RollingSting

AI Engineer

[RollingSting @RollingSting]

这就是大号韭菜，美国葛兰

[The Assembly @InTheAssembly]

Cathie Wood must be one of the worst investors of all time.

She bought $32.8M of $CBRS at $310, one day after the IPO closed at $311.

The stock now trades at $198. Down 37% in one month.

[RollingSting @RollingSting]

这个人是 Meta 的前 VP，以及 Deepmind 的前 VP

[Mat Velloso @matvelloso]

All day using GLM 5.2. Didn't miss much. First open model that passes the bar as a daily driver. Things are not going to be the same.

Damn, now I want to buy some serious hardware.

[RollingSting @RollingSting]

@LogicalThesis Meta will fail because of it's AI

[RollingSting @RollingSting]

有点意思，通过 Anthropic Fable 做 3d。结合 3d 打印机打破次元壁垒

[adam @adamdotnew]

We’re very excited to announce that Anthropic’s Fable 5 is a SOTA model at mechanical engineering tasks!

It can generate intricate working assemblies and mechanisms in a single prompt.

[RollingSting @RollingSting]

@caozlog 长期复利太难。投资比较反人类，不适合绝大多数人。在企业家身上是优点，在投资领域可能是缺点

[RollingSting @RollingSting]

@aleabitoreddit If DeepSeek catch up, Anthropic valuation will collapse

[RollingSting @RollingSting]

@antirez Cannot agree more

[RollingSting @RollingSting]

@CuiMao @DarioAmodei 1. 为啥提示词触发从 cc 转 codex？是 cc 对多机位 的调用做的不好么？2.你现在是通过 codex 转成多机位脚本再调用 seedance 生成？你是只描述场景么 3.从图做视频，每次抽卡，你是咋预览的？

[RollingSting @RollingSting]

Both dynamics will feed back into lower token usage growth.
I do not believe AI will turn everyone into an OPC, just as the mass adoption of paper, pens, and compulsory education did not turn everyone into writers.

[RollingSting @RollingSting]

Ultimately, people will realize that building is not the hard part — distribution and sales are. The replacement of human labor by token consumption will be constrained from two sides.

[Polymarket @Polymarket]

JUST IN: Uber’s COO says heavy AI spending is getting harder to justify, as higher token usage fails to show a clear payoff in consumer features.

[RollingSting @RollingSting]

On the production side, effective use of AI still requires experienced human operators. In many real-world scenarios, AI is not capable of getting things right in one shot. If allowed to operate without sufficient guidance, it does not just increase wasteful token spend;

[RollingSting @RollingSting]

On the consumption side, many AI-built products may fail to convert into real revenue. When that happens, users will become more rational and far more selective about what they choose to build, asking whether the expected value of a product can actually cover its token costs.

[RollingSting @RollingSting]

it can also introduce severe, even destructive, side effects. Human labor substitution will therefore plateau once these operational constraints become binding, likely showing up in the form of more failures, more instability, and higher error rates.

[RollingSting @RollingSting]

这个 Meta 被裁的兄弟说，这次裁员并不是裁绩效最差的百分之十，而是基于角色，团队和绩效，甚至可能包括薪酬在内的总体评估

[RollingSting @RollingSting]

@FT if u trust him......

[RollingSting @RollingSting]

@miles_pxgc Meta’s model effort is doomed to fail.
Top AI labs are led by top researchers.
Meta put a businessman with an outsourcing background in charge.
That tells you everything.
Meta will survive — but because of its social platforms, not because it wins in foundation models.

[RollingSting @RollingSting]

Google has DeepMind and still shipped Gemini.
Google I/O was supposed to show strength. It mostly showed how weak Google’s AI products still are.
Maybe Google isn’t winning AI product.
Maybe it’s just becoming the arms dealer. $Googl

[RollingSting @RollingSting]

The issue isn’t spending money on AI. The issue is when the spending is the only product.

[RollingSting @RollingSting]

There are two kinds of AI companies: product AI and financial-statement AI.
Product AI ships things people actually use — better models, better products.
Financial-statement AI ships capex, GPU purchases, org charts, and layoffs — then calls it “AI transformation.” $Meta $Goog