Alex Yeh

Founder & CEO of @GMI_Cloud I Nvidia Reference Platform Cloud
Board Director @Robostrategy

[Marc Andreessen 🇺🇸 @pmarca]

Interesting. (Founder of https://t.co/PY9jRPdR6L, creator of GLM AI models.)

[IT Guy @T3chFalcon]

The reason is quite hilarious 😂😂.

Microsoft put $50 billion into Anthropic.

FIFTY billion dollars.

they are a Project Glasswing partner. Fable 5 runs inside Azure. Microsoft sells Claude to its own enterprise customers through Microsoft 365 and GitHub Copilot.

and they won't let their own employees use it.

here's why.

under Anthropic's new Mythos-class data retention policy, every prompt you type and every response you get is stored for 30 days. automatically. no opt out.

if their safety classifiers flag anything in your session, anything, they keep it for up to two years.

you don't get told when that happens, what was flagged or who can see it.

Microsoft employees paste confidential contracts into these things. customer data. internal roadmaps. acquisition strategies. legal documents. source code.

all of it sitting on Anthropic's servers for 30 days minimum. flagged sessions for two years.

so the company that invested $50 billion looked at that policy and told its staff: actually hold on.

other Claude models still work internally. under Zero Data Retention rules. the normal ones are fine.

just not the most powerful one they helped fund.

and one more thing.

the Pentagon listed Anthropic as a supply chain risk in March and banned defense contractors from using its products.

Microsoft funds Anthropic. sells Anthropic's models. runs them on Azure. helped build the most powerful one.

won't let employees use it.

the Pentagon won't let defense contractors near it.

the safeguard that makes Fable 5 safe enough to release publicly is the same safeguard that lets Anthropic keep your data for two years.

the guardrail is a data retention policy.

but you can use it. it's in your browser right now. 🌚

have fun.

[GRITCULT @GRITCULT]

Learn to confidencemaxxx and watch reality bend your way. Winners act like winners. This is your sign to start believing in yourself. This is your sign to start becoming delusionally confident in yourself and the fact that it will all work out.

[Phoenix Yin @Phoenixyin13]

认真读完了OpenAI 研究员 Noam Brown 今天的长帖，一个被行业严重低估的现实。
LLM 的真实能力天花板，远高于当前任何 benchmark 所显示的水平。

原因，是给它的test-time compute太少了。而随着模型迭代，这个问题会越来越突出。

GPT-5.5 发布初期，benchmark 只比 5.4 略有提升，很多人觉得没那么惊艳。
但实际体验是明显的 step-change。
为什么？因为评估时没有控制 test-time compute。
一旦把 tokens 放在 x 轴画曲线，5.5 在相同计算预算下明显更强。单个数 benchmark 把这个差距掩盖了。

更关键的观察是，更强的模型，从额外 test-time compute 中获益更大。
性能提升曲线更陡，plateau 被不断推远，甚至可能不存在。
Karpathy 的 autoresearch 实验几百次实验后性能仍在上升。
AI Security Institute 的网络攻防评测在 1 亿 tokens 后，顶级模型仍在快速进步。
这说明能力天花板可能比我们想象的远得多。

这对 AI 安全与 preparedness 的影响非常大。
当前大部分安全评测只在低 inference budget 下进行。
但真正危险的能力，尤其是长 horizon、多步规划、复杂 agent 行为，很可能只在高计算预算下才显现。
Gemini 3 Deep Think 的争议，暴露了这个问题。scaffold加上高 inference 能让普通用户接近高能力，但 system card 没反映这条曲线。

Noam 指出一个更棘手的现实，即长 horizon 任务的评估越来越困难。
有些能力只能通过实际长时间运行才能验证，而模型迭代周期可能已经短于这个验证时间。
未来 labs 可能会面临一个两难，要么大幅延迟发布，要么在不完整的安全评估下发布。

两年多过去了，从 o1 开始，行业依然习惯报单个数，安全组织依然会对多花 100x inference 就能显著变强感到惊讶。
这其实是 post-o1 时代最被低估的 paradigm shift。
未来模型之间的真实差距，更会体现在谁能更好地理解和利用 intelligence-compute 曲线。
对于做 agent、长 horizon 推理、高风险部署的人来说，重新理解这个曲线，已经成为了重中之重的必选项。

