林宥恒

Solo builder road to 100k/MRR

Entrepreneur & Investor. Building in public.
Tech & Ai Ventures.

[林宥恒 @TR__Lin]

韩国股民杠杠爆仓了跑外卖想回本结果发现不让上杠杆了😅

[林宥恒 @TR__Lin]

我昨天刚刚花了一个重置额度🙂

[Tibo @thsottiaux]

Oops... I did it again.

Enjoy reset usage limits for all paid users for Codex and ChatGPT Work. Super grateful for an incredible team who is iterating at lightspeed and keeping the infra up as we scale faster than ever.

Enjoy the weekend!

[林宥恒 @TR__Lin]

我发现用fable5，提到kimi他就给我降级opus4.8了…正是严防死守啊

[林宥恒 @TR__Lin]

@dontbesilent 我做了个skill可以实时全量拉取得，微信这边对聊天记录拉取没啥限制。

[林宥恒 @TR__Lin]

新加坡小归小，每天日出日落还是好看的🤭

[林宥恒 @TR__Lin]

@Corsica267 wc之前看到乌老师发的结合访谈自己做了个，没想到直接开源了。无私奉献的乌老师😍

[林宥恒 @TR__Lin]

@BabybusFL 推一推让我下单一下

[林宥恒 @TR__Lin]

知道什么叫绝望吗？

[林宥恒 @TR__Lin]

分享一个有意思的网站：
https://t.co/Bg6BH5RaMC

Kimi K3 自己做的 4399。

据说都是一句话 prompt 做出来的。

那个 macOS Web Simulator，超级丝滑，跟在真实 Mac 上操作一毛一样。

[林宥恒 @TR__Lin]

特别是Kimi K3出来以后，50倍的价差下大家一定会比较的。如果我能以2%的价格获得最接近甚至超越最前沿大模型的性能，为什么要付那么高的价格？anth说下周上opus5，我并不认为会超过fable5，那就完全是苟延残喘了，我月初刚刚续费的max20x啊！！！心痛。

[qinbafrank @qinbafrank]

一周前聊到的大模型价格战是真的来了，今天几家 AI 大模型公司不约而同做了同一件事：
1）Anthropic 宣布重置所有用户的 5 小时和每周使用限制；
2）OpenAI 再次提高 Codex 和 ChatGPT Work 的使用额度。注意“再次”两个字，说明不是第一次放开限制，而是近期连续多轮上调；
3）xAI 不仅重置了 Grok 的使用限额，还开放了 Grok Build；
4）Cursor更直接，将所有订阅计划的模型使用额度翻倍。

看到这类动作，大家最关心两个问题：为什么要这么做？影响是什么？
1、先说为什么要这么做？
核心还是这两周一起聊到的https://t.co/uiVFMdYxdd 企业采用AI进入分场景、多模型的工程化阶段，企业自己也需要把高频、低复杂度和低价值场景分配给高性价比token、这也倒逼了解前沿闭源大模型主动应战，价格战永远是最有效的策略。

2、再说有何影响？
1）短期来看，大家担心是不是影响大模型的利润率。毕竟价格下降意味着单次调用收入降低，单位 Token 的商业价值下降，短期毛利率和盈利能力自然会承压。
2）但中长期，则是要看能激发出多大的幅度的需求量增长。如果需求爆发式增长、token的消耗量增长幅度大大抵消token单价的降低，最后总盘子变大了。上周在这里https://t.co/mrdtZbY8gT也有聊过：“越多企业采用AI，成本工程化做的越好，高端模型的份额被部分蚕食，但会不会带动商业化份额绝对值应该还是在快速增长？因为盘子更大了。”

但这里最关键的就是这个短期和中长期之间的时间差有多大。降价毛利和盈利能力会立马下去，需求的爆炸是不是能立马补上来，还是需要一点时间才行？

[林宥恒 @TR__Lin]

@zephyr_z9 @scaling01 How can opus 5 compete with 5.6 sol and Kimi k3 if it is not even better than fable5?

