Why SPAR Exists
🇬🇧 English Version
Why SPAR Exists
Most AI assistants are designed to agree with you.
They’re helpful. Polite. Encouraging. They validate your thoughts and offer solutions.
But they won’t challenge you. And they won’t remember you.
The Real Problem
You don’t think in one sitting.
Good ideas come in fragments:
- A shower thought on Monday
- A coffee chat insight on Wednesday
- A late-night realization on Friday
But here’s what happens with most AI tools:
You ask:
“Do you remember that idea we discussed last week?”
They answer:
“I don’t have access to previous conversations.”
Every session is isolated.
Past insights get buried.
You’re the one who has to remember, organize, and connect.
You become the memory system. AI is just the answering machine.
This is backwards.
Thinking Needs Two Things
1. Friction (Challenge)
Good ideas don’t come from validation. They come from being challenged.
- Scientists need peer review — not cheerleaders
- Startups need critical investors — not just believers
- Athletes need tough coaches — not just supporters
Your thinking needs the same thing.
But most AI assistants are too agreeable. They say “Great idea!” when they should ask “Why do you think that?”
SPAR pushes back.
It questions your assumptions. Finds holes in your logic. Forces you to defend your ideas. Makes you think harder.
Not to be mean. Not to be contrarian for its own sake.
But because sharp thinking requires resistance.
2. Memory (Continuity)
But challenging you isn’t enough.
Your best thinking happens over time:
- Week 1: First draft of an idea
- Week 2: You forget about it
- Week 4: A related thought appears
Most AI tools can’t connect these. You have to.
SPAR remembers for you.
When you come back to an idea weeks later, SPAR knows:
- What you said before
- What’s still unresolved
- How new thoughts relate to old ones
You don’t finish in one sitting. SPAR picks up where you left off.
This Is What SPAR Does
SPAR (砺) means “whetstone” in Chinese — the thing that sharpens a blade through friction.
SPAR is an AI that does two things:
It Challenges You
- Questions your assumptions
- Finds holes in your logic
- Forces you to defend your ideas
- Makes you think harder
It Remembers for You
- Tracks your in-progress topics over weeks
- Automatically links new thoughts to old ideas
- Surfaces relevant past discussions when you need them
- Connects fragmented insights into coherent strategies
Not just a sparring partner. A thinking partner with a memory.
Real Example: The Restaurant Project
Week 1 (Monday):
You: “Thinking about opening a family restaurant…”
SPAR: “Why a restaurant? What experience do you have? Have you run the numbers?”
You: “Um… not yet. Just thinking out loud.”
SPAR: “What problem are you solving that existing restaurants aren’t?”
Week 2: [You forget about it]
Week 4 (Friday):
You: “Maybe I should start with a small cafe instead…”
Most AI tools: “That’s a great idea! Here’s how to start a cafe…”
(Doesn’t know about Week 1. Fresh start.)
SPAR: “You mentioned ‘restaurant’ 3 weeks ago. Now ‘cafe’. Same vision or pivoting?”
You: “Oh right… I forgot about that. Maybe cafe is more realistic?”
SPAR: “Why more realistic? Is it capital constraints or operational complexity?”
This is the difference:
- Standard AI would treat Week 4 as a new conversation
- SPAR connects it to Week 1 and keeps challenging
You get both: confrontation + continuity
The Design Choices
I made two unconventional choices:
1. Confrontational instead of friendly
I could have made SPAR encouraging and supportive.
I chose confrontational instead.
Why? Because in my testing:
- People say they want encouragement
- But they need to be challenged
- And after being challenged, they think more clearly
SPAR doesn’t make you feel good. It makes you think better.
2. Persistent memory instead of fresh starts
I could have made SPAR like most AI: every conversation a clean slate.
I chose persistent memory instead.
Why? Because:
- Good thinking takes time
- Ideas evolve over weeks, not minutes
- You shouldn’t have to remember what you said last week
SPAR is your external brain. You think. It remembers.
What Makes It Different
1. Confrontational by design
Not “That’s a great idea!” — Instead: “What evidence do you have for that?”
2. Persistent memory across time
Not “New conversation, who dis?” — Instead: “You mentioned this 3 weeks ago. Still thinking about it?”
3. Automatic topic management
You don’t organize. You don’t tag. You just talk.
SPAR figures out what belongs together.
4. Voice-first experience
Speaking forces you to think in real-time. No time to polish your argument. Raw thinking, exposed and examined.
Who This Is For
SPAR is for people who:
- Have important decisions to make
- Want to stress-test their thinking
- Know that “yes-men” are dangerous
- Prefer clarity over comfort
It’s not for everyone.
If you want validation, there are plenty of AI assistants that will give it to you.
If you want your ideas sharpened and remembered, there’s SPAR.
The Belief
I believe:
- Most people’s thinking is fuzzier than they realize
- Clarity comes from being forced to defend your ideas
- The best thinking partners are the ones who disagree with you
- And remember what you said
SPAR is that thinking partner.
Try It
Your ideas deserve to be challenged. And remembered.
Not destroyed. Sharpened.
Not scattered. Connected.
That’s what SPAR does.