[Alex Yeh @alex_yehya]

caught a mini me in the left bottom corner 😆 from @NVIDIAGTC

[Alex Yeh @alex_yehya]

Interesting take

[看不懂的SOL @DtDt666]

炸裂！马斯克说出了所有经济学家都不敢承认的话：现有整个经济体系即将崩塌，且没有任何力量能够阻止。

马斯克：“价格会猛烈崩盘。”

不是小幅下跌，是彻底崩塌。

人工智能和机器人并非在创造增长，它们正在摧毁经济学赖以成立的“稀缺性框架”。

马斯克直言：“它会像超音速海啸一样砸向我们。”

生产呈指数级爆发，货币供应却只是线性增长；生产率维持持续的两位数扩张，那些听上去不可思议的数字，即将变成基本常态。

这不是渐进式进化，而是彻底的替代。

马斯克：“价格会猛烈崩盘。”

不是小幅下跌，是彻底崩塌。

AI 剥离人力成本，消灭生产失误，清除所有推高商品价格的低效环节。
任何商品的边际生产成本都将趋近于零，同时产品品质还在持续提升。

各国政府只会本能应对：印钱、放水、刺激经济。
这套专为稀缺型经济设计的套路，撞上它们完全无法理解的物质极大丰裕，会彻底失效。

马斯克：“GDP 指标早已失去意义。”

所有传统经济模型，都建立在劳动力有限、产出受限、效率缓慢提升的假设上。
而 AI 根本不在这个框架内运行——它直接把这些前提变量抹掉。

生产暴增，央行狂放水，但价格依然会崩盘。
因为实物供给的爆发速度，远超任何货币干预能对冲的程度。

生产浪潮的速度，永远跑在政策前面。

在所有历史模型里，通缩都是危机信号。
但这一次，不是需求崩塌，而是供给走向无限。

经济没有失灵，它只是彻底超越了那些为衡量“稀缺”而设计的工具。

真正的权力，会落到掌控无限产出系统的人手里。
当生产成本趋近于零，金钱就会变成次要东西。
决策者还在用适配“资源有限”的仪表掌舵，而那些限制早已不复存在。

这一切已经开始。
而那些主导经济规则的人，对于自己整个行业一夜之间被淘汰的结局，完全没有答案。

[Alex Yeh @alex_yehya]

lol

[Vivek Sen @Vivek4real_]

GEORGE HOTZ: “YOU CAN BREAK AI DOWN INTO 5 TIERS.”

“DATA CENTERS — TIER 1, FABS — TIER 2, NVIDIA/AMD — TIER 3, OPENAI/ANTHROPIC — TIER 4, AND COMPLETELY WORTHLESS THINGS LIKE CURSOR AND WINDSURF, WHICH ARE TIER 5.”

“OPENAI AND ANTHROPIC WILL EAT ALL THE VALUE FROM THE CURSORS AND WINDSURFS OF THE WORLD.

I ARGUE THAT THE TIER 4S (OPENAI/ANTHROPIC) AREN’T EVEN GOING TO HAVE VALUE.”

[Alex Yeh @alex_yehya]

Great progress for the AI community

[Sheel Mohnot @pitdesi]

Anthropic exp $10.9B Q2 revenue, up from $4.8B in Q1, and $559M of operating profit for the Q.

Compute is 56 cents per dollar of revenue, lower than I expected.

Q2 projection:

$10.9B revenue
-$6.1B compute
-$2.7B sales & marketing
-$1.5B other costs
=$559M operating income

[Alex Yeh @alex_yehya]

@WarcMeinstein @RoboStrategy Congrats bro. Look forward to work more closely together

[Alex Yeh @alex_yehya]

@Rewkang @adcock_brett @Figure_robot @jeffcardn @Apptronik @Lindon_Gao @DynaRobotics @evanbeard @standardbots @taochenshh @DexmateAI @PathRobotics @cixliv @REK @benedektasi @AllonicRobotics Wbm

[Alex Yeh @alex_yehya]

Still tilted 🤣

[Deedy @deedydas]

The vibes in SF feel pretty frenetic right now. The divide in outcomes is the worst I've ever seen.

Over the last 5yrs, a group of ~10k people - employees at Anthropic, OpenAI, xAI, Nvidia, Meta TBD, founders - have hit retirement wealth of well above $20M (back of the envelope AI estimation).

