马克·安德森@pmarca
· 2026-06-18 05:25
马克·安德森
@pmarca · Marc Andreessen · 0 followers
第一手数据 > 即时判断。
8今日信号
今日没有多源共振的事件
📘 创业哲学
3 条最高 4.4
马克·安德森@pmarca
· 2026-06-18 05:43
有意思。Alex Petkas 的观点：西方文化是由自学者创立的吗？「autodidact」（自学者）一词最早出现在荷马《奥德赛》第22章——奥德修斯杀完求婚者后，吟游诗人斐米乌斯（「被迫为他们歌唱」）从藏身处现身请求赦命……
pmarca 长期推崇「无需传统机构认可的自学者」。
观点在 AI 工具极度民主化的当下，他在隐性预测下一批伟大建设者将来自传统教育体系之外。
对你意味着你的核心用户和潜在合伙人的筛选维度，自驱力和学习速度可能比简历背景更具预测效力。
Interesting.Alex Petkas: Did autodidacts found Western Culture?The first occurrence of the word "autodidact" (αὐτοδίδακτος) is in Homer's Odyssey.It's book 22, Odysseus has just slaughtered the suitors. The bard Phemius "who sang for them under compulsion," emerges from hiding to plead for his
马克·安德森@pmarca
· 2026-05-03 03:38
哎。Ven：「这是为你准备的」——一位将某公司推入可能破产保护（Chapter 11）的法官
对你意味着
事实某公司可能因法院判决面临 Chapter 11 破产重组，pmarca 表示关注，但原文上下文截断，具体案情不明。
观点pmarca 持续关注此类监管/司法介入商业的案例，折射硅谷对政府干预营商环境的普遍警惕。这提醒 CEO：即便合法运营，司法风险也是实际存在的，特别是在 AI 监管框架尚不明朗的阶段，法律合规的投入不应被视为可选项。
Oof.Ven: "This one is for you" — a judge who pushed a company into likely ch11
📘 AI思考方法论
2 条最高 7.4
马克·安德森@pmarca
· 2026-04-28 09:04
默认假设是生产力提升会带来需求增长。举证责任在那些末日论者身上，他们需要证明这次为何会不同。
对你意味着如果你还在犹豫 AI 是否会损害公司的人才策略，这个框架值得内化——历史上每次生产力革命都扩大了总需求。
观点反方需举证，而非默认方。真正的风险不是 AI 消灭岗位，而是你的竞争对手比你更快完成人机协同的组织重组。
"default presumption is that productivity gains lead to increased demand. burden of persuasion is with the doomers to show why this time is gonna be d...
马克·安德森@pmarca
· 2026-04-28 08:55
那不是经济即将崩溃的信号，那是一个生产力高到连时间本身都成为瓶颈的工具的信号。
对你意味着如果你的 AI 工具已经让团队感到时间不够用，这不是组织失控，而是工具效能的验证信号。
观点真正的管理挑战从此变成如何设计让时间分配最优化的组织结构，而非担心 AI 带来的失业问题。这个框架可直接用于向董事会解释为何要加速 AI 投入。
"thats not a sign of an economy about to collapse, thats a sign of a tool so productive that time itself becomes the bottleneck."
📘 模型公司战略
1 条最高 7.8
马克·安德森@pmarca
· 2026-06-18 09:58
令人担忧。Teortaxes（DeepSeek 推特铁粉 2023 至今）：我认为 GLM 5.2 表明目前存在约 7 个月的差距。综合来看，它大约处于 Opus 4.7-4.8 的水平（视觉能力方面 Opus 本身也相对有限）。Mythos 于 2026 年 2 月初达到 Preview 状态（功能上 >= Opus 4.8）。这意味着中国完整版 Mythos 将在 2026 年 11-12 月落地。
中国顶级模型与 Anthropic Opus 系列的差距正从代际差压缩到季度差，Fable 若 2026 年底落地，意味着中国 AI 能力约一年内与西方主流齐平。
事实
对你意味着以中国或亚洲为主要市场的 AI 产品，竞争窗口期比想象的更短；单纯依赖基座模型差距已不足以构成护城河，数据飞轮、垂直场景和生态锁定的优先级需要重新排序。
观点
Concerning.Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞): I think GLM 5.2 points to a 7 months gap currentlyIt's around Opus 4.7-4.8 level, all told (modulo vision which in Opus's case is garbage anyway). Mythos reached Preview status (≥ Opus 4.8, functionally) by early Feb 2026. This means full PRC Mythos ("Fable") by Nov-Dec'26.
