Santi Garza
① 他说过的话
10 条
我喜欢用的一个类比是:如果模型是引擎,那 harness 本质上就是引擎周围的那辆车。
指向原始笔记的链接
The analogy I like to use is that if the model is the engine, the harness is essentially the car around the engine.
—— Santi Garza · [03:50]
我想说,当前沿变化如此之快时,锁定只用一家提供商是有点危险的。
指向原始笔记的链接
Dangerous, I would say, to lock into just one provider when the frontier is changing so quickly.
—— Santi Garza · [14:58]
我要说,没有一个模型能在所有类别中都获胜。
指向原始笔记的链接
And I’ll say that no one model wins every category.
—— Santi Garza · [17:23]
你在为你不需要的大量参数付费。
指向原始笔记的链接
You’re paying for a ton of parameters you don’t need.
—— Santi Garza · [20:43]
它在用五个词推断你想说什么这件事上做得非常好,但它不能读懂你的心思。
指向原始笔记的链接
It does a really good job at inferring what you’re trying to say with five words, but it can’t read your mind.
—— Santi Garza · [27:09]
随着时间推移,这实际上会让你花钱,因为每一轮都在喂入所有这些来自失败尝试和出错轮次的额外上下文。
指向原始笔记的链接
That is actually going to cost you money over time because every turn is feeding all this extra context from failed attempts, turns gone wrong.
—— Santi Garza · [28:12]
AI 开发中最大的罪过之一——我甚至见过专家也这么做——就是永远地养着一个聊天。
指向原始笔记的链接
One of the greatest sins in AI development, which I see even experts doing sometimes, is like nursing a chat forever.
—— Santi Garza · [34:56]
首先,你以为你保留着的所有细节,你其实没保留住。
指向原始笔记的链接
First of all, all the details you think you’re keeping, you’re not keeping.
—— Santi Garza · [35:25]
智能体喜欢猜测。这有点是它们 DNA 里固有的。它们是猜测者。
指向原始笔记的链接
Agents like to guess. That’s kind of built into their DNA. They’re guessers.
—— Santi Garza · [38:22]
我用 Composer 构建,0.12 美元。相比之下是 32 美元。
指向原始笔记的链接
I built with Composer, $0.12. When compared with 32 US dollars.
—— Santi Garza · [57:02]
② 出现在这些集
1 集
③ 他谈到的
点进去有真内容 —— 本页主要出口
Martin · Cursor · 智能体 · token · harness · 上下文窗口 · 提示词 · 缓存 · 压缩 · 模型选择