🇨🇳 中文版本
为什么做 SPAR
大部分 AI 都是设计来同意你的。
它们有礼貌、很有用、很鼓励人。验证你的想法，给你答案。
但它们不会挑战你。也不会记住你。
真正的问题
你不可能一次就把事情想清楚。
好想法都是碎片：
- 周一洗澡时冒出来的
- 周三喝咖啡聊出来的
- 周五半夜突然想到的
但大部分 AI 工具是这样：
你问：
“上周咱们聊的那个想法，你还记得吗？”
它说：
“抱歉，我看不到之前的对话。”
每次都是新的开始。
之前的想法全埋了。
你得自己记、自己整理、自己串起来。
你成了记忆系统，AI 只是个答题机器。
这不对。
思考需要两样东西
1. 摩擦（有人挑战你）
好想法不是夸出来的，是怼出来的。
- 科学家需要同行评审 — 不是粉丝团
- 创业的需要质疑你的投资人 — 不只是信你的人
- 运动员需要严格的教练 — 不只是加油的
你的想法也一样。
但大部分 AI 太好说话了。该问”你凭什么这么想？”的时候，它说”好主意！”
SPAR 会怼你。
质疑你的假设。找你逻辑的洞。逼你把想法说清楚。让你想得更深。
不是杠精。不是为了反对而反对。
是因为好的思考需要阻力。
2. 记忆（能接上话茬）
但光怼你还不够。
你最好的思考都是跨时间的：
- 第一周：想法刚冒出来
- 第二周：忘了
- 第四周：又想起相关的东西
大部分 AI 连不起来。你得自己记。
SPAR 帮你记。
几周后你回来继续想这事儿，SPAR 知道：
- 你之前说了啥
- 哪些还没想清楚
- 新想法和旧想法怎么接上
你不用一次想完。SPAR 接着上次的继续。
SPAR 是干嘛的
SPAR（砺）就是磨刀石 — 靠摩擦把刀磨快的那个。
SPAR 是个会做两件事的 AI：
挑战你
- 质疑你的假设
- 找你逻辑的漏洞
- 逼你把想法说清楚
- 让你想得更深
帮你记
- 追踪你正在想的事，哪怕过了几周
- 自动把新想法和旧想法连起来
- 需要的时候把之前聊过的翻出来
- 把碎片化的想法整成完整的思路
不只是个对手。是个有记忆的思考伙伴。
真实例子：开餐厅这事儿
第一周（周一）：
你：”在想要不要开个家庭餐厅…”
SPAR：”为啥是餐厅？你开过吗？算过账没？”
你：”呃…还没算。就是想想。”
SPAR：”你要解决啥问题是现在的餐厅解决不了的？”
第二周： [你忘了]
第四周（周五）：
你：”要不先开个小咖啡馆？”
普通 AI：”好主意！开咖啡馆要这样…”
（不知道第一周的事。重新开始。）
SPAR：”你 3 周前说的是’餐厅’。现在’咖啡馆’。换方向了还是同一个事儿？”
你：”哦对…我都忘了。咖啡馆可能更靠谱？”
SPAR：”为啥更靠谱？钱不够还是觉得餐厅太复杂？”
这就是区别：
- 普通 AI 把第四周当新对话
- SPAR 连回第一周，继续怼你
你得到的：既有人怼，又能接上话茬
我的设计选择
我做了两个不走寻常路的选择：
1. 怼你，不哄你
我本来可以让 SPAR 温柔鼓励。
我选了怼人路线。
为啥？因为测试的时候发现：
- 大家嘴上说想要鼓励
- 但其实需要被挑战
- 被怼过之后，想得更清楚
SPAR 不让你爽。但让你想得更明白。
2. 一直记着，不是每次重来
我本来可以让 SPAR 像其他 AI：每次都是白板。
我选了一直记着。
为啥？因为：
- 好好想个事需要时间
- 想法得演化几周，不是几分钟
- 你不该记自己上周说了啥
SPAR 是你的外部大脑。你想，它记。
哪里不一样
1. 设计上就是要怼你
不说”好主意！” — 说”凭什么？”
2. 跨时间记忆
不说”新对话，你谁？” — 说”你 3 周前提过这事儿，还在想？”
3. 自动整理话题
你不用分类。不用打标签。你就说。
SPAR 自己判断啥和啥是一回事。
4. 语音优先
说话逼你实时想。来不及修饰。原始的想法，直接暴露，直接检验。
适合谁用
SPAR 适合这种人：
- 有重要决定要做
- 想测测自己的想法经不经得起推敲
- 知道”应声虫”害人
- 要清楚，不要舒服
不适合所有人。
想要认可，一堆 AI 等着夸你。
想磨练想法还能记住，就 SPAR。
我的想法
我觉得：
- 大部分人的思考比自己以为的更模糊
- 想清楚得靠被逼着把想法说明白
- 最好的思考伙伴是那些跟你意见不一样的人
- 而且还记得你说过啥
SPAR 就是这样的伙伴。
试试
你的想法值得被挑战。也值得被记住。
不是被打烂。是被磨快。
不是散了。是串起来。
这就是 SPAR 干的事。