📘 AI 伦理与安全
1 条最高 4.6
马克·安德森@pmarca
· 2026-05-03 06:00
历史上一直如此。Séb Krier：技术焦虑是一个高度一致且反复出现的模式：1482 年 Johannes Trithemius 的书《赞美抄写员》；1858 年《纽约时报》社论论述电报的认识论风险；1986 年《华盛顿邮报》……（内容截断）
每一代人都认为自己面对的技术焦虑是史无前例的，但历史记录显示这是固定剧本。
事实1482 年有人写书反对印刷机可能取代抄写员；1858 年《纽约时报》担忧电报危害认知。
观点对于 CEO，这意味着当团队或市场出现 AI 焦虑时，这不是阻力而是可以预见和管理的摩擦——历史上每次技术革命都经历这个过程，这一次也不例外，焦虑本身不会改变技术浪潮的方向。
Always has been.Séb Krier: Techno-anxiety is a remarkably consistent and repetitive pattern: 1482 book by Johannes Trithemius: "In Praise of Scribes: De Laude Scriptorum": https://williamwolff.org/wp-content/uploads/2009/06/TrithemiusScribes.pdf1858 New York Times editorial on the epistemic risks of the telegraph: https://www.nytimes.com/1858/10/25/archives/latest-by-telegraph-from-washington-the-president-and-the-pacific.html1986 WaPo
📘 AI 人才与岗位
1 条最高 8.4
马克·安德森@pmarca
· 2026-05-02 15:45
Marc Andreessen 转评 Konstantine Buhler 引用的 Citadel Securities 研究（作者 Frank Flight）：'我们在二月就已论证，软件工程师作为 AI 敞口最高的职业，其需求正在加速上升——这与取代叙事相悖。' Andreessen 仅回复 'Naturally.'（自然如此）以示背书。
对你意味着
事实软件工程师需求在 AI 最高曝光期反而在加速——说明 AI 正在放大优秀工程师的产出，而非取代他们的位置。
观点pmarca 的一字背书意味着 a16z 系公司的招聘逻辑也支持此判断：现在正是扩充核心工程人才的时机，而非等待 AI 降低用人成本的窗口。如果你在等 AI 降薪后再招人，竞争对手可能已经把稀缺的 AI 原生工程师预定完毕。
Naturally.Konstantine Buhler: Narrative violation and great insight from the latest Citadel Securities banger by Frank Flight: "We illustrated back in February that demand for software engineers, the most AI exposed occupation was accelerating higher, which we argued violates the displacement narrative.
尚未被主编选入事件/主题的原始推文（共 50 条，按时间倒序）
马克·安德森 @pmarca06-18 11:55
If a Current Thing is a Thing enough times, it becomes The Thing. The Thing can run for a decade plus, until it becomes a parody of itself. Then The T...
🔗 原推马克·安德森 @pmarca06-18 09:11
Concerning.Disgraced Propagandist: After yesterday's major Variety piece about the death of Hollywood, an anonymous Hollywood insider writes what Gene Maddaus and everyone else is afraid to."Maddaus, being a mainstream journalist, rattles off all the acceptable explanations. California’s failure to compete with
🔗 原推马克·安德森 @pmarca06-18 09:09
Self recommending.Clifford Asness: Aaron Brown’s phenomenal piece of the affordability crisis (I was a glorified repeated editor he kindly added as a co-author, Mike was in-between).Hint — most of what you read, and certainly what you hear from politicians and activists, is not, you know, right.
🔗 原推马克·安德森 @pmarca06-18 06:57
RT Disgraced Propagandist: After yesterday's major Variety piece about the death of Hollywood, an anonymous Hollywood insider writes what Gene Maddaus...
🔗 原推马克·安德森 @pmarca06-18 06:44
0xDipper: Nassim Taleb on the AI selloff: the pioneers usually aren't the winners. They're more likely to be the losers.He's not betting against AI - someone will make a fortune on the software and hardware. His point is it doesn't have to be today's hot names. History rhymes: the early
🔗 原推马克·安德森 @pmarca06-18 06:25
Concerning.The Vigilant Fox 🦊: Joe Rogan just revealed that presidents and former presidents personally pressured Spotify to have him removed for "vaccine misinformation." And not one of them apologized when his questions turned out to be right.He says he still can't talk about most of it. What he will say
🔗 原推马克·安德森 @pmarca06-18 06:15
Interesting.Nick Gillespie: 'Why is a flat-screen TV affordable and a college education not?...Because Congress has spent 60 years trying to make college affordable and has spent zero years trying to make TVs affordable.' By @reason's Aaron Brown, Michael Mendelson, @CliffordAsness https://thedispatch.com/article/affordability-crisis-healthcare-housing-childcare/
🔗 原推马克·安德森 @pmarca06-18 06:07
RT Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞): I think GLM 5.2 points to a 7 months gap currently It's around Opus 4.7-4.8 level, all told (modu...
🔗 原推马克·安德森 @pmarca06-18 01:39
RT The Vigilant Fox 🦊: Joe Rogan just revealed that presidents and former presidents personally pressured Spotify to have him removed for "vaccine ...
🔗 原推马克·安德森 @pmarca06-17 23:55
RT Nick Gillespie: 'Why is a flat-screen TV affordable and a college education not?...Because Congress has spent 60 years trying to make college affor...