Everyone outside that group feels like they can work their well-paying (but <$500k) job for their whole life and never get there.

Worse yet, layoffs are in full swing. Many software engineers feel like their life's skill is no longer useful. The day to day role of most jobs has changed overnight with AI.

As a result,
1. The corporate ladder looks like the wrong building to climb.
Everyone's trying to align with a new set of career "paths": should I be a founder? Is it too late to join Anthropic / OpenAI? should I get into AI? what company stock will 10x next? People are demanding higher salaries and switching jobs more and more.

2. There’s a deep malaise about work (and its future).
Why even work at all for “peanuts”? Will my job even exist in a few years? Many feel helpless. You hear the “permanent underclass” conversation a lot, esp from young people. It's hard to focus on doing good work when you think "man, if I joined Anthropic 2yrs ago, I could retire"

3. The mid to late middle managers feel paralyzed.
Many have families and don't feel like they have the energy or network to just "start a company". They don't particularly have any AI skills. They see the writing on the wall: middle management is being hollowed out in many companies.

4. The rich aren’t particularly happy either.
No one is shedding tears for them (and rightfully so). But those who have "made it" experience a profound lack of purpose too. Some have gone from <$150k to >$50M in a few years with no ramp. It flips your life plans upside down. For some, comparison is the thief of joy. For some, they escape to NYC to "live life". For others still, they start companies "just cuz", often to win status points. They never imagined that by age 30, they'd be set. I once asked a post-economic founder friend why they didn't just sell the co and they said "and do what? right now, everyone wants to talk to me. if i sell, I will only have money."

I understand that many reading this scoff at the champagne problems of the valley. Society is warped in this tech bubble. What is often well-off anywhere else in the world is bang average here.

Unlike many other places, tenure, intelligence and hard work can be loosely correlated with outcomes in the Bay. Living through a societally transformative gold rush in that environment can be paralyzing. "Am I in the right place? Should I move? Is there time still left? Am I gonna make it?" It psychologically torments many who have moved here in search of "success".

Ironically, a frequent side effect of this torment is to spin up the very products making everyone rich in hopes that you too can vibecode your path to economic enlightenment.

[AYi @AYi_AInotes]

前两天我发了篇推文说Meta是AI界最遗憾的公司，今天就刷大前Meta LLaMA团队的Peter Pang发的这篇3500字长文，是我今年看到的最有实战价值的AI-first落地报告。

它把所有喊了半年的AI-first空话，变成了可复制的步骤和可验证的数据。

现在所有人都在说AI辅助开发，说Copilot能提效20%，Pang认为这根本不是AI的正确用法。

在旧流程上叠AI工具，最多只能带来10-20%的效率提升，结构一点没变，
真正的AI-first，是先假设AI才是主要的代码构建者，然后把整个工程架构、CI/CD、测试流程、甚至组织分工，全部推倒重来。

最反直觉的一点来了：
当大多数人都以为AI-first的瓶颈是模型不够强，是工程师不会用工具，
他用真实数据告诉我们 AI写代码的速度从来不是瓶颈。

上下游的人工流程，才是卡死所有效率的死穴：

• PM写一份详细需求要两周，AI写完代码只要两小时

• QA测一轮要三天，AI部署只要两小时

• 再怎么招人，也永远追不上大厂的headcount增速

也就是说AI再快，只要有一个环节是人在拖，整体效率就还是人的效率。

所以他们根本没给工程师配更多Copilot，他们花了全部精力，建了一整套让AI能独立、安全、可靠干活的harness（驾驭系统）：
统一monorepo让AI能看到整个系统的所有代码，

