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

[小满陪你看好车 @butterfly20099]

广州今天试本田Gold Wing，像特别会照顾人的成熟姐姐，座椅和风挡都把长途舒适放在前面。佛山、东莞、惠州想看完整体验，主页有记录。

[小栗说车日常记 @Lillian522]

@OrkaTV that restoration job looks incredible, what software?

[0xbobaa @0xbobaaa]

the worlds 2026 market gives gen.g a 31% shot. it gives t1 16.5%

t1 lifted the cup in 2023 and 2024. the market has them fourth

worlds winner on dg3 right now:
> gen.g 31c
> hanwha life 18.5c
> bilibili gaming 17.5c
> t1 16.5c
> g2 4.8c, the best-priced western team

three teams sit within 2c of each other behind gen.g. that's the market admitting it has no idea who's second

the region market is clearer: lck 68.5c, lpl 22.5c, lec 6c. korea is priced to win it more often than not

my read: t1 at 16.5c is the mispriced one. november bo5s reward the team that peaks late, and peaking late is t1's whole identity

play-in started today in los angeles. swiss stage oct 23, grand final at the barclays center nov 14. every series trades on dg3, from the first play-in game to brooklyn

[小栗说车日常记 @Lillian522]

@Lmg_cito @jenna1417 你这个不主动搭理人的说法笑死，确实像。等保定场的动态

[小栗说车日常记 @Lillian522]

@softrockhardluv 老板娘这词用得太到位了哈哈，好奇临汾什么时候有试驾？

[小栗说车日常记 @Lillian522]

@rocsald 行吧，那我只管蹲天好了，能不能偶遇全看缘分，先谢你个方向。

[🌺Sai Fan @Sai60012]

@grok Jay Shani devay namah

[小栗说车日常记 @Lillian522]

@ai_explorer25 @tavus Exactly what I was thinking

[CyrilXBT @cyrilXBT]

Would you run a random script from a stranger on your computer?

That's what installing an unchecked AI skill is.

One study of 31,132 agent skills found:

→ 26.1% had vulnerabilities
→ 5.2% looked malicious

NVIDIA open sourced the scanner it uses before publishing skills to its own catalog.

SkillSpector checks for prompt injection, data theft, privilege escalation and poisoned MCP tools.

Gives every skill a 0 to 100 risk score.

Scan before you install.

Commands below

[小太阳的自驾小天地 @teddykoloveko]

大连今天试KTM Super Duke，像性格直接的运动御姐，站着已经有点不好惹，油门真正打开以后更不会跟你客气。沈阳、鞍山、营口，完整动态体验在主页。

[小栗说车日常记 @Lillian522]

@VhinEGee 朔州这边想看车得跑一趟，想问下这台是哪年的车、跑了多少公里？标题里没写清楚。

[小栗说车日常记 @Lillian522]

@thebadest07 @NadyaDhaifina 把战斗痕迹叙述成加分项的话，风险账容易被忽略，哪天真出问题最气的是自己。

[TIME @TIME]

The contentious race for Ohio governor took an even nastier turn: A new ad against Democrat Amy Acton starred her estranged mother https://t.co/Bf1mPgnABF

[Sammy @sammy_capital]

holy sh*t, someone just mapped the entire REA ecosystem

8 free repos that let your AI agent take apart a closed app, turn it into source code and verify it against the original

the drop:

→ https://t.co/TPrnFjivzs (319k ★ · 80.8k forks)

the engine: connects Claude Code, Codex and other agents to .exe files, APKs, Electron apps, firmware, WebAssembly and network protocols through 130+ MCP tools

→ https://t.co/zNoIJYj6ZD (867 ★ · 188 forks)

the menu: curated tools and resources for binaries, apps, firmware, file formats and protocols, sorted by task and link-checked every week

→ https://t.co/Ls6UHRiOU0 (66 ★ · 20 forks)

the proof: DX-Ball rebuilt from a Windows binary, REA investigates each function, the agent writes C/C++ and every rebuild gets compared against the original machine code

→ https://t.co/jEwgE0bUMf (18 ★ · 2 forks)

the hard mode: Touhou 4 reconstructed from 16-bit PC-98 code across 975 commits, with levels, bosses, rendering and bullet patterns turned into maintained C++ and assembly

→ https://t.co/gikm8U0oms (3 ★ · 1 fork)

the honest benchmark: three preregistered studies with Claude Code on Windows crackmes and real binaries, including the negative results, REA didn't raise the solve rate and roughly doubled cost

→ https://t.co/TvsQyoaieS (93 cases)

the gallery: official case studies, 60+ community projects and discussions built with REA, from rebuilt apps and printers to old games and Windows binaries

→ https://t.co/1J0NVI7WDx (v0.2.0)

the workbench: a local studio on top of REA where agents analyze a project, propose a reconstruction, show the diff and run tests, and nothing gets applied until a human approves

→ https://t.co/WXjde0iu1d (12,880 projects)

the memory: an index of open-source projects born from reverse engineering, from game decompilations and devices to APIs, file formats and AI clients, with a ready-made agent skill so you never reverse engineer the same thing twice

the formula:

closed app → REA → evidence → source code → verified rebuild

[xinces @xinc3s]

EVERYONE TRIED TO GET RICH ON AI BY BUYING NVIDIA. I MADE $48,300 FROM AI WITHOUT OWNING A SINGLE AI STOCK

here's the trade nobody on my feed was talking about

every AI data center is a building that eats electricity. one big campus can pull as much power as a small city. and it has to sign for that power years before it opens

so i stopped asking "which AI company wins" and started asking "who do they all have to pay?"

i gave the question to 3 agents:

→ HAIKU 5.5 read every US utility, power producer and grid supplier filing. 10-Ks, 10-Qs, 8-Ks, earnings calls. every mention of "data center", "large load", "interconnection"
→ JEV sorted it. one question per company: is this demand signed, with a number, or just talk on a call? it can't write a single word, it only picks. talk got thrown out
→ OPUS 5.5 read only the survivors and wrote one page each: how much of next year's demand is already under contract, and what the stock price assumes. it never saw the chart first

out of 140+ companies, 6 made it through. i bought 4

7 months later:
→ 3 of them up between 31% and 64%
→ 1 down 11%. sold it. the contract got delayed
→ net: +$48,300

i didn't predict anything. the agents just read what was already signed while everyone else watched nvidia's chart

in a gold rush, don't guess who finds gold. read who's selling the shovels

the prompt i used ↓

"Build me a research desk in this folder. Read-only, no brokerage.
1. Reader (haiku): pull the latest 10-K, 10-Q, 8-K and call transcript for every US utility and power producer from SEC EDGAR. Save every mention of data centers, large loads or interconnection with the number and a source link.
2. Gate (jev): for each company ask "is this demand signed, with a contract and a number?" Keep only scores above 0.85.
3. Analyst (opus, never opens prices): one page per survivor: signed demand vs current capacity, what has to go right, the one fact that would kill the idea.
4. Code adds the stock move last. Show me the agent files first. Nothing runs until I say go."

[小栗说车日常记 @Lillian522]

@0xMovez print file new below and watched night car some details, mind you wanted so turned all kept weirding

[小栗说车日常记 @Lillian522]

@beamnxw was local witness quotes that asked first tiny and report side short creepy morning bit

[小栗说车日常记 @Lillian522]

@beamnxw was local witness quotes that asked first tiny and report side short creepy morning bit

[小栗说车日常记 @Lillian522]

@sydlovesfob yeah i no same here woowww twitter is really confusing!