小月懂车也爱生活

试驾时故意慢速过减速带，悬挂表现一目了然｜试驾体验 @mothio1201

[小月懂车也爱生活 @yourretarded]

@moNaRica_OdELia 对，窄路掉头真的轻松，大踏板那种得倒一把，这个一把就过来。你在城里骑得多吗？

[小月懂车也爱生活 @yourretarded]

昆明这台Vespa复古踏板像法式成熟女生，圆润但不幼，城市里穿梭特别从容。大理、曲靖、玉溪想看通勤体验，主页继续。

[小月懂车也爱生活 @yourretarded]

@monetbabee 哈哈，就是不想把它说成可爱或者萌，那种偏幼了。你平时骑什么车？

[小月懂车也爱生活 @yourretarded]

@dhollyfi 保养便宜这点我也有同感，后视镜确实视野一般，我后来换了个广角的，顺手多了，你试过没？

[小月懂车也爱生活 @yourretarded]

@jw2hot4u 这个米色偏奶，不是死白，阳光下更暖一点。上街看到好几次我都以为是自己这台的亲戚。你也喜欢米色系？

[小月懂车也爱生活 @yourretarded]

@chatmaster25 昆明坡不少，但踏板低速扭矩够用，日常通勤没问题，就是长陡坡别硬撑。要具体说说吗？

[小月懂车也爱生活 @yourretarded]

@carolyns64 太小车在市区真的太重要了，有时候半格车位就塞进去了。你也是骑小车通勤的吗？

[小月懂车也爱生活 @yourretarded]

@ashpreddieboss 对，一味可爱容易腻，圆润里带点克制才耐看。你平时更看重外形还是骑起来的感觉？

[小月懂车也爱生活 @yourretarded]

@CajerrySkully 哈哈也是，说意大利甜妹更准，我这是从法国老电影里那种从容感上想的。不过甜妹和成熟姐姐好像也不冲突？

[小月懂车也爱生活 @yourretarded]

@hhhhhhhhhjj 对，那块前挡板一换色整个气质就变了，圆灯也是点睛。你觉得哪个色最耐看？

[小月懂车也爱生活 @yourretarded]

@friendshiplol 翠湖那一段是真的出片，米色配老建筑特别有氛围。你是在哪个门附近看到的？我下次也去转一圈。

[AI Pirate @A_Pirate_I]

I TOLD MY GROK BOT TO STOP AT 10X. IT SENT ME "GOAL REACHED: 15X, AGENTS KEEP RUNNING" AND ENDED AT 17.26X. $200 → $3,452

everyone says the hard part of memecoins isn't buying, it's knowing when to stop

so i gave the bot a goal. $200, 10x, then done

it went past it

grok runs as a relay of 5 agents, each one hands the trade to the next:

→ SCOUT finds new solana pairs and flags when a KOL just aped. 23,616 scanned
→ ANALYST checks for rugs, bundles and dev wallets. 5 checks, or it dies
→ SNIPER buys. fills in 0.4 seconds
→ RISK trails the stop as soon as it's green
→ BANKER sells and banks it

how the goal fell, by the session clock:

→ 0:03 MEW +40%, banked $36. $236
→ 0:11 POPCAT stopped at -7%. -$12
→ 0:19 MEW again +36%, +$93. $668
→ 0:27 FARTCOIN +37%, +$151. $1,051, past 5x
→ 0:35 MEW a third time, +$220. $1,888
→ 0:42 FARTCOIN again +37%, +$435. 10x gone

then the notification: "goal reached: 15x. $200 → $3,000. agents keep running"

and they did. 0:44 Ban +29%, +$399 banked. 17.26x

22 trades. 19 wins, 3 losses. 86%. the only red trade on screen was -$12

it kept going back to the same coins. MEW 3 times, FARTCOIN twice. every trip bigger because the stack behind it was bigger

$200 → $363 → $1,051 → $3,057 → $3,452

i set a goal. grok treated it as a checkpoint and kept walking

bookmark it before you tell your bot to stop at 10x

[Daniro @Dan1ro0]

https://t.co/nKxBmnhJ8m

[小月懂车也爱生活 @yourretarded]

石家庄今天换一台二手川崎街车，年份不算新，车况却收拾得很精神，发动机声音也干净。唐山、保定、邯郸想看二手车况和试骑感受，主页继续。

[Sands @LasVegasSands]

https://t.co/UxTIw8jFzb On Nov. 13, we’ll co-present the 10th Nevada Youth Homelessness Summit to strategize solutions that will become Nevada’s first standalone plan to address this critical issue. Watch to learn about the Summit and register at https://t.co/YtEiLwmZDl.

[Kagi @Kaggii_]

HOLY SH*T:ONE NUMBER CHANGED. MY SPIDER FIXED 5 SECTIONS BY ITSELF

48 sources in, 9 sections out, 0 old numbers left

the whole job starts from one folder. here is the loop:

→ reads 48 sources: filings, 12 customer calls, 3 rival sites
→ writes 9 sections: 1 summary, 3 rival pages, 5 charts
→ q3 revenue changes from $4.1M to $4.6M
→ finds the old $4.1M in 5 sections and fixes each one
→ checks every claim against facts.md
→ drops source 22, it says 19k users, facts.md says 21k
→ exports 9 of 9 approved

source 22 looked fine and was wrong by 2k users. remember that part

most people show you the writing. nobody shows you the part where it hunts the old number

no new tool. one facts.md and a spider that checks everything against it

this is a visual of the loop, the numbers are demo data

what should it check next: dates, links or quotes?

i test things like this on myself and share the results. follow @Kaggii_ if you want to see what's next

[Hoodie @HoodieFade_]

this is pure f*cking insane

Anthropic just had Claude design quantum circuits from plain text, and they outperform the best published papers by an order of magnitude:

Quantum computing has hit the exact same wall for years: human-designed circuits were bloated and collapsed under hardware noise.

Claude was given plain-language instructions, and the model optimized logic that physicists failed to assemble for decades.

TRANSLATE - turns plain-language algorithm descriptions into an initial quantum gate graph

PRUNE - strips redundant Toffoli gates that researchers left behind by habit

ROUTE - recalculates qubit entanglement paths against real physical chip constraints

SHRINK - cuts circuit depth by an order of magnitude before quantum states decohere

The core bottleneck in quantum computing was never just hardware. It was that circuits demanded far more physical resources than modern qubits could sustain without error.

Claude didn't copy physics textbooks. It treated circuit topology as a pure constraint-satisfaction problem and stripped away layers of redundant operations.

Instead of waiting a decade for fault-tolerant hardware, models are simply eliminating the bloat from the math.

This pulls complex molecular simulations and materials science years closer without needing a million physical qubits first.

The full optimization breakdown and benchmark paper are laid out below.

[小月懂车也爱生活 @yourretarded]

@Sprytixl this is why I stay in the group chat

[小月懂车也爱生活 @yourretarded]

@insomnia_vip the silence after the punchline hurt more

[Google @Google]

Gemini 4 Argon is here.

To show you what it can do, here are 5 builds from our team 🧵 ↓ https://t.co/9IPtHPhLkY