全链路结构化可观测让AI能自己定位错误，Claude 3-pass AI评审代替人工code review，

六阶段确定性CI/CD加Statsig feature flags和一键kill switch，

甚至做了一个自愈引擎，每天自动聚类生产错误，自动建工单，能修复的问题自己就解决了。

结果有多夸张呢？只用了14天，这个25人的团队，就做到了每天3到8次生产部署，出了坏功能，当天就能发现并回滚。

更反常识的是，部署频率翻了十几倍，用户指标和转化率反而上升了不少。

整个公司的分工被彻底重构，没有了天天写CRUD的工程师，人类只剩下两种角色：

架构师，负责设计规则和SOP，批判AI的输出，

验证者，负责判断风险和质量。

而且这不是只在工程部门，营销、产品发布、客户支持，全公司所有职能都在往AI-native转。

CTO花在日常管理上的时间，直接从60%降到了不到10%。

最扎心的是，这一切没有任何黑科技，所有工具都是开源的，所有流程都写在了文章里，任何人都能抄。

所以，真正的门槛从来都不是技术，是你愿不愿意承担转型的真实成本：比如员工的焦虑，资深工程师的抵触，连续几个月每天18小时的试错。

大多数公司宁愿守着10%的提效舒适区，也不愿意打碎自己运行了十几年的旧系统。

最让我震撼的是，这套工程逻辑，1:1平移到个人身上也完全成立。

你的笔记库就是你的个人monorepo，
你的AI助手就是你的专属agent，
你需要的也许更多的AI工具，而是一套属于你自己的认知harness和自愈循环。

让AI每天帮你扫描思考的漏洞，聚类你的认知gap，自动生成迭代计划。

当然，这只是他们一家的早期经验，不是所有公司都能直接复制。

但也给我们指明了一个清晰的方向：
未来的竞争，不仅看谁会用更多的AI工具，还要看谁先愿意把自己从一个执行者，彻底变成一个架构师和批判者。

[Alex Yeh @alex_yehya]

Congrats!! Hope it will be open-sourced soon 👀

[Alexandr Wang @alexandr_wang]

1/ today we're releasing muse spark, the first model from MSL. nine months ago we rebuilt our ai stack from scratch. new infrastructure, new architecture, new data pipelines. muse spark is the result of that work, and now it powers meta ai. 🧵

[GMI Cloud @gmi_cloud]

Day 0 support for GLM-5.1 is LIVE on GMI!

This new upgrade means
✅ No Plateaus: It stays productive over 100s of rounds.
✅ Agentic Engineering: SOTA on SWE-Bench Pro (58.4).
✅ Persistence: Handles 1000s of tool calls without losing context.

The era of "vibe coding" is over. The longer it runs, the better the result.

Congrats to the @Zai_org team!

[Alex Yeh @alex_yehya]

here is their technical report for anyone who is interested: https://t.co/FBnSoVHlG2

[Alex Yeh @alex_yehya]

Been reading https://t.co/iOhvGh3AFf’s technical report on GLM-5 and their latest blog on GLM-5.1, and one thing stands out:

In just over a month since launching GLM-5, @Zai_org reports a 28% gain in coding capability with the 5.1 update.

Beyond the 58.4 SWE-Bench Pro result, what caught my attention more is its endurance on long-horizon agentic tasks.

It doesn’t seem to stall after 50 turns. It keeps going through 600+ iterations and 6,000+ tool calls.

That kind of persistence matters.

Good job https://t.co/iOhvGh3AFf team!

[Z.ai @Zai_org]

Introducing GLM-5.1: The Next Level of Open Source

- Top-Tier Performance: #1 in open source and #3 globally across SWE-Bench Pro, Terminal-Bench, and NL2Repo.
- Built for Long-Horizon Tasks: Runs autonomously for 8 hours, refining strategies through thousands of iterations.

Blog: https://t.co/hmyDe4Nel3
Weights: https://t.co/CuUjXcPKJD
API: https://t.co/fz6reja4fb
Coding Plan: https://t.co/Nk8Y98HNhU

Coming to https://t.co/WCqWT0qCQb in the next few days.

[Alex Yeh @alex_yehya]

the benchmark here is quite stunning, tbh

sad that they are not releasing it to public

[Alex Albert @alexalbert__]

We released Claude Opus 4.6 just two months ago. Today we're sharing some info on our new model, Claude Mythos Preview.

[Alex Yeh @alex_yehya]

As someone in semi, I can validate this

[Dwarkesh Patel @dwarkesh_sp]

If Huawei hadn't been banned from TSMC in 2019, @dylan522p thinks it would have already eclipsed Apple as TSMC's biggest customer.

And that it would have better AI chips than Nvidia.

Before it was banned, Huawei was actually the first to ship a 7nm AI chip — two months before Google's TPU and four months before Nvidia's A100.

[Scale at GMI @scale_at_gmi]

Can’t find us? Want the full picture of our programming for Scale?

We’re here: https://t.co/JlVsLUt0UV

Our mission is always:
• Help founders build faster
• Share real startup insights
• Support builders worldwide

Building an AI startup? Follow us. Big things coming.