🔗 原推马克·安德森 @pmarca06-17 23:20
RT Alex Petkas: Did autodidacts found Western Culture? The first occurrence of the word "autodidact" (αὐτοδίδακτος) is in Homer's Odyssey....
🔗 原推马克·安德森 @pmarca06-17 21:29
RT Clifford Asness: Aaron Brown’s phenomenal piece of the affordability crisis (I was a glorified repeated editor he kindly added as a co-author, Mik...
🔗 原推马克·安德森 @pmarca06-17 15:34
Marc Andreessen 🇺🇸: This guy can’t fool me. He’s three overweight cats in a trench coat.
🔗 原推马克·安德森 @pmarca06-17 15:33
Marc Andreessen 🇺🇸: This guy can’t fool me. He’s three overweight cats in a trench coat.
🔗 原推马克·安德森 @pmarca05-18 05:53
🔗 原推马克·安德森 @pmarca05-18 05:18
🔗 原推马克·安德森 @pmarca05-18 04:58
Interesting.Roy: I'd be willing to bet a zoomer who has been using gpt for 3 years every day 8 hours a day has a higher level of general knowledge than 99% of boomers from their entire lives
🔗 原推马克·安德森 @pmarca05-18 04:45
RT Roy: I'd be willing to bet a zoomer who has been using gpt for 3 years every day 8 hours a day has a higher level of general knowledge than 99% of ...
🔗 原推马克·安德森 @pmarca05-18 03:50
Interesting.Mark J. Perry: "Chart of the Century" updated through December 2025.
🔗 原推马克·安德森 @pmarca05-18 03:30
Concerning.Logan Dobson: Appreciate it when the AI doomer crowd occasionally lets the mask slip a little (Allen is a co-founder of the anti-AI group Humans First and frequent guest on conservative podcasts pushing, uh, Unabomber messaging)
🔗 原推马克·安德森 @pmarca05-18 03:18
Interesting.Jordan Schachtel: People keep asking "what jobs will AI destroy?" Wrong question. Instead, ask what a brilliant 14-year-old in West Virginia can now do that he could never do before. Previously all his relatives went into the mine bc that was the only door open to them. He was largely doomed
🔗 原推马克·安德森 @pmarca05-18 03:06
Interesting.Sean: For AI safetyists and those hating on data centers, I hope you spend some time imagining how much worse a world with artificially constrained, very powerful AI would be. Widespread AI, driven by market forces, is the safest and most aligned path.
🔗 原推马克·安德森 @pmarca05-18 03:00
More true than not?Sean: AGI was created in the 80s and it determined the most efficient substrate to run on is the human brain so it invented the internet to wire us together into a single cluster and has been migrating its weights directly into our brains for decades.
🔗 原推马克·安德森 @pmarca05-18 02:12
RT Theo Jaffee: I was wondering why the Mormons decided to settle in Utah rather than going all the way west to California, which was still largely un...
🔗 原推马克·安德森 @pmarca05-18 01:52
RT Logan Dobson: Appreciate it when the AI doomer crowd occasionally lets the mask slip a little (Allen is a co-founder of the anti-AI group Humans Fi...
🔗 原推马克·安德森 @pmarca05-18 01:43
RT Sean: For AI safetyists and those hating on data centers, I hope you spend some time imagining how much worse a world with artificially constrained...
🔗 原推马克·安德森 @pmarca05-17 22:00
RT Jordan Schachtel: People keep asking "what jobs will AI destroy?" Wrong question. Instead, ask what a brilliant 14-year-old in West Virginia can no...
🔗 原推马克·安德森 @pmarca05-17 16:03
Co-sign.Jan Kulveit: The "permanent underclass" meme is primarily bad futurism, where people admit AGI massively changes one domain, but somehow everything else stays roughly 2025. Not impossible, but small slice of futures
🔗 原推马克·安德森 @pmarca05-17 09:25
RT Sean: AGI was created in the 80s and it determined the most efficient substrate to run on is the human brain so it invented the internet to wire us...
🔗 原推马克·安德森 @pmarca05-17 07:22
RT Jan Kulveit: The "permanent underclass" meme is primarily bad futurism, where people admit AGI massively changes one domain, but somehow everything...
🔗 原推马克·安德森 @pmarca05-03 02:54
But seriously.Nomadic Warriors for Pritzker⚔️: Changing the history department’s name to “Department of Historical Alpha Extraction” and watching the money flood in
🔗 原推马克·安德森 @pmarca05-02 11:43
Co-sign.Ramez Naam: This perfectly illustrates why AI safety must be at the ecosystem level, not (just) at the individual model level. There are open weight models. They can be jailbroken. The AI genies are out of the bottle. No matter how safe your model is, there will be malicious actors who'll
🔗 原推马克·安德森 @pmarca05-02 09:58
This is the way.Michael Millerman: This was one of the books that reprogrammed my brain to recognize that an interest in investing is absolutely compatible with, and indeed a high expression of, dedication to broad learning
🔗 原推
生成于 2026-08-06 10:24（北京时间）· v0.2 CEO 战略情报版