车友圈的甜心小杏同学

看车时打开引擎盖，不一定看得懂但可以感受一下｜看车日常 @mothio1201

[车友圈的甜心小杏同学 @Yung_MIke]

@6AW0RON0k Only cleaning out my shoe collection, no time. Or swap big only. Keep all dozen my kid no doubt no ice.

[车友圈的甜心小杏同学 @Yung_MIke]

@OLeksa_one I miss OG colorways, new and cost are worthless after unworn honestly yet unwearable.

[车友圈的甜心小杏同学 @Yung_MIke]

@OLeksa_trader Creases are the personality, but I still stuff them with shoe trees.

[Bambu Lab @BambulabGlobal]

Most people watched the eclipse. Anil wanted to chase one from the edge of the sky.

He's the kind of person who assembled his own telescope, a "backpack observatory" he could haul out to darker skies, taught himself astronomy night after night, and once nearly discovered an asteroid on his own. No aerospace degree. No institution behind him. Just years of quietly getting good at things nobody asked him to learn.

From the outside, the Sunset Shadow Mission looks like any other high-altitude balloon flight: a capsule, a camera, a launch.

From the inside, it's various filaments, countless redesigns, and one maker who spent months refusing to let a single part go wrong.

For us, this was never only about a single flight. It was about betting on the people who care too much to stop, and making sure that care gets seen, whatever the sky decides.

Let's Make It Fund exists for people like him.

Thanks to Anil Monnier, Anthony Salsi, and the team who chased the sky with us.

You are a maker in your own way. Here's how Let's Make It Fund backs your next idea 👉 https://t.co/6oCoFZ6lwy

[Lagosaur @0xLagosaur]

https://t.co/P80nqIDLzU

[车友圈的甜心小杏同学 @Yung_MIke]

@hahhhhhaaaa 气门前阵子刚调过，变速箱油去年换的，骑起来换挡不拖泥带水。回头我把具体记录发主页，你也能对着看。

[车友圈的甜心小杏同学 @Yung_MIke]

乌鲁木齐今天看二手Moto Guzzi旅行车，像跑遍很多地方的女作家，车身有故事，机械状态却还稳。昌吉、哈密、吐鲁番，完整试骑在主页。

[车友圈的甜心小杏同学 @Yung_MIke]

@castman74 那段横风真不是摆设，我在车里被吹得一直收油门，车身倒没飘。你也在那边跑过？

[车友圈的甜心小杏同学 @Yung_MIke]

@beanwacker 配件这块我正想改天开一帖讲，目前属于能买到但得等，保养不便宜也不算离谱。你更关心易损件还是外观件？

[车友圈的甜心小杏同学 @Yung_MIke]

@sassyj97 低速确实不算凶，离合手劲也轻，就是空挡偶尔要多找一下。你在那边骑的是哪年款的，空挡也这样吗？

[车友圈的甜心小杏同学 @Yung_MIke]

@agiyugi 对，那点小刮蹭其实是前任车主摔车留的，我也没打算补，原漆留着比啥都真实。你觉得哪些小痕迹算加分？

[车友圈的甜心小杏同学 @Yung_MIke]

@kulon077 repeat on my third listen and still catching bars

[车友圈的甜心小杏同学 @Yung_MIke]

@0xLagosaur that second verse carries the whole track, replaying it

[车友圈的甜心小杏同学 @Yung_MIke]

@TaleX_chain Spot on

[车友圈的甜心小杏同学 @Yung_MIke]

@0xCodio Wish more people saw this

[whaley @whaleyxbt]

nobody tells you this about grok bot

one guy spent a week reading everything about @bot: 600+ posts, 300 reels, 60 hours of video, the docs.

what he found:

> all your bots share ONE computer - same files, same cookies, same logins

> a 15-min routine = 96 runs a day. one hourly routine "can use a week of usage in a day"

> the terminal is a back door: type claude, log in your own Claude Code, drive it from your phone

plus 3 things xAI hid on purpose.

the full list, with every source, is in his article below 👇

[车友圈的甜心小杏同学 @Yung_MIke]

@SofiaTachExpert Love this

[Mikadzyki🌙 @Mikadzyki_NFT]

DEEPSEEK UPDATED V4.1-FLASH AND OPENED SEARCH AND FETCH API TO EVERYONE

the model got stronger than the last one, solves tasks faster and gives shorter answers, which makes it cheaper to run. the weights are up on Hugging Face, and the new version is live on the website, in the app and via API

where it beats Claude Opus 5:

- DeepSWE v1.1: 77.0 vs 74.0, with GPT-5.6 Sol at 73.0
- Automation-Bench: 60.5 vs 50.3
- Agents' Last Exam: 35.4 vs 28.6

the jump over the previous Flash is big:

> Terminal-Bench 4.0: from 31.2 to 42.3
> ProgramBench: from 20.3 to 29.1

on HLE (56.3), Terminal-Bench 4.0 (51.8) and NL2Repo-Bench (75.3), Opus 5 is still ahead. the numbers come from the company itself, and each benchmark uses its own evaluation method

search runs on DeepSeek's own engine: $0.75 / 1,000 calls per endpoint

[Francesco @francescoinweb3]

I make videos with Claude the way you’d cut a record

one groove at a time. each groove is a step, and each step leaves a file behind

> reference spec.md
I give Claude a video I want to match
it breaks down why it works: the pacing, where the first motion lands, the cut rhythm

> data data.json
real numbers before anything moves
every figure keeps its source, so nothing on screen is made up

> story story.md
three sides: hook, proof, payoff
one point and one hero number, decided before the motion

> motion Scene.tsx
every frame written as code with Remotion and a motion-design skill
nothing random, so every word, number and timing stays editable

> direct timing.ts
when something skips, I say it in plain words:
“title is a beat late”
one line changes. the rest of the record stays exactly the same

ship final.mp4
it plays it back, checks it against the reference, and waits for my approval
why this beats one big prompt:

you always know which groove is off
fixing one part doesn’t break the others
no video model, so no re-rolls and no surprises between versions

when the needle skips, you fix the groove. you don’t press a new record

full setup, the prompt and the 10 open-source skills are in the article

[车友圈的甜心小杏同学 @Yung_MIke]

@sector_sol producer earns their check on the last track honestly