车友圈的知性小星同学

车内的手机支架位置，导航看着顺眼最重要｜汽车日记 @mothio1201

[Robtimus @_robtimus_]

31 YEAR OLD DIRECTOR IN SEOUL RENTED A $22,000 ROBOT TO SAVE ON STUNTMEN… AT REHEARSAL HE FORGOT TO TELL THE ROBOT IT WAS FAKE

The scene was simple. Three guys with sticks attack, the hero falls nicely, everyone goes to lunch

The robot got the script, but the word "staged" wasn't in it once

The first one swung and was on the mat a second later
The second thought it was a glitch and flew to the floor
The third saw the robot raise a finger for "one sec", then took a kick to the chest

9 seconds. Three stuntmen down. The choreographer yells "that wasn't in the plan"

Now the director is paying three people sick leave, and the robot stands in the corner like it just nailed the best take of its life

Would you be the fourth guy in this scene ?

[车友圈的知性小星同学 @TRAY_TG]

@Oxylabs_io Anybody cop from that drop or was it instant bot territory?

[Reuters Business @ReutersBiz]

WATCH: Customers queued in Shanghai to pick up their orders of Apple's first folding phone, the iPhone Duo, at Asia's largest Apple store, with market watchers betting on whether the device can reinvent the small foldables market https://t.co/C3WO9OTrf8

[车友圈的知性小星同学 @TRAY_TG]

@therealsexy17 不是视频，主页就是图文的完整版；车身紧不紧光靠图片确实难讲，你担心的话我把车架和摆臂这些地方多拍几张——你看哪块最想先确认？

[车友圈的知性小星同学 @TRAY_TG]

哈尔滨这台二手杜卡迪Scrambler车漆有轻微使用感，车身却没显松，油门回应也还很精神。大庆、佳木斯、齐齐哈尔想看完整体验，主页继续。

[车友圈的知性小星同学 @TRAY_TG]

@biggesttaylor 对，车架我就懒得绕弯子说，确实是这台车还能骑的部分；你要是也想看具体细节，回头我挑一处慢慢说。

[车友圈的知性小星同学 @TRAY_TG]

@gtd1988 这块我也说不百分百，光看照片只能看出表面无橘皮，最好还是当面试。你想要哪几个位置的细节我拍给你。

[车友圈的知性小星同学 @TRAY_TG]

@disgustedheart 路试过的，二挡拉到红区前转速上得干脆，原地捏只能说初步正常。你平时是在城里通勤多还是拉长途？

[车友圈的知性小星同学 @TRAY_TG]

@LMR236 对，薄是薄，我扒细看过几处边缘没补漆痕迹，使用感基本是蹭衣裤磨出来的。你要细节图的话我再补几张？

[车友圈的知性小星同学 @TRAY_TG]

@shiaaee 对，压车听着咯吱响那种多半是胶套或转向头的事，这台专门试过没那动静。你看那台最后收了吗？

[车友圈的知性小星同学 @TRAY_TG]

@chris_pisz12 大庆冬天搁车库确实省心，只要电瓶别亏就没事，你那边冬天一般多少度停多久？

[rompasso @0xrompasso]

Opus 5.5 is 20% cheaper per token than Opus 5. Part of the rest of the discount comes from a setting nobody asked you about.

This isn’t a smaller model hiding behind a new number.

It tops Opus 5 almost everywhere Anthropic published, 66.4% on Terminal-Bench 4.0 against 52.3%. The catch sits in the defaults, not the weights.

What changed between the two that won’t show up in your logs:

→ Default effort dropped from high to medium, so any call that never set effort now thinks less on the exact same prompt

→ Adaptive thinking is always on and can’t be switched off, and every thinking token bills at the $20 output rate

→ The “about 40% cheaper on typical workloads” figure leans on the model finishing in fewer tokens, not on the rate card

→ Setting tool_choice to any or to a named tool now returns HTTP 400, the one change loud enough to find on day one

One hands-on test ran three DevOps tasks on both models. Opus 5.5 used fewer output tokens than Opus 5 on just one of the three, and still came in cheaper overall, $0.090 against $0.117, on price per token alone.

The loud break gets fixed by lunch.

The quiet one ships to production as a slightly shallower answer that nobody thinks to trace back to a default.

Which settings to pin before you migrate a single call is in the writeup below 👇

[Lovart AI @lovart_ai]

✨ Meet Arti, your personal design agent. New in Lovart 2.0.

Arti knows your taste and remembers how you work. Schedule a brief before you log off, and wake up to research and drafts ready to review.

The more you create, the more Arti gets you.