[林宥恒 @TR__Lin]

@Corsica267 @StrangerSsan wc乌老师是老乡啊，我周天回福州🤣

[林宥恒 @TR__Lin]

@Pluvio9yte 烧的太快了，我攒了四次充值一天用一个

[林宥恒 @TR__Lin]

@wallen97754394 那么就是半导体设备相关了🤣klac，uctt那些

[林宥恒 @TR__Lin]

打算开始用X记录我的创业+投资过程了，来新加坡十年了，马上也要从学校毕业了，给自己定下了以后的路线。

我本身是商科毕业，大学期间学的商务分析，所以也算ai+投资圣体了🤣

其实从去年年初开始，我就开始用ai做一些项目，并确确实实的把“做事情”这件事给提效了很多。加上我个人从17年到现在一直有在美股投资，也希望能有人多交流交流，只是之前一直不想公开记录这些。

开始分享内容一方面是我想给自己的思考留下一些痕迹，也多一些不同视角的碰撞。

我会在这里记录对 AI、创业和一些市场的思考，当自己的日记好了

[Navalism @NavalismHQ]

Don't write to make money, write to build relationships with like minded people that you haven't yet met.

@naval

[Citrini @citrini]

Since CPI is either going to be a clearing event or a thanos snap event, I figured it might be a decent time to be helpful.

If you’re relatively new to investing, then you’re bound to learn about what happens when your portfolio goes down.

Now, this is coming from someone who’s portfolio is up a paltry 20% YTD and hasn’t been gunning it on risk recently. Although also someone who’s been trading and investing their own money for the better part of a decade and has managed to not go bust (except for one, very painful time early on).

If you’re constructing a portfolio it’s important to realize it is its own position rather than a collection of positions. A stock is not just a company but the sum of its valuation, shareholder base, its sensitivity to liquidity, crowding, financing/rates and its catalyst calendar. High beta stocks are often five different trades in one ticker.

You should always have a working idea of your “tilts”. Does your portfolio go up/down more if tech rallies, if certain countries outperform, if a specific thematic is validated etc etc.

In general, you should not have a portfolio of 20 different stocks that all act the same. You might think you won’t, but getting the value of your book cut in half will make you do stupid things. (As an aside, this is also why even though buying the dip is generally a good strategy, progressively buying the dip early into a drawdown can make you mess up at the exact lows.)

High beta stocks come in all shapes and colors but in general they are selling the distant future. In good markets, the time out to that future is cheap. In bad markets, you start paying rent. That rent tends to appear as a lower multiple even while estimates stay the same.

A lot of times people will tell you the only thing that matters in a drawdown is “is the thesis intact?”. That’s one aspect, the other two are “how have expectations changed?” and “did you size like an idiot?”.

Don’t average down just to improve your cost basis, the market doesn’t care about your cost basis. Only add if you can truly underwrite the expected return improving, and that means taking a view that goes beyond a default return to multiples that may be unsustainable.

Price can become a fundamental and technical sell offs can manufacture fundamental problems - reflexivity cuts both ways.

The best question you can ask yourself in a drawdown is “from here, what is the range of outcomes and what’s the best use of the next dollar?”. If early in a drawdown you note that every time one sector goes up your portfolio goes down, it could be a decent idea to add exposure to that sector.

The ultimate goal is not to avoid every drawdown but to make sure no single drawdown takes away your ability to act on real opportunities when they arise.

No amount of truisms will help make anyone a better investor, but there is something you can do right now. If this is one of your first few drawdowns, you can observe how you react. Take notes on it. Find out what mistakes you make and then optimize your portfolio, sizing, strategy etc to compensate for those shortcomings. It’s a lot easier to do that than try to fight your own psychology - and anyone who pretends there’s a one size fits all answer to that is lying.

[林宥恒 @TR__Lin]

🐱

[徐冲浪 @cyrilxuq]

https://t.co/TF46Z8pIIa

还是推荐投资新手好好看看这几本

youtuber kol普遍喜欢装神弄鬼，什么买了5年tsla所以财富自由之类的，这种都是运气而已，你万一买到某个东西归零了呢？

我可是把美国本科教材直接给你了，当然，这个东西看悟性，但是我觉得对大部分人来说，《货币金融学》都非常值得看，你可以别的不看，但是这本一定要看。

[林宥恒 @TR__Lin]

@Corsica267 乌老师不在之前那个日内群了吗