Phaidros

[Phaidros @phaidrosliang]

https://t.co/PNJpUQT5iw

[Phaidros @phaidrosliang]

https://t.co/6yBjMR3TEW

[Phaidros @phaidrosliang]

@indigox 666

[Phaidros @phaidrosliang]

@thsottiaux @Leooweb3 'call me dad', -- that should be your line.

[Phaidros @phaidrosliang]

AI evaluation today feels a lot like A/B testing in the internet industry circa 2012–2014. As an industry, we’re still at a very early stage in understanding how to do it well.

[Phaidros @phaidrosliang]

Distillation is a terrifying business model — and I mean distillation in the broadest sense.

[Phaidros @phaidrosliang]

This is the master of memes!

[brett goldstein @thatguybg]

that feeling when you're not in the group chat but you have anthropic shares

[brett goldstein @thatguybg]

that feeling when you're not in the group chat but you have anthropic shares

[Phaidros @phaidrosliang]

@DeepGapJournal @DavidSacks haha, you need to change the constitution first

[Phaidros @phaidrosliang]

@nicoletang0717 @thsottiaux @claudeai Are they dance battling???

[Phaidros @phaidrosliang]

I once made the analogy that top AI researchers are as rare as world-class footballers—and should be paid like superstars.

There’s another implication of that analogy: the resources you’ve consumed over your career—training facilities and competitive play for footballers; compute and tokens for AI workers—largely shape your capabilities, which in turn shape your market value.

In this era, tokenmaxing is simply rational.

If you’re ambitious, you should probably feel a little ashamed of leaving your quota unused.

[Phaidros @phaidrosliang]

Start with the free research and decide whether it is useful to you. Investing involves risk; my writing is general research, not personalized advice.
https://t.co/dax2aVQvHn

[Phaidros @phaidrosliang]

The question I want each note to answer: what would make this view wrong? Agreement is less useful when everyone is repeating the same source. I want the reasoning to remain visible when an outcome disappoints.

[Victor E. Nunez @victornunez]

hey Astra, how fast are you really at using a computer?

[Y Combinator @ycombinator]

Paul Graham on what drives founders through the hard parts:

“Being dutiful is not enough. The obstacles are too fearsome for mere dutifulness.”

“You know what actually motivates founders day to day? The fear of failure.”

“If your server’s crashing, you’re not thinking, ‘I got to deal with this because if I deal with it, I’ll become a billionaire.’ You’re thinking, ‘Oh no, the server’s crashing. It will be a disaster. The engine of my model train set is falling off the edge of the table. I need to go and save it.’”

[Phaidros @phaidrosliang]

One thing I’ve noticed about today’s LLMs in engineering: ideas that would once have been dropped immediately — just because it looked too resource-consuming, like migrating a large codebase or a legacy system — is now feasible.
That is real productivity.

[AIGCLINK @aigclink]

最近身边很多创业做FDE的朋友，非常执着于研究palantir的Ontology，甚至号称自己做了Ontology，我一直认为palantir本体论是一种倒退，不管是从技术层面还是业务层面，这个东西本质是个中台。

是不是有点似曾相识，中台这个东西都被搞烂了，前两年阿里就因为中台被嘲笑都是概念派（对比拼夕夕），今���需要的result as a service，而不是再倒退回去谈中台这种中间态。中台是为了方便后续开发和落地，以及降低边际成本搞的东西，而今天ai时代需要的不是这个，而是被验证的闭环过的ai员工，作为fde应该多搞能够交付结果且验证过能被审计的ai员工，而不是为了炒作概念搞本体论。

当前处于软件日抛的时代，稀缺的不再是软件或者中台，而是能确保交付结果且可被审计回滚的AI员工，这才是FDE的职责，为终端客户基于模型能力封装适合其的个性化AI员工。

国内现在人人都在喊自己是FDE，前段时间人人都在喊自己是OPC，最近喊opc的少了，是因为大多数被淘汰了少部分被留了下来，FDE也会经历大浪淘沙，但方向错了一切就都错了，且行且珍惜，祝愿fde都能找到自己的卡位。

[lody @lody_ai]

The developers contributing to Lody have given us an incredible amount of energy.

Our team has embraced open source since @loro_dev . It’s hard to explain the appeal—but building something with like-minded people and solving shared problems together feels genuinely amazing.

https://t.co/FQUvIbYOmc

[Phaidros @phaidrosliang]

This is actually a very simple problem: for data centers to remain sustainably profitable, the global economy—or at least the world’s major economies—must experience rapid growth driven by AI as the engine of a new industrial revolution. Otherwise, there simply won’t be enough demand to pay for all that capacity.

Such rapid economic growth would manifest itself in the emergence of unprecedented new products and entirely new sources of demand, much like the sudden arrival of railways, automobiles, and airplanes in previous industrial revolutions. Ten years from now, perhaps middle-class families will be spending their weekends vacationing on the Moon.

Conversely, if AI fails to drive meaningful growth in the broader economy, the idea that data centers can somehow continue enjoying extraordinarily high returns is pure fantasy. The obvious endgame would be a debt crisis.

So rather than worrying about the sovereign debt problems of developing countries, as some of you have suggested, you would be better off worrying about the AI bubble.

[Phaidros @phaidrosliang]

This is actually a very simple problem: for data centers to remain sustainably profitable, the global economy—or at least the world’s major economies—must experience rapid growth driven by AI as the engine of a new industrial revolution. Otherwise, there simply won’t be enough demand to pay for all that capacity.

Such rapid economic growth would manifest itself in the emergence of unprecedented new products and entirely new sources of demand, much like the sudden arrival of railways, automobiles, and airplanes in previous industrial revolutions. Ten years from now, perhaps middle-class families will be spending their weekends vacationing on the Moon.

Conversely, if AI fails to drive meaningful growth in the broader economy, the idea that data centers can somehow continue enjoying extraordinarily high returns is pure fantasy. The obvious endgame would be a debt crisis.

So rather than worrying about the sovereign debt problems of developing countries, as some of you have suggested, you would be better off worrying about the AI bubble.