scholarship to the Bristol Old Vic Theatre School, launching a career as a character actor.
He had been bald since he was 19 — alopecia runs in his family.
His father, a former regimental sergeant major in the British Army, was prone to outbursts of “repeated violence against my mother,” as Stewart wrote in a newspaper piece. “Violence is always a choice a man makes and he alone is responsible for it.”
In his memoir, “Making It So,” Stewart wrote about how he and his older brother Trevor braced for the nights when their dad came home drunk and angry, and attacked their mother. “Sometimes it was with an open hand, other times a closed fist. Always he aimed for her head.”
The son would later become a patron for Refuge, an organization that provides shelter for domestic abuse victims.
Loving Shakespeare proved fruitful as Capt. Picard
“The stage would prove to be a safe space, a refuge from real life in which I could inhabit another person, living in another place and time,” he would write.
Stewart was accepted into the Royal Shakespeare Company and stayed for 14 years, playing everything from Mark Antony to Henry IV to Shylock to Oberon. “People who were not in the company would say to me, ‘Give it a break. Why don’t you go somewhere else?’ And I would say, ‘To do what?‘” he told the AP. “Telly?”
Stewart eventually left to look for more varied work, including the 1981 film “Excalibur.” He had just done a production of “Who’s Afraid of Virginia Woolf?” in 1987 when an interesting offer came along. “I found myself in Hollywood shooting a syndicated science fiction series, which I was assured would fail.”
It didn’t, of course. Thanks to Stewart’s rich baritone and noble bearing, the “Star Trek” franchise ran for seven years, 178 episodes and produced four feature films. He was known for his slogan “Make it so.”
His adoration for Shakespeare would serve him well when he began to play Jean-Luc Picard, a 24th century starship commander.
“There is a formality to the way they speak and comport themselves that remind me of numerous Shakespearean situations I’d been in onstage. I should play Jean Luc, I realized, as if he were a character in ‘Henry IV,’ which is about brave men,” Stewart wrote in his memoir.
Despite the Borg and Klingons, Stewart still managed to make room for blank verse. He established the Paramount Shakespeare Company, a workshop on Saturday and Sunday mornings for any actor interested. (Stewart revisited Picard in the TV series “Star Trek: Picard.”)
Stewart died just a few days after Walter Koenig, another “Star Trek” icon, who played Ensign Pavel Chekov on the original NBC TV show from 1966 to 1969.
“What a heartbreaking week it has been,” William Shatner, who played Captain James T. Kirk in the franchise, wrote on Instagram. “Sir Patrick Stewart was a friend, a colleague, and a magnificent actor. From the stage to the bridge of the Enterprise, he brought intelligence, grace, and a deep sense of humanity to everything he did.”
“Star Trek: The Next Generation” gave him higher visibility, which led to more work, including roles in the films “Conspiracy Theory,” “L.A. Story” and on Broadway, where he starred in “The Tempest,” “The Caretaker” and David Mamet’s “A Life in the Theater.” He was knighted by Queen Elizabeth II in June 2010. Stewart was also an ambassador for the International Rescue Committee.
A brave and commanding persona. But sexy?
In July 1992, he and supermodel Cindy Crawford were on the cover of TV Guide’s “Top Turn-Ons,” based on a readers’ poll. In his memoir, Stewart wrote: “Absurd.”
Stewart was first introduced as the character Xavier in the original X-Men trilogy, appearing in three films from 2000 to 2006. He later returned in four more movies before Professor X’s death in 2017’s “Logan.” He reappeared in “Doctor Strange in the Multiverse of Madness” in 2022 and reprised the role in the
Unlimit founder backs Phase blockchain with $45M as mainnet launches

Phase’s new network lets smartphone owners earn tokens for validating transactions and sharing internet connections with AI agents.