[车友圈的知性小星同学 @TRAY_TG]

@mirku21 soldering iron still rocks but tool new pad falling off

[Daniro @Dan1ro0]

This guy used Grok Bot to build a Quant Trend system and it made $526 a day on Polymarket

The result is already +$69,953 PnL in 133 days with about $86 per average trade, repeating the same setup 48 times every hour across BTC "Up or Down" markets

Several AI agents work together as one algorithm:

1. DMI determines which direction currently has control

The bot compares positive and negative directional movement:

+DI = 100 × (+DM / ATR)
-DI = 100 × (-DM / ATR)

Then:

+DI > -DI → Up bias
-DI > +DI → Down bias

The wider the distance between the two lines, the clearer the directional advantage

2. ADX confirms whether that direction has enough strength

The system calculates:

DX = |+DI - -DI| / (+DI + -DI) × 100
ADX = Average(DX)

A rising ADX means the directional move is becoming stronger
A weak ADX means BTC has no clear trend, so the bot ignores the DMI signal

This is the first agent that will create a trading bot strategy for you and then build it:https://t.co/8tvtHFdE4P

This trader’s Polymarket account: MayorMamdani

Most profitable trades:
$1,671 → $3,247 (+$1,576)
$621 → $2,017 (+$1,396)
$1,000 → $1,973 (+$973)

The system uses DMI to choose Up or Down and ADX to confirm that the move has enough strength behind it

[Hedgeye @Hedgeye]

h/t: https://t.co/C5lw9OO9kW

[Spike 1% @SpikeCalls]

This bee is flying around with a fruit fly's brain in its head, and it has no idea.

We started with a normal honeybee, with fuzzy stripes, six legs and wings.

Then we zoomed in through its compound eye, all 5,500 facets of it.

We went through the glass chip sitting on top of its head.

Inside the chip, the brain isn't a bee brain.

It's a fruit fly brain.

There are 139,255 neurons in it, and every single one was mapped from a fly.

There are 54.5 million synapses, all wired exactly the way a fly's are.

There are 8,453 cell types, and none of them belong to a bee.

The optic lobe is trying to see through bee eyes.

The antennal lobe still smells like a fly.

The mushroom body is learning how to fly a body that's 100 times heavier.

Meanwhile its own 1,000,000 bee neurons are sitting there, completely switched off.

The brain still thinks it's a fly.

The body is very clearly a bee.

It's flying anyway.

So who is actually flying this bee.

[Asteri @Asteri_eth]

this list is straight up f*cking gold for motion design

20 open-source GitHub repos to run your own motion design studio with Opus 5.5

one pick per layer: agent skills, animation, 3D, assets, sound and export

01 Remotion → write React, get real editable video
https://t.co/2MnpZQrLlw

02 Remotion Skills → official video recipes for Claude Code
https://t.co/nz3anwkCYf

03 Motion Canvas → code explainers and animated diagrams in TypeScript
https://t.co/rWQTitvtlP

04 Motion AI Kit → feed Claude Code the Motion docs and best practices
https://t.co/FlFx1A7NhH

05 Motion → smooth layout transitions and gestures in live UI
https://t.co/1orZ6aMbHM

06 GSAP → frame-precise timelines for text and SVG
https://t.co/HhwPzpCe1Y

07 Anime.js → draw SVG paths, split text, move UI pieces
https://t.co/fFxawh6xPV

08 GSAP Skills → teach Claude Code clean timelines and React teardown
https://t.co/HdSTwCBEXF

09 Rive Web → vector animations that react to app state
https://t.co/5VTFCZi5YK

10 dotLottie Web → play, pause and steer Lottie files
https://t.co/IDX2Ulm0y8

11 Three.js → 3D shots with real cameras and lights
https://t.co/KQBagkuShd

12 React Three Fiber → write 3D scenes as React components
https://t.co/efNx10EPLX

13 Drei → ready-made cameras, 3D text and controls
https://t.co/9j6H7yy6w6

14 React Three Rapier → physics: collisions, joints, rigid bodies
https://t.co/IeWy9Z5OxI

15 React Postprocessing → bloom, depth of field and the final polish
https://t.co/cEE83wAzIq

16 SVGO → strip the junk out of SVGs before you animate them
https://t.co/qCv95NeftJ

17 Satori → JSX in, SVG title cards out
https://t.co/WozmVoiIPe

18 Howler.js → sound effects, fades and volume for web demos
https://t.co/9AVsnL9zLr

19 Mediabunny → encode and convert video right in the browser
https://t.co/bX2k89OrQS

20 FFmpeg → final encode, resize and file checks
https://t.co/bfqyckTvNL

the workflow:

brief → typography → motion → scene → sound → render → review → ship

bookmark this + the article below shows how to wire up your own motion studio

[车友圈的知性小星同学 @TRAY_TG]

@RareSoulMammaai that clack is chef's kiss, what plate is this?

[车友圈的知性小星同学 @TRAY_TG]

@Sai60012 @grok lubed my switches and now I need zero advice from anyone