小栗说车日常记

原来不同路况要用不同的驾驶模式｜车知识 @marilyn_1003

[小栗说车日常记 @hamunya]

@ballooskee 是的，这种反差光看照片体会不到，得上手。你看完之后最意外的是哪一点？

[小栗说车日常记 @hamunya]

兰州这台八成新旅行SUV停着和开起来是两种感觉。停着时，后排和后备厢都照顾得很到位，像健身教练姐姐；真正把车带起来以后，视野开，动力也不拖，城市和快速路之间切换得很自然。细节上使用痕迹真实但不乱，反而更容易判断车况。白银、定西、天水想继续看，更多日常使用感受我放主页了。

[小栗说车日常记 @hamunya]

@manshine 定西目前这台还是得从兰州看，顺路的话我也可以帮你问问附近有没有同款。你那边预算和年份大概啥要求？

[小栗说车日常记 @hamunya]

@Melissarivera12 对，试过才明白，停着不起眼，开起来反而舒服。好多人看车只看外装了，你啥时候来兰州开一圈感受下？

[小栗说车日常记 @hamunya]

@CourtneyLamar 对，低速窜高速肉那种最累人，这台给油收油都挺线性的。你平时开车最看重这块的哪种表现？

[小栗说车日常记 @hamunya]

@zara1437 并线那一下我特意试过，80到100加速不算猛但跟脚，踩多少给多少，迟滞不太明显；你平时开什么车，我好拿它作个参照。

[小栗说车日常记 @hamunya]

@mzcent 哈哈停着确实容易被当奶爸车，底盘我跑快速路并线时没什么虚位，你之前是不是也开过这类车？

[小栗说车日常记 @hamunya]

@Royking55 同款车友握个手，后排坐仨大人也不至于挤，婴儿车确实不用硬塞，反而还有余量。

[小栗说车日常记 @hamunya]

@skatinkaiden 对，就是这个意思，我这车踏板和挡把的磨损跟里程对得上，看着心里有底。

[小栗说车日常记 @hamunya]

@tgarner09 白银的先记下，市区油耗我这两天堵车多点会具体测一下，你那边平时跑白银城区多不？

[Gyomei @Gyome1_]

holy sh*t, someone just dropped REAP, the prompt library for REA

1.7k stars in 5 days, and it turns "how do I even ask REA this" into copy-paste prompts

https://t.co/9WJMfuCf1V

the sections:

→ function tracking and binary analysis
→ software cloning and game logic reconstruction
→ porting to WASM
→ firmware, Android APK and kernel
→ Electron apps
→ vulnerability research, anti-cheat and DMA research for CTFs and security work

17 sections, swap the <placeholders> and run

every prompt forces the same rule: files, functions and addresses for every claim, confirmed facts kept apart from guesses

ask about one function at a time, the narrower the question, the more you can trust the answer

save this ⭣

[Marlen @Marlenuii]

3 AI agents working on the same codebase without stepping on each other's changes

Grok, Gemini, and Claude can operate as a coordinated coding team instead of 3 isolated chat windows

The interesting part isn't having more models writing code

It's giving them a shared execution system where every action can be checked before it reaches the codebase

Here's the workflow shown in the demo

→ SCOUT explores the repository, identifies relevant files, and maps dependencies before changes begin

→ BUILDER handles implementation, turning the discovered context into actual code changes

→ AUDITOR reviews the output, flags conflicts, and sends problems back into the execution loop

All 3 operate against one shared disk with a live dashboard tracking progress, active tasks, and agent activity

The real advantage is the feedback loop

One agent discovers the problem, another works on the fix, and a third checks what could break

Instead of manually copying context between Grok, Gemini, and Claude, the coordination layer becomes the system connecting them

The next step for AI coding isn't making one model write everything faster

It's making multiple models work on the same project without creating a bigger mess

Save this if you're experimenting with multi-agent coding workflows

[VALIX @RetroValix]

This guy used Grok Bot to build a directional repricing system on Polymarket

Result: +$149,920

It trades 5 minute BTC, ETH, and SOL Up/Down markets using fair value estimates and dynamic hedging

Several Grok Bot AI agents work together as one algorithm:

1. Looks for an undervalued outcome

It uses data from external sources to continuously update P(Up) and P(Down)

The model factors in the distance between the spot price and the opening level, time remaining, momentum, volatility, order flow, and a latency edge

For example:

> Down price = 67¢
> Model’s estimate for Down = 74%

It enters a position when it identifies an edge

After entry, the model keeps updating every few seconds. If confidence increases, the algorithm buys more of the same outcome

2. Hedges when the signal changes

If the edge starts to disappear, it may sell part of its original position

If the signal completely reverses, it sometimes buys the opposite outcome

Its Polymarket account: trinity42

The algorithm’s main edge comes from the gap between its own fair value estimate and the current market price:

> Buys undervalued outcomes quickly
> Increases position size as confidence grows
> Reduces exposure when the original signal starts to change

[小栗说车日常记 @hamunya]

@irongiantXBT Test pressing labels handwritten almost always grabbed my money deeper dollar stack.

[小栗说车日常记 @hamunya]

武汉今天碰到二手Moto Guzzi，像会讲故事的文艺轻熟女，车龄有了，那一下独特轻晃却仍然很有性格。宜昌、襄阳、荆州，完整试驾主页见。

[小栗说车日常记 @hamunya]

@sciencegirl ask about more clack than practical benefit ends with one of these

[小栗说车日常记 @hamunya]

@SpikeCalls clicked command for an hour just enough